Burnac Corporation v. United Dominion Industries Corporation

Burnac Corporation v. United Dominion Industries Corporation Court Supreme Court Date 2007-02-08 Citation 2007 NSSC 40 Docket SH 160320 Judge/Registrar/Adjudicator Coughlan, C. Richard (Honourable Justice) Document Type Decision Relations Library Sheet - Burnac Corporation v. United Dominion Industries Corporation - 2007 NSSC 40 - 2007-02-08 - Library Sheet Decision Content IN THE SUPREME COURT OF NOVA SCOTIA Citation: Burnac Corporation v. United Dominion Industries Corporation, 2007 NSSC 40 Date: 20070208 Docket: SH 160320 Registry: Halifax Between: Burnac Corporation Plaintiff v. United Dominion Industries Corporation Defendant Judge: The Honourable Justice C. Richard Coughlan Heard: October 2-4, 10-12, 16-19, 23-25, 30-31, November 1-2, 6-8, 14-16, 2006, in Halifax, Nova Scotia Decision February 8, 2007 Counsel: Aidan J. Meade and Kevin D. Gibson (Genevieve Dawson on October 10-11, 2006 only), for the Plaintiff Michael J. Wood, Q.C., Ronald S. Noseworthy, Q.C. and Cory Withrow, for the Defendant Coughlan, J.: [1] The roof of the Sears store at the Village Mall in St. John’s, Newfoundland collapsed February 9, 1987. Then, in January, 1995, the roof of the Centrac warehouse in Mount Pearl, Newfoundland collapsed. Both buildings were constructed with open web steel joists fabricated by Robb Engineering of Amherst, Nova Scotia. The Association of Professional Engineers and Geoscientists of Newfoundland notified professional engineering associations across Canada there could be a structural problem with open web steel joists fabricated by Robb Engineering between 1970 - 1985. The Association of Professional Engineers of Nova Scotia provided the Province of Nova Scotia with a list of contracts in Nova Scotia that contained Robb Engineering open web steel joists. The Nova Scotia Department of Housing and Municipal Affairs wrote owners of buildings containing Robb joists, recommending open web steel joists fabricated by Robb Engineering between 1963 and 1985 be inspected by a qualified professional engineer. [2] The owners of the Mayflower Mall, Sydney, Nova Scotia received the notification. Consulting engineers were engaged to verify the presence of Robb joists in the Mall. An inspection was carried out and a report prepared. [3] The report stated that of 1,900 open web steel joists which could be inspected, a total of 199 or 10.5% were found to have at least one panel point with a broken or missing weld. The average for buildings inspected in Nova Scotia containing Robb joists was 2.28%. The evidence of experts for all parties is a cracked or missing weld is a danger and must be repaired. The report set out a proposed remediation scheme, which was the same as implemented by the Nova Scotia Department of Transportation and Public Works for the Government’s own buildings. The owners, concerned with safety issues arising from the report, followed the consultants’ recommendation and undertook the remediation of the Mall. [4] Burnac Corporation, the owner of the Mayflower Mall, claimed against United Dominion Industries Corporation, of which Robb Engineering was a division, for the cost of the inspection and remediation of the Mall. United Dominion Industries says the actual repairs to the welding and webs of the open web steel joists were necessary, but the reinforcing of the joists undertaken in the remediation was not necessary to repair defects to the joists and mitigate the danger. [5] The issues for the Court are: - Were there defects in the open web steel joists manufactured by the defendant? - Was there a real and substantial danger to the occupants of the Mall? - Was that danger caused by the defendant’s negligence? - Were the repairs and remediation undertaken by Burnac required to alleviate the danger? [6] The Mayflower Mall was constructed in 1979. United Dominion Industries Corporation, formerly United Dominion Bridge Company Limited, admits it fabricated, transported and erected the open web steel joists in the Mayflower Mall. [7] Robb Engineering, a wholly owned subsidiary of United Dominion Bridge Company Limited (Dominion Bridge), was located in Amherst, Nova Scotia. Robb had various divisions. One division was engaged in the fabrication of open web steel joists. The joists at the Mayflower Mall were fabricated by Robb Engineering. [8] Carl Freeman Smith was Welding Supervisor at Robb Engineering from approximately 1967/1968 until he retired in 1985. He supervised welding in all Robb Engineering divisions. Approximately 25% of the welding at Robb Engineering was on the joist line. As a general rule, less experienced welders worked on the joist line. There were welding data sheets prepared by the Engineering Department to provide welders with information concerning a particular weld a welder was to make, including type of joint, number of passes to be made and type of electrode to be used. Mr. Smith kept the data sheets on the walls of his office so the welders could see them. There were no welding data sheets for joists. [9] The joists had top and bottom chords, each chord had two lengths of steel plate on angles called chord members, with round bars between them called web members. The web members connected the top and bottom chords. The web members were welded to the chord members. Where the web and chord members are connected by welds are called joints, panel points or nodes. Joists were of various lengths. [10] Joists were assembled in a jig. The top and bottom chords on one side were placed in the jig. Then the web rods were put in place and tack welded. Next, the remaining top and bottom chord angles were put in place. The joist was clamped in the jig and flipped. The joist was then standing up. Two welders welded one side of the chord angles. One starting at the end of the joist and working toward the centre, the other starting at the centre and working toward the other end. When one side was finished, the joist was flipped and the other chord was welded. The welds used were flare beveled and puddle welds. [11] Once the welders were finished, the joists were put on a trolley and went to be painted and dipped in a tank of paint. There were no inspection of the joist between welding and painting. [12] There was no standard way to pile joists for inspection. The joists were in a pile which allowed an inspector to see either the top or bottom chord, depending on how the joist was placed on the pile. What the inspector could not see was not inspected. The amount of work in the plant determined how much Mr. Smith inspected. If he was busy, he had no time to inspect joists. Mr. Smith said Norman Ellis was an inspector who he thought inspected joists once a day. Mr. Ellis did inspections other than joist inspections. There was another person who did joist inspections, but Mr. Smith did not know how often he conducted inspections. Mr. Smith agreed on cross-examination there could be joists not inspected. [13] Michael John Pates was employed by Dominion Bridge from June, 1976 until September, 1986. A metallurgical and welding engineer, he became welding engineer for Robb Engineering in May, 1978. Prior to that he had worked on the welding at Robb Engineering. Mr. Pates visited Robb Engineering about four or five times a year, each visit lasting two days. Robb Engineering was one of Dominion Bridge’s smaller operations. [14] At Robb Engineering, the shielded metal arc welding process, also known as stick welding, was used in the fabrication of joists. [15] Mr. Pates thinks his first visit to Robb Engineering was on May 19-20, 1977. He observed problems with the welding of open web steel joists. The joists could not be inspected to determine if a weld was acceptable as the welds were covered with slag. During welding, a weld is in a molten state. Slag floats to the surface and covers the weld. Slag should be removed. If slag is not removed, it hardens and makes it impossible to inspect the weld. With slag present, one could not see if there was a cracking or fusion problem with the weld. He had been told by his colleagues at Head Office there was a problem with welding at Robb Engineering. [16] Following his visit, Mr. Pates wrote a letter dated May 31, 1977, addressed to Mr. L. H. Naylor, Robb Engineering, Amherst, the subject of which was “visit re welding practice and procedure”, which stated in part: As you are aware, the situation regarding the control of the quality of welding during the fabrication of joists is not satisfactory. This is due, in part, to the non-removal of slag after shielded metal-arc welding of the flare-groove and puddle welds, which makes it virtually impossible to inspect for weld length and freedom from defects. Whilst appreciating that a program does exist to provide an ongoing quality surveillance, this clearly cannot be performed successfully on welds which are masked by slag deposits. Further areas of concern are: 1) The control over throat sizes of flare-groove joints when no approved weld data sheets exist designating welding parameters proven to have produced deposits meeting the design throat thickness. 2) The use of Jetweld #3 electrodes which are not certified by the Canadian Welding Bureau. . . . . Successful joints could be established on the basis of direct throat measurements on sectioned welds deposited under controlled welding conditions. Acceptable welding parameters could then be submitted to the Canadian Welding Bureau for approval. [17] Mr. Pates was trying to get the joist line under control. [18] He then visited the Robb Engineering facility on September 29-30, 1977. The welding in the joist line was the same as during his previous visit. [19] The joist line remained a problem during Mr. Pates’ visit to the Robb Engineering plant on January 24-25, 1978. There was no control over how the welds were made. [20] Mr. Pates made a further visit to Robb Engineering on April 18-19, 1978. A meeting was held, attended by Mr. Pates, Lionel Naylor of Robb Engineering, Carl Smith, Robb Welding Supervisor, R. Fisher and R. Golec, Seniors Managers of Robb. Mr. Pates summarized the discussions concerning the joist line in a letter to Mr. Naylor dated April 24, 1978 as follows: Joist Line The comments of our initial meeting held in the earlier part of this year were pursued further. In addition to ourselves, R. Golec, S. Fisher and C. Smith were present during these discussions. Various aspects relating to the joist fabrication were discussed and it was agreed that welding consumables used on the line must be brought in line with code requirements. It was decided that the GMAW process was the best candidate, and that Robb should forward flare groove throat sizes, lengths, etc. which would satisfy the latest addition of W59.1 where specific reference to flare groove joints is made (see attached copy). Once this design information is available then work on GMAW deposits, using representative joints, may proceed in the laboratory. The distortion problem, where some joist (™40') were experiencing a sweep in the top chord and an “S” type sweep in the bottom chord was traced back to the rigidity of the jig being used in the shop. This jig needs to be redesigned in order to maintain chord tolerances. [21] Mr. Pates again visited Robb Engineering June 14-15, 1978. At that time the joist line was not in operation due to lack of work. Mr. Pates stated that from his recollection while other areas of the Robb Engineering facility were not bad, the joist area felt like it was off limits. It was dirty. [22] There was an ongoing attempt to improve the welding process on the joist line. A joist line program was established to investigate changing the welding process to bring the line within the National Building Code without the need to “deslag” the welds. [23] Tests were carried out and a report entitled “Joist Assemblies - Robb Engineering. An Investigation into Present and Alternate Welding Processes” dated July, 1979 was prepared. Mr. Pates was an author of the report. The purpose of the test and report was to determine if an alternate welding process could be used to meet the Code without the need to deslag. Samples of production welds from Robb Engineering were sectioned and examined. Weld metal cracks were evident in the samples examined. The report concluded, a strict quality control program needed to be established at Robb Engineering. Such a program was not implemented during Mr. Pates’ involvement with Robb Engineering. The report recommended all joists be deslagged to allow for routine inspection for weld quality and weld length. Mr. Pates stated the report raised concerns in his mind, but he was told by Messrs. Fisher and Golec there was no need to deslag as there had never been a failure, load tests had been performed and joists performed satisfactorily. Prior to the report, no deslagging was allowed. [24] During a visit by Mr. Pates to Robb Engineering on September 25-26, 1979, a presentation was made to Robb senior management dealing with joist line problems. It was decided action was to be taken to improve the quality of welding on joists. In a letter to Mr. L. Naylor dated October 2, 1979, Mr. Pates set out the course of action agreed upon: Further tests on production joists also showed that many of the end joints, where the larger diameter bars had been used, contained similar linear defects. After discussions with yourself and Carl Smith, a presentation was made to R. Golec, R. Crossman and S. Fisher. The outcome of this meeting concluded that, effective immediately, the following steps would be undertaken to bring the welding of joists to a quality acceptable to Engineering. 1. Electrodes will be changed to E7014 and all craters will be filled. 2. The three critical end joints will be deslagged and inspected on all joists. These will be subsequently inspected by MPI until we have the assurance that the cracking problem has been overcome. 3. One joist per shift will be deslagged completely and inspected by Carl Smith. MPI inspection may be performed at Carl’s disgression (sic). Additionally, the following work will be performed: a) Carl will carry out tests to find the best parameters for flare groove joints. Data sheets will be made up for these and welders will carry out trial welds which will be sectioned at the centre and end to assure that penetration is achieved. Welders will carry out such tests at regular intervals; this approach to checking the quality of their work will mean that it becomes unnecessary to section production joists at this time. b) Engineering would look into the possibility of increasing the bend radius of the webs to allow for less downhill welding. c) Management indicated that they would be looking into the procurement of new bending equipment in the coming year. This would be extremely beneficial in reducing bad fit up and would therefore increase weld quality. d) All joists incorporating 1" diameter and 15/16" diameter bars at the critical end joints, and which are now lying in the yard, will have these welds repaired according to the procedure agreed between Carl and myself. e) A document incorporating welding data sheets and correct welding and inspection procedures will be drawn up by ourselves to act as a control document for the short span joist line. The meeting concluded that the situation regarding inspection should be reviewed with time once the weld quality had improved to our satisfaction. It is my opinion, however, that the decisions made at the meeting represent a minimum inspection and that we should maintain such a system for all future joists. During my visit, we also discussed the puddle joints on the joist line, and it was thought necessary that we should carry out some mechanical testing of typical joists in the R&D Centre at Lachine. I enclose a copy of Report No. L106/1/73 which describes the testing of flare groove joints and this may be of help to you in the design of a relevant jig. [25] Testifying about the above items, Carl Smith stated with regard to item numbered one, the electrode was changed and some of the craters filled. With regard to item numbered two, the three end joints were deslagged and inspected on all joists. The MPI inspection was carried out. With regard to item numbered three, one joist was completely deslagged and inspected by Mr. Smith on a few occasions. Due to his workload, Mr. Smith was unable to say how many times he conducted such an inspection. Mr. Smith testified the work set out in items a) and e) was not done. Items b), c) and d) were done and, with regard to item d), insufficient welds on the end joints were found and the welds were increased. [26] Mr. Smith was aware Robb Engineering advertised their joists as complying with CSA (Canadian Standards Association) W59 Standard (Welded Steel Construction (Metal Arc Welding)). He was aware the joists were not fabricated in accordance with it. He was told by Robb management they could not make joists in compliance with W59 and stay competitive. [27] Mr. Pates complained that contrary to CSA W47.1 (Certification of Companies for Fusion Welding of Steel Structures), Carl Smith had too many duties. The Standard required a welding supervisor be employed full-time in that position. [28] Mr. Pates stated he had concerns joist welding may be a danger to the public, but did not inform his supervisors of his view. Mr. Pates tried for over two years to have data sheets for joist welding produced. Although Mr. Smith knew CSA W59 required data sheets be prepared, welding data sheets for joists were not prepared. [29] An Advisory was issued by the Department of Housing and Municipal Affairs and sent to owners of buildings thought to be constructed with Robb open web steel joists. As a result of the Advisory, the owners of the Mayflower Mall commissioned an inspection of the Mall. CBCL Limited, assisted by fga Consulting Engineers Limited, was retained to conduct the inspection. The inspection took place from September to November, 1998. To limit the disruption of Mall operations, for the most part the inspections took place between 11:00 p.m. and 6:00 or 7:00 a.m. James Bradley Kennedy, a civil engineer employed by CBCL Limited, was lead engineer on the project and involved in preparing the inspection report. The purpose of the inspection was to inspect the joists in the Mall and look for cracked, broken or missing welds. Mr. Kennedy also inspected some joists. Most of the joist inspections were conducted by Darrin Frederick McLean, a professional engineer employed in the Sydney office of CBCL, and by Kevin James Pinhorn, a Level II welding inspector, certified by the Canadian Welding Bureau. Mr. Pinhorn was also a certified welder and pipe fitter. Mr. Pinhorn was employed by fga Consulting to work on the project with CBCL. Each of Messrs. McLean and Pinhorn testified he inspected about 50% of the joists in the Mall. An inspection sheet was prepared for each weld inspected. The inspectors were looking for bent web and chord members and missing or cracked welds. [30] The same procedure was used by all inspectors. The bottom chord of the joists was approximately two feet above the 14 foot ceiling, with the top chord 26 to 30 inches above the bottom chord. Ceiling tiles were removed and the inspectors climbed stepladders which were as high as 14 feet. As there was no light above the ceiling tile, flashlights were used. Inspectors also used prybars, approximately 18 inches in length. The inspectors looked at a joint to determine if a weld was present. If a weld was present, the prybar was placed between the web and chord member and pressure applied to determine if there was separation. If there was movement, that indicated a cracked weld. If the weld was sound, nothing happened. [31] All accessible panel points on all joists were inspected. There were difficulties in conducting the inspections. Mr. Kennedy testified some of the joists were inaccessible and could not be inspected. Some of the welds on the top chord could not be seen as the deck of the Mall’s metal roof was sitting on the top chord. Where possible, on the top chord, the top of the weld was inspected with the aid of a small inspection mirror. [32] Mr. McLean testified not all welds were visually accessible because of duct work, piping or where the corrugated metal roof rested on the top of the joist. In addition to resting on the top of the joist, the roof was spot welded to the joist. The joists were painted which affected their visibility. There were also cracks on the bottom chord not visible without using a prybar. In general, it was difficult to see cracks on the top chord. The inspection mirror could not be used to examine welds covered by the roof deck. [33] Where the roof deck was attached to the top chord, the length and width of the weld could not be measured. The inspectors were not able to inspect all welds or joists, and some welds or joists were enclosed by drywall or duct work. [34] Mr. Pinhorn testified it was physically impossible to measure the length or width of welds on top chords as he could not get a tape measure there. Where the roof deck rested on the top of the joist, he could not see welds. Where the roof deck was raised, he could use a mirror. Not all areas of the joist could be accessed because of obstructions, such as gyprock, conduits and wiring. In those areas he could not see the welds. There was slag all over the joints on the bottom chords, as well as the ones he saw on the top chords. Where slag covered welds, he could not see the welds. [35] While Mr. Pinhorn was not asked to determine the quality of the welds, he observed weld quality was poor, with porosity, overlap and undercut present. He did not record all porosity, slag or undercut, but testified it was persistent. He described some of the welds as being as “thin as an eggshell”. [36] After inspection of the joists was completed, CBCL prepared a report and fga Consulting commented on it. The report stated, “In general, welding quality was poor with weld failures and poor fit up noted throughout.” Of the 1,900 open web steel joists inspected, a total of 199 or 10.5% were found to have at least one panel point with a broken or missing weld. Also 84 joists were found to have one or more bent web members and 40 joists had one or more bent chord members. It was Mr. Kennedy’s opinion the majority of the bent members occurred during construction. One web member was missing and another was broken. The report recommended the remediation scheme implemented by the Nova Scotia Department of Transportation and Public Works be adopted for the Mayflower Mall. [37] Gary Follett, President of fga Consulting Engineers Limited, was qualified as an expert in welding engineering matters and structural engineering, entitled to give opinion evidence in those areas. Mr. Follett is a professional engineer. He has extensive experience with Robb joists. [38] After the partial collapse of the roof at the Village Mall, fga Consulting was hired to inspect the part of the Mall that had not collapsed. The company became more involved until the Mall owners relied solely on fga Consulting concerning engineering matters respecting the collapse. Fga Consulting carried out a detailed inspection. The company found cracked joists and welding errors. Welds full of porosity and with very little penetration were observed. There were a lot of faults with the welding. He saw puddle welds from the Village Mall which were little more than a blob of metal on an end of a rod. [39] In January, 1995, the Centrac roof collapsed. Fga Consulting was hired to assist with the inspection and advise regarding welding matters. The quality of the welds inspected was poor. As Mr. Follett said, there were more problems than normal, even for a building constructed with Robb joists. [40] Mr. Follett was appointed to the Open Web Steel Joist Review Committee for the Province of Newfoundland. The purpose of the Committee was to determine how the problem concerning open web steel joists was to be addressed. The Committee visited every Newfoundland government building containing open web steel joists. If a building contained Robb open web steel joists, it was inspected and a remediation program developed. [41] Subsequently, Mr. Follett was hired as a consultant to the Open Web Steel Joist Working Committee established by the Province of Nova Scotia. The Committee’s mandate was to assess the Robb open web steel joist issue, determine if a problem existed in Nova Scotia and recommend the appropriate action to correct any problems. A report dated July 8, 1998 titled “Open Web Steel Joist Investigation Interim Remediation Report” was produced. Mr. Follett authored the report. The report dealt with the proposed remediation of buildings owned by the Province of Nova Scotia constructed with Robb open web steel joists. [42] The inspection of public buildings in Nova Scotia constructed with open web steel joists fabricated by Robb Engineering resulted in the determination 2.28% of joists had at least one failed joint, that is, at least one cracked or non-welded joint. The comparable failure rate in Newfoundland was 8.3%. [43] The report stated there are in excess of 22,000 joists in public buildings containing more than 65,000 panel points or joints. It was assumed every second or third top chord panel point was inaccessible for inspection due to the down flute of the building’s roof and, therefore, between 113,000 and 168,000 joints could not be inspected. Additional joints were inaccessible due to the location of mechanical equipment. [44] Section 4.0 of the Nova Scotia report contains a discussion of inspection results and states in part: The theory of brittle flare bevel welds contributing to the weld cracks is supported by the random nature of the location of these cracks. Figure 5 seems to imply that the cracks have occurred throughout the joist as opposed to the more highly stressed location which in turn seem to suggest that many of the cracks are not service related. Joists in length categories 5 and 6 (10 m to 14 m) have more cracks on a percentage basis. Initially this would appear to be service related. Such may not be the case, these failures may not be related to the susceptibility of longer joists to cracking during handling, or to the fact that the longer joists have larger members, which means faster cooling rates and thus higher hardness. Once again further investigation to determine the mode of failure would be advisable. Figure 7 certainly does indicate a marked increase in failures above length category 4. This suggests additional remediation may be advisable around the 9 m or 10 m length and above. [45] A remediation scheme was proposed with buildings categorized on the basis of the percentage failure and age. In buildings that had satisfactorily performed for twenty years and had a failure rate of 1% or less, only failed welds or members would be repaired. The remediation scheme was divided into various levels as follows: Proposed Joist Remediation (Roof & Floor) Category Description Suggested Remediation (Prel.) Level I Includes any joist: 1) with span exceeding 9m, or 2) in a snow build up area (ie. high low roofs) or 3) in high floor load areas (i.e. libraries, stockroom, etc.) or 4) in an area with excessive failures, or 5) that has more than two (2) failed joints. 1) Carry out 100% inspection and make necessary repairs to joints. 2) Install spacers in compression chords where force is $50% of the maximum force. 3) Reinforce members/panel points on each end until resultant shear force $50% of end value. Level II Includes any joist: 1) with span between 4m and 9m inclusive, and 2) that is in a non-snow build up area. 1) Carry out 100% inspection and made necessary repairs to joints, and 2) Install spacers in compression chord where force is $50% of the maximum force. Level III Includes any joist: 1) with span less than 4, and 2) that is in a non-snow build up area. 1) Carry out 100% inspection and make repairs. Level IV Includes any roof joist: 1) designed to 0.6 snow load factor and, 2) with failures or obvious deficiencies. Note for joist designed to 0. with no failures or obvious deterioration, treat at Level I, II, or III above. 1) Analyse and upgrade to 0.8 snow load factor Note: 100% inspection for joint failures must be carried out for all buildings, regardless of category. [46] During remediation, Mr. Follett visited the Mayflower Mall. He looked at existing welds, as well as the remediation. The existing welds were typical of what he saw elsewhere. He saw undercut, porosity, slag not removed, cracked welds and missing welds. There was inconsistent profile of welds present. The weld would meet requirements and then get narrower. He observed craters. Mr. Follett did not see any evidence of base metal flaws or overstressing at the Mall. [47] Mr. Follett said, in construction, one first turns to the National Building Code, which is a Code of minimum requirements for buildings or structures. An engineer should design to the Code. A company should build to the Code. The Code requires conformation to CSA S16.1 (“Steel Structures for Buildings”), which requires welding design and practice to conform to CSA W59 (“Welded Steel Construction”) and that the welding be carried out by a company certified to CSA W47.1 (“Certification of Companies for Fusion, Welding of Steel Structures”). This was summarized by Mr. Follett in his report “Open Web Steel Joist Remediation, Mayflower Mall, Sydney, Nova Scotia” dated January, 2004 as follows: Essentially, for structural steel units such as OWSJ’s to be termed “safe” it should have at least been welded by a company certified to CSA W47.1 who should have followed welding procedures which were approved by the Canadian Welding Bureau (CWB) as the W47.l Standard administrator. Through the approval process for these procedures, the company should have verified that a competent weld could be produced. The certified company should have then used welders who had current certificates of competency, certifying that they were capable of completing the required weld. This should have been done under the supervision of a welding supervisor who was accepted by CWB as possessing sufficient knowledge and experience as stipulated by CSA W47.1. [48] The flare beveled welds and puddle welds used in Robb joists were not pre-qualified. If a fabricator intended to produce a weld not pre-qualified by the Canadian Welding Bureau, the following procedure should be followed. The fabricator first should satisfy itself it could consistently develop a weld to comply with the Standard and design. Then the fabricator would prepare a test of the weld witnessed by a representative of the Canadian Welding Bureau. The Bureau, in turn, would test the weld to determine if the procedure was valid and whether the welder could receive a “ticket” to make that weld for that particular company. Subsequently, other welders would be tested by the Bureau to determine if they could receive “tickets” to make the weld. [49] Welding data sheets are prepared, which are, as Mr. Follett described them, instruction sheets as to how to make a weld. They contain instructions to welders as to how to make a particular weld. Mr. Follett testified the accepted procedure is to be consistently followed to ensure there were no cracks in roots of welds as, in the field, there is no non-destructive way of determining if there is a crack in the root of a weld. [50] Mr. Follett said joists with cracked and missing welds would not get out of the Robb plant if proper procedures were followed. He said it was shoddy practice. He did not consider the welding adequate, nor the Mayflower Mall safe. If a weld was missing or cracked, it had no capacity. Mr. Follett said anyone can make a mistake, but to have 19 joists containing a joint not welded is astonishing. It calls into question the fabricator’s attention to quality. Anyone should see if a weld is missing. 10.5% of joists with a missing or broken weld, as at the Mayflower Mall, is a monstrous number. [51] In the Mayflower Mall 1.6% of joists had three or more broken or missing welds. Many of the failures were not in the most stressed areas, which indicated to Mr. Follett cracks occurred as a result of poor welding. Mr. Follett testified with Robb joists in general he noticed failures occurred in groups, with no apparent reason for the failure. The failures were random. [52] Mr. Follett testified if he knew the welds had been made by welders following approved procedures, properly qualified, with all checks in place, he could have confidence in the welding. The problem with Robb joists was he had information welders had failed welding tests, there were no approved procedures, a non-standard weld was being produced without acceptance by the Canadian Welding Bureau. Measuring the length of welds could be a valid approach to determine the strength or capacity of welds if you knew the fabricator followed approved procedures in making the welds. However, given Robb Engineering’s practices, he could not understand how measuring the outside of a weld could give one confidence. Many of the welds he saw could have been measured, but were cracked. [53] Mr. Follett testified there were problems in inspecting joists. The roof deck of the Mayflower Mall rested on the top chord of the joists, which made much of the top chord inaccessible and impossible to inspect. He also testified it was near impossible to measure welds on the top chord. [54] The remediation program provided for placing spacers on the top chord of joists. In addition to connecting web members to chord angles, welds held chord angles together. The spacers were added as the engineers considered the welds inadequate to prevent lateral movement. [55] Commenting on the use of a prybar to inspect welds, Mr. Follett said such little force is used it should not break a weld. He said he would like to see the man who could take an 18 inch prybar, stand on top of a ladder and apply enough force to break a weld. He had never seen such a man. [56] Emad Pasha, Director of Construction for Burnac Corporation, was involved with the Robb joist issue at the Mayflower Mall. He reviewed the Mayflower Mall joist inspection report. The company was concerned about safety issues arising from the report’s findings of joint failures four times higher than the Nova Scotia average. Safety was not to be compromised. He reviewed the remediation program. He was of the opinion the remediation recommended by the report should be implemented. Burnac should do what the Province was doing to their own buildings. Burnac authorized the recommended remediation. [57] Tender documents were prepared and the job was tendered through requested bids from companies with experience with Robb joists. Bids were received, but were considered high. The consultants made suggestions to reduce costs. Lower bids were received, but still considered high. The bidders were told the numbers were high and lower bids were received. The job was awarded to Metal World Incorporated. [58] CBCL and fga Consulting were involved in the remediation project. [59] George Byron Crocker , a Level II welding inspector, certified by the Canadian Welding Bureau was employed by fga Consulting in association with CBCL on the remediation project. He commenced work on the project in January, 2000. Mr. Crocker carried out inspections on joists which had been previously inaccessible, as well as joists after remediation was performed. When conducting an initial inspection, he visually inspected the chord when he could see it. He could not see all the welds on the bottom chord. Sometimes the welds were covered by slag or paint and he could not see whether the weld was cracked. Mr. Crocker applied pressure with a prybar to determine if the weld was cracked. He used the prybar on both the top and bottom chords of the joists. It was harder on the top chord, as sometimes you could not apply enough pressure. He did not want to apply too much pressure on the top chord as he might break a spot weld which attached the roof deck to the top chord, or bend the chord. Mr. Crocker testified he had never broken a good weld with a prybar. He described the welding at the Mayflower Mall as of poor quality throughout. There were missing welds. He saw burn through, undercut, overlap, porosity, crater cracks, unfilled craters, undersized welds and blobs of metal. [60] Dr. Ralph Edward Southward, a professional engineer, President of Southward Consultants Limited, was qualified as an expert entitled to give opinion evidence in structural engineering and in welding engineering as that relates to welded steel construction, welding design, welding procedure and welding practices, and the quality and capacity of welds in construction, including open web steel joists. Dr. Southward was a member of CSA W47.1 “Certification of Companies for Fusion Welding of Steel Structures” from 1979 until 1996, and a member of CSA W59 “Welded Steel Construction (Metal Arc Welding)” since 1979. He has chaired CSA W59 since 1994. He has a doctorate in structural engineering from Cambridge University. [61] Dr. Southward was retained to review the report “Mayflower Mall Joist Inspection” prepared by CBCL and fga Consulting, and comment on the reasonableness of the proposed remediation. He was also asked to comment on the original welding of Robb open web steel joists. [62] Dr. Southward testified there is a unique welding system in Canada. Most countries are working toward our system. The National Building Code places obligations on companies. Quality in production is needed to ensure the public is safe. The public assumes buildings are safe. CSA W47 was written in 1947. It establishes the protocol in Canada for quality control in steel fabrication shops. The Standard requires engineers be involved in and take responsibility for the design and work of steel fabricators. To ensure proper procedures and designs are in place, CSA created the Canadian Welding Bureau to manage the Standard. Professional engineers take responsibility. The fabricator has to have a full-time welding supervisor. In addition, welders have to pass Standard tests and be tested every two years. Welders are tested for the process, position of welding and type of welding. Welders have to be certified by the Bureau on each type of welding they do. Robb Engineering did shielded arc welding. There are standard welding configurations in the industry. Welders’ “tickets” for flare beveled welds are normally transferable from company to company if the welder changes jobs, but the Canadian Welding Bureau provides “tickets” for flare beveled welds in open web steel joists are not transferable. Canadian buildings generally have a very high safety level. The fabricator does the engineering, supervision and ensures competent welders, all supervised by the Canadian Welding Bureau. [63] There are welds the Bureau has pre-qualified. There is an obligation on the fabricator to do the work properly. If the Standard is followed, one can assume welders could do a proper job. A fabricator is not restricted to using pre-qualified welds. If the fabricator wishes to use a different joint or weld, it would have to prove it. The fabricator prepares a procedure and a test joint and submits it to the Canadian Welding Bureau. The Bureau reviews the procedure, tests the weld and determines if the weld is acceptable, and if so, qualifies the fabricator to make the weld. [64] About a year after the roof of the Sears store in the Village Mall collapsed, Sears asked Dr. Southward to review information concerning the collapse. After reviewing the joists involved, he had real concerns about the welding at Robb Engineering. Then, after the Centrac collapse, Dr. Southward was asked to give his opinion of the welding of the open web steel joists used in the roof. What he saw outside the collapsed area at Centrex was consistent with what he observed at the Village Mall - examples of what you do not do in welding. He could not rely on the quality of Robb joists when there are joists with missing or cracked welds. It is unheard of in Canada to have all these joists with missing or cracked welds. [65] Dr. Southward visited the Mayflower Mall in July, 1999. At the Mall, he observed missing welds and lack of fusion. For example, he saw a puddled weld welded across two web rods with minimal or no welding to connect the angle on the left side, and there was no fusion between the web members and the right chord. He also observed poor fit-up of web members to the chord angles. Poor fit-up is a concern. In the fabrication shop, the web members are pre-bent and brought to the welder. The web rods should be of a fixed length. Therefore, if there is poor fit-up in one chord, there will be a problem with poor fit-up in the other chord. With poor fit-up on the bottom chord, Dr. Southward has a concern for the welding on the top chord. [66] The only practical way to inspect welds in open web steel joists is visually. Radiographic and ultrasonic methods of testing are not appropriate for welds in open web steel joists. In commenting on the possibility of determining the effective throat or capacity of a weld by measuring the weld from the outside, Dr. Southward referred to figure 21 in Mr. Pates’ July, 1979 report “Joist Assemblies - Robb Engineering. An Investigation into Present and Alternate Welding Processes”, in which there appears to be a large gas pocket and crack at the root of the weld. There is no way to see that crack in the field. An outside measurement would not tell you the strength of the weld because of the defect in the root of the weld. It is difficult, virtually impossible, to examine welds non-destructively. [67] Both CSA S16 and CSA W59 require slag be removed from welds at the time the welds are made. Slag prevents the inspection of welds. [68] Dr. Southward testified cracks are not acceptable. The Standards do not allow for cracks to exist. A crack means a joint cannot carry a load. One cannot rely on a weld with a crack in it. Cracks have to be corrected. A missing or cracked weld effects the strength of welds. [69] Dr. Southward testified the Village Mall, Centrac and the Mayflower Mall had the same systematic problem. It is difficult to have any confidence in welds if a quality control process allows joists shipped with missing or cracked welds - it puts the rest of the welds in question. He could not answer the question as to what was the chance of failure if the welds in the Mayflower Mall were not reinforced. No one knew the extent of the lack of fusion at the Mayflower Mall. He could not rely on the quality of Robb joists because of the missing and cracked welds. Dr. Southward did not attempt to determine the capacity of welds at the Mall. [70] Considering the results of the Newfoundland “Open Web Steel Joists Load Testing and Inspection Program - Final Report”, Dr. Southward concluded in his report: Since a “broken” weld should not exist in a properly designed and fabricated open web steel joist, these findings represented a very real and substantial risk of collapse, and a concern for the safety of the public. [71] In response to a question as to what effect a prybar may have on a sound weld, Dr. Southward stated he did not think an inspector with an 18 inch prybar would generate enough force through his body weight and strength to fracture welds if they were sound welds of adequate size to do the job they were expected to do. [72] Dr. Southward accepted the remediation approach adopted for the Mayflower Mall. He said it was not practical to do rewelding on the top chords as it was hard to get at the chords. The remediation program was necessary. The Mall contained a higher number of missing or cracked welds than the Nova Scotia average - 10.5% versus 2.28%. Welds were covered with slag. Considering Robb Engineering’s welding procedures, there was no way to know if the welds had strength. There was no method available to test the strength of the welds that remained at the Mall. Just repairing the missing or cracked welds at the Mall was not sufficient to determine the joists were safe. To Dr. Southward, the remediation was a matter of public safety, the joists had to be remediated to attempt to establish the reliability expected of Canadian buildings. It was necessary to ensure the remaining welds were all right. [73] With Level I remediation for joists thirty feet or more in length, 25% of the joints at each end of the joists were to be reinforced and the missing and cracked welds repaired. Engineers made the decision, not every joint had to be reinforced. At the Mayflower Mall slightly less than 50% of joists were reinforced. Dr. Southward testified it would have been easier to replace the joists, but the decision was made the remediation would protect the public. [74] Spacers were also placed in the top chord in the Level I remediation. It was not practical to reweld the cracked welds on the top chord with the roof deck sitting on the top chord. It was not acceptable to weld over a cracked weld. W59 is clear the crack is to be removed. The chord members of Robb joists were designed so as not to require spacers. Spacers were included in the remediation program to provide for missing welds or welds that may crack. Spacers were used to prevent chord members from buckling. [75] Dr. Southward said the testing at Memorial University reported in “Open Web Steel Joists Load Testing and Inspection Program - Final Report” only provided information about the welds tested. He did not place a lot of weight on the test. Nothing in the report indicated remediation was unnecessary. The purpose of the testing was to give confidence in the remediation scheme. [76] In the 2003 edition of CSA W59, flare beveled welds are now pre-qualified for applications other than open web steel joists. There was not any discussion to extend flare beveled welds’ pre-qualification to open web steel joists. In his report, Dr. Southward sets out his conclusions as follows: 1. The rate of failure of Robb Engineering open web steel joists indicated something was fundamentally wrong. . . . . 2. Since a “broken” weld should not exist in a properly designed and fabricated open web steel joist, these findings represented a very real and substantial risk of collapse, and a concern for the safety of the public. . . . . 3. The problem with Robb Engineering open web steel joist was the welding practices and lack of quality control associated with the fabrication of their joists. Robb Engineering joists did not meet the requirements of CSA Standards S16, W47 and W59. . . . . 4. The Mayflower Mall, located at 800 Grand Lake Road in Sydney, Nova Scotia was identified by Robb Engineering as being constructed using their open web steel joists. . . . . 5. An inspection of the welding of the open web steel joists in the Mayflower Mall indicated the welding was of greater concern than the welding typically found on other Robb Engineering joists in Nova Scotia. 6. For public safety, the issue of the welding of the Robb Engineering joists at the Mayflower Mall had to be addressed. . . . . 7. In our opinion, the remediation program devised for the Mayflower Mall was necessary, and justified, because of the welding practices of Robb Engineering. 8. The remediation program was based on a method of repair that was suitable for joists that had already been installed in buildings. The program recognized the joists had existed for some time, and categorised the level of remediation required relative to the risks associated with the potential failure of each joist. [emphasis in original] [77] Dr. David James Laurie Kennedy, a professional engineer and professor of engineering since 1956, was qualified as an expert entitled to give opinion evidence in structural engineering and in the research and testing of structural components and welded connections, as well as the design and investigation of steel structures including welded connections in open web steel joists. Dr. Kennedy has a doctorate from the University of Illinois in structural engineering. He was a member of CSA S16 “Steel Structures for Buildings” from 1962 to 2001, and Chairman of S16 from 1968 until 2001. Dr. Kennedy developed a strong interest in steel structures in 1960 and did fundamental work on open web steel joists in 1965. Besides his teaching duties, Dr. Kennedy has been engaged in consulting engineering. [78] Dr. Kennedy stated he was not a welding engineer and he did not assess welding practices or procedures in his consulting practice. He agreed Dr. Southward was the expert for W47 and W59. [79] Dr. Kennedy had been involved with Robb’s joist issue since the late 1980's. His involvement included testifying in the trial arising from the Centrac collapse. He was also involved with the Village Mall, the Mall in Summerside and the North Sydney Mall. He has not tested a Robb joist, nor Robb flare beveled welds or puddle welds. [80] Dr. Kennedy visited the Mayflower Mall on July 6, 1999 with Greg Saunders of SGE Acres and Joseph Schneider, a former employee of Dominion Bridge. Prior to the visit, Dr. Kennedy had a copy of the inspection report prepared by CBCL and fga Consulting. The purpose of the visit was to determine if they could corroborate the results of the CBCL/fga Consulting inspection and determine how they could participate in the remediation program, if needed. Messers. Schneider and Saunders visited the Mall and inspected joists prior to Dr. Kennedy’s arrival. Dr. Kennedy read notes prepared by Mr. Schneider which corroborated the CBCL/fga Consulting inspection report. Dr. Kennedy wanted to measure the size of the flare beveled weld and puddle welds, and look at the Mall’s roof. He was on the roof about one hour, but his observations there did not play an important part in his opinion. Using a small measuring tape, approximately four millimetres wide, and calibrated calipers, Dr. Kennedy looked at seven joists. He did not inspect all panel points or joints on any joist. None of his measurements were made on the top chord of any joist. None of his measurements were made at the end of a joist. Dr. Kennedy’s opinion was the capacity of welds could be determined by measuring the face length and width of welds. [81] Dr. Kennedy testified missing welds, broken welds or damaged end joints are gross defects and constitute a danger in a building. They must be repaired, he said. In an article he wrote, “Evaluation of structural weld defects”, Dr. Kennedy stated: Cracks. Perhaps of all the defects, cracks are the most serious and under no circumstances should such defects be tolerated. . . . . ... Careful control must be provided over welding procedures to prevent cracking in, for example, the case of making welds between heavy parts that are severely restrained. ... Under almost any condition of loading the strength of a structural member or welded joint is likely to be reduced substantially by the presence of even a very small crack. ... Cracks can be prevented by conforming to proper welding procedures. ... [82] Dr. Kennedy did not know if the welds at the Mayflower Mall were filled flush. Welds transfer shear forces and hold chord angles together in the top chord. The joints he measured at the Mayflower Mall was a very small sample. The question was, what load can a Robb joist carry, and Dr. Kennedy said he had a lot of data from the Newfoundland report, “Open Web Steel Joists Load Testing and Inspection Program - Final Report”. [83] The top chords of joists were more difficult to examine than bottom chords as the roof deck was welded to the top chord. Dr. Kennedy testified it was a problem to determine if a weld was of sufficient size in the field. On cross-examination, Dr. Kennedy reluctantly agreed you cannot physically measure length and width of welds on top chords. He gave the following evidence: Q. Now if someone wants to measure ... sorry, someone wants to calculate the capacity of a weld, you need to measure how much weld is present? A. That is correct. Q. And you need some device to measure, obviously? A. That’s correct. Q. And you used a measuring tape to measure lengths ... A. That’s correct. Q. ... in the Mayflower Mall? And you say you used calipers to measure the face width? A. That’s correct. Q. And when measuring face widths, you have to measure a number of locations along the length of the weld if the width is not constant? A. That is correct. Q. And then you ... you’d add them all up and average them? But I ... my friend walked into that last week. That’s basically what you do? A. That’s correct. You measure them at a number of locations when the ... it appears that there’s a visible variation in the width. Q. Now for the top chord, Dr. Kennedy, you’d need a flashlight to see up there? A. Yes. Q. And you need your tape measure to measure? A. But I did not measure any top chord welds. Q. Fair enough. I understand that. A. Okay. Q. You’d also need a mirror? A. That would be a help, yes. Q. Because you can’t see the top of the welds from the top chord because of the presence of the roof deck? A. That’s correct. Q. You can’t get your head up there? A. Yeah. Q. So in order to measure let’s say a flared bevel weld that’s located in an up flute ... A. Yes. Q. ... of a corrugated deck, you’d be at the top of your ladder just about? A. That’s correct. Q. And you have flashlight in one hand? A. I’m going to have three hands here pretty soon. Q. You ... you see where I’m going, Dr. Kennedy. A. Yes. Q. So you’ve got three hands and ... but most of us are short one. We got a problem there, don’t we? Correct? A. I didn’t ... and I didn’t measure any top chord welds. Q. No, I ... I ... I acknowledge that, you didn’t measure any top chords. A. Yeah. Q. So you’ve got your flashlight in one hand, your mirror in the other hand, and your calipers ... we’re missing a hand to do that or ... or the tape measure, whichever, correct? A. Yes. Q. Now, if the ... of course, if the down flute is on top of a weld, you can’t get up there to measure anything anyway, can you? A. If a ... the down flutes are ... are not as ... harmful is not the right word ... as they’ve been made out to be, in my opinion. Let’s take the flare beveled groove weld that we’re talking about. It goes around the corner. The corner is quite a gradual corner. It starts here, goes up, around and down. I can still look at the ends. I agree with you that I cannot get a proper estimate of the width of the weld, but I can see both ends. Q. So you’d have to estimate the length of the weld? A. You can ... yes, you’d ... you have to estimate the length. You can make a mark on the rod at right angles to its length at this end and at that end, and then get that length from ... Q. Okay. A. ... from the inside of the rod. Making allowance for the thickness of the rod, the dam of the rod and you have the length. It is more difficult. Q. Okay, so. Yeah.. Would you agree that it’s ... you cannot practically measure face widths and lengths at the top chord? A. It’s more difficult ... difficult ... Q. It’s ... A. ... and I ... I will ... Q. It’s very difficult. For example, how do you get your caliper up there to measure the face width? A. You ... you can’t get the caliper, only on the end. And we’ve discussed that issue already. Q. Yeah. So if there are variations in the width, you’re going to be in some trouble with the measurements of the face width. A. So what you do is you say, all of these flare beveled groove welds were made in the same manner. They were made with the welder standing over the joist and looking down and going wooooozzzzz zzzzzzzzz, and then the joist is flipped over and those welds that he just did are now the bottom chord welds, and now we do the top chord welds zzzzzzzzzz zzzzz. So knowing the sequence of welding and measuring the bottom chord welds, it is not very difficult to draw the conclusion that top chord welds are like bottom chord welds, and if I can make some measurements on the bottom chord welds is you ... I have a grasp of the top chord welds. Q. Okay. Coming back to my question though, it’s not practical to actually measure the top chord welds. You’ve just described as ... a ... a way of estimating the top chord welds, but you cannot practically measure them, can you? A. Well, I’ll ... I’ll say it again. I can look at the ends of the welds, and then I’m using my knowledge of the variation in the width of such flare beveled groove welds to say, is the measurement at this end and that end a likely indicator of the average width of the weld face. Q. My question is very simple, Dr. Kennedy. A. And I thought my answer was. Q. You can’t physically measure them at the top chord - the whole weld? A. That is correct. [emphasis added] [84] Dr. Kennedy said if an engineer is looking at any problem, he or she should obtain all information available, for example, if one is looking at Robb joists, the National Building Code, CSA S16, snow data and the Newfoundland testing report. He did not review Mr. Follett’s report, Dr. Southward’s report or the plaintiff’s List of Documents prior to preparing his report. [85] Dr. Kennedy placed a great deal of emphasis on the testing done at Memorial University. He said the Newfoundland report shows the reinforcing done at the Mayflower Mall was unnecessary. In the report at p. 31, s. 3.5.3, the following statement is made: The open web steel joists were of two distinct levels of quality. The joists that were obtained from Holy Cross School contained better welds than those obtained from the Mall. The joists from the Mall, however, had many repairs to connections to make them suitable for testing. [86] Dr. Kennedy, who was not present for the testing in Newfoundland, was not prepared to accept that the joists tested were of two distinct levels of quality. He agreed the joists were lying on a test bed when tested, with access to all panel points and chords for inspection and repair. He also agreed the rates of broken and missing welds varied from building to building in Newfoundland. Most of the joists tested in the Newfoundland report were repaired prior to testing. Dr. Kennedy agreed a fair amount of repairs were made in accordance with W59 and the welds, when tested, were no longer Robb welds. [87] In his report, Dr. Kennedy did two sets of calculations concerning the strength of welds in the Mayflower Mall. Table III was based on factors only added to CSA S16 in 2004, and the equations were not available at the time of the remediation of the Mayflower Mall. The second calculations in Table IV used older factored resistance to calculate effective throat. On cross-examination, Dr. Kennedy agreed the calculations in Table IV were incorrect. [88] Dr. Kennedy testified crater cracks do not propagate in statically loaded structures such as the Mayflower Mall. He stated open web steel joists are designed with redundancies built in, that is, different ways to carry load. In dealing with redundancies in his report, Dr. Kennedy stated at p. 3: These three redundant paths for carrying load exist but are not relied upon in the basic design and fabrication. Neither are they used to avoid a necessary repair. But, with the redundancies, an isolated or unknown defect in all probability would not lead to collapse. ...” [emphasis added] [89] Having reviewed the Newfoundland report, Dr. Kennedy testified the puddle welds were fantastic and the flare beveled welds doing their job. In dealing with the Robb joist problem, he said he would ask himself what he has to do to repair the joists to prevent failure. He would repair defects. He would either measure the welds or, on the top chord, use a prybar to determine the strength of welds. His opinion is the scheme devised should be the minimum which would allow the joists to do their job, that is, repair major defects. [90] In his report, Dr. Kennedy came to the conclusion 43% of the broken welds at the Mayflower Mall were broken during the course of the inspection by CBCL and fga Consulting. [91] Dr. Kennedy’s opinion was the Level I remediation which required reinforcing of 25% of the joists at each end was over and beyond what was required for the structural sufficiency of the joist. The Newfoundland report showed this reinforcing was unnecessary. In the tests in Newfoundland, all welds carried the loads they had to carry. The reinforcing of welds at the ends of joists would not affect welds at mid-span. The puddle welds demonstrated that they carried load in badly damaged situations. Dr. Kennedy was of the opinion the spacers provided for Level I remediation were unnecessary. The spacers have nothing to do with transferring shear from chord to web members, but are to prevent lateral buckling. The Robb joists were engineered not to require spacers. If there was a broken weld, it should be identified during the inspection. [92] Dr. Kennedy’s conclusions concerning the Mayflower Mall are set out in his report as follows: Summary and Conclusions 1. CBCL and FGA carried out close to a 100% inspection of the joists of Mayflower Mall in 1998 which revealed a miscellany of defects. 2. At the time of the inspection the roof structure had withstood “the test of time” for about 20 years. 3. A limited inspection of the joists by Saunders, Schneider and Kennedy as reported herein indicated that about 43% of the welds reported as broken by CBCL and FGA were broken by excessive prying forces in the attempt to determine which welds were cracked, broken or missing. Damage - seven inelastic deformations - in the chord angles provides evidence that the prying forces were excessive. 4. Field measurements of the weld face width and length reported herein indicate that the flare bevel groove welds have calculated shear strengths greater than required to transfer the maximum force in web rods to the chord angles. There is no evidence in the CBCL and FGA report that any attempt was made to determine the throat areas of welds and thus to assess or estimate their strengths. 5. The “Newfoundland” report contains two significant series of tests that demonstrate (a) joists with the minimum level of repairs as given by CBCL and FGA carried loads greater than the factored load level and are therefore structurally sufficient and (b) joists with a puddle weld severely damaged including the most critical puddle weld at panel point 3 were also structurally sufficient. 6. Since CSA Standard S16.1-1974 came into effect, the capacity of the top chord of Robb joists in their RF series has been based on skew axis buckling of the individual angles and therefore no spacers are needed. They are irrelevant. 7. CBCL and FGA recommend three levels of remedial work. The least extensive, Level III consists of 100% inspection together with repairs to failed joints and bent members. The intermediate Level, Level II, in addition, requires top chord spacers to be installed in accordance with CSA S16.1-94. Much more work is required for highest Level I with significant reinforcement of members and panel points over about one-half of the joist span. We consider this remediation scheme based on that proposed in the Newfoundland Report to be inconsistent with the favourable joist strength test results. 8. Based on the two significant series of joist strength tests in the “Newfoundland” Report and corroborated by weld face and length measurement, we opine that the necessary and sufficient remedial work to bring all joists to the desired state of structural sufficiency comprises, following 100% inspection, repair of failed welds and straightening of bent compression members. [93] Stig Rolf Skarborn, a professional engineer, President of Skarborn Engineering Limited, was qualified as an expert entitled to give opinion evidence in structural engineering and in welding engineering as that relates to welded steel construction, welding design, welding procedure and welding practices, and the quality and capacity of welds in construction, including open web steel joists. [94] Mr. Skarborn is also a Level III welding inspector and is the retained welding engineer for a number of companies. His firm was retained by the Department of Supply and Services of New Brunswick to inspect buildings with Robb joists, to provide an opinion as to whether the joists were capable of carrying the load of the building. Mr. Skarborn inspected thirteen buildings for the Province of New Brunswick, as well as two shopping centres, the Lancaster Mall in Saint John, New Brunswick and the Cumberland Mall in Amherst, Nova Scotia. In addition, Skarborn Engineering Limited conducted inspections in Nova Scotia as sub-agent for CBCL Limited. [95] The procedure Mr. Skarborn used was to inspect the first three or four panel points or nodes on one-third of the joists. If there was a problem, all joists were inspected. In areas with snow load, the first three or four panel points on all joists were inspected. At a school he inspected in Saint Louis de Kent, New Brunswick, there was a large number of cracked and missing welds, as well as other problems with the joists. [96] When doing his inspections in the late 1990's, Mr. Skarborn was proceeding on the basis that welding procedure data sheets for Robb Engineering existed. It came as a surprise to him that Robb Engineering did not have flare beveled welds approved by the Canadian Welding Bureau. Mr. Skarborn was unaware of Robb’s internal correspondence about the problems with welds. Although he was not at first successful in obtaining documentation concerning Robb’s welding procedure, he eventually obtained welding procedure documentation date stamped 1992. [97] Mr. Skarborn testified a crack is a defect which is never acceptable and must be repaired, and the problems with Robb joists had to be repaired. [98] Mr. Skarborn wanted to determine the strength of the flare beveled welds and puddle welds used in Robb open web steel joists. He had six Robb flare beveled welds and five Robb puddle welds. That was not a sufficiently large sample to make an engineering assessment. Additional welds were simulated for testing. The simulated welds were of superior quality to the Robb welds. Mr. Skarborn tested the welds and concluded if the ratio between the face width and effective throat thickness of a weld is established, the resistance of a flare beveled weld can be determined by inspection of the weld’s face length and width. Mr. Skarborn’s position was set out in a paper by himself and Gary Daneff entitled “Shear Resistance of Flare Beveled and Puddle Welds in Open Web Steel Joist Applications” presented to the Annual Conference of the Canadian Society for Civil Engineering in June, 1998. Mr. Skarborn testified his paper assumed competently made welds. [99] The August, 2003 CSA W59-03 Standard adopted a face width and effective throat ratio identical to that used by Mr. Skarborn. The ratio did not apply to open web steel joists. [100] Mr. Skarborn did not know the required lengths of welds when he conducted his inspections. [101] In commenting on Mr. Skarborn’s report, Dr. Kennedy said he was comfortable with Mr. Skarborn’s data. The welds simulated for Mr. Skarborn’s testing were the same as the welds made by Robb Engineering. Dr. Kennedy had confidence in the welds. [102] Mr. Skarborn testified it was impossible to measure the length of welds on the top chord where the roof deck obstructed it. In his report, Mr. Skarborn stated, “Welds on the top chord are almost completely obscured by the deck.” He testified it was impossible to measure lengths of welds which were so obscured. He had difficulty inspecting welds on the top chord - it could not be done where the roof deck rested on the top chord, where the roof deck was raised an inspection mirror could be used. He stated it was not an ideal methodology, but was the one they used. There were cracked welds on the top chords when Mr. Skarborn conducted the inspection of the Ellerslie Elementary School in Prince Edward Island. The welds were difficult to repair, but a pencil grinder could be used to remove the welds prior to rewelding. [103] Welds connect the two top chord angles to prevent buckling. If a weld is not present, the length for assessing buckling is longer. [104] All welds should have slag removed to be inspected under normal conditions. Mr. Skarborn normally insists slag be removed before he inspects a weld, but he testified, he is a practical man and Robb joists were a special case. The presence of slag helped in the inspection. Slag is very brittle and if there was movement, the first thing to happen was the slag would crack. [105] Mr. Skarborn had met Gary Follett at a number of conferences. He contacted Mr. Follett when he started working on the Robb joist issue. He attended a meeting in Fredericton, New Brunswick, he thinks in late 1997 or 1998, which was attended by Mr. Follett and a representative of the Nova Scotia Government, where the New Brunswick approach up to that time was explained. [106] When acting as CBCL’s sub-agent in conducting inspections in Nova Scotia, Mr. Skarborn used the Nova Scotia methodology. Mr. Skarborn agreed it is possible for reasonable engineers to disagree and approach the same problem differently. [107] Michel Comeau, a professional engineer, testified and was qualified as an expert entitled to give opinion evidence in structural engineering and the design of buildings, and the investigation and remediation of buildings containing open web steel joists. Mr. Comeau is a Principal and Director of Engineering of Campbell Comeau Engineering Limited. He has been a member of CSA S16 Technical Committee Steel Structures for approximately eighteen years. [108] Mr. Comeau has dealt with Robb joists since the early 1980's. His firm has been involved in thirty buildings containing Robb joists. Starting in 1998, Mr. Comeau looked after half of the buildings himself. The firm inspected buildings for the Royal Bank of Canada. The welding in the Royal Bank buildings was not particularly bad. In buildings inspected by his firm, only a small number of defects were found. He inspected the South Centre Mall where the joists had a higher percentage of defects. Mr. Comeau stated defective welds are a function of workmanship of the joist fabricator. With poor workmanship, more defects of various types can be expected. One could expect defects inside welds of poor workmanship. [109] Mr. Comeau assumed Robb Engineering was a qualified fabricator, with qualified welds and welders. He assumed compliance with all Standards: W59, W47 and S16. The use of Standard W59 and S16 is based on the assumption all welding was done in accordance with W59; if not, the use of S16 for further calculations may not be accurate. [110] In preparing his report dated June 9, 2006, Mr. Comeau did not review Mr. Follett’s or Dr. Southward’s reports. He agreed an engineer should take into account available information when preparing a report. [111] Mr. Comeau visited the Mayflower Mall for one-half day on August 10, 1999. He looked at the interior of the Wal-Mart store, the roof area of the mezzanine of the Wal-Mart store and was also on the Mall’s roof. He does not recall if a ladder was used during the visit. His inspection was visual, to look at the defects set out in the CBCL/fga Consulting report and the repairs carried out. The repairs he observed were reinforcing plates in the area of broken welds. Vertical plates with welds added to reinforce joists. Joists with three or more defective welds were repaired. Broken welds in joists were observed. Mr. Comeau looked at a dozen joists. He did not inspect all panel points on any one joist. He did not measure welds or calculate the capacity of welds at the Mayflower Mall. He did not calculate the forces on the Mall or look at design drawings for the Mall. [112] In discussing the remediation scheme for the Mayflower Mall, Mr. Comeau testified broken or missing welds and bent web members must be repaired. A joist with missing or cracked welds should never leave a fabricating shop. He stated, the reinforcing of 25% of each end of joists 30 feet or more in length is not necessary if all other defects in the joist have been repaired. [113] Mr. Comeau stated if there is a crack at the root of a weld, it should be deducted in determining the strength of the weld; but as it cannot be seen from the surface, the strength of the weld would be over-estimated. Mr. Comeau testified the reinforcing of the joists would increase the strength of the joist as it would reinforce 25% of the joist at each end. The spacers added to the top chord stiffen the top chord and increase the strength of the joist if the weakest link is on the top chord. Welds from panel point to panel point determine the buckling length of a joist. If a weld is broken, the buckling length is the distance between sound welds. Spacers limit the buckling length of the joist if there is a missing or broken weld. [114] Mr. Comeau stated it is difficult to see welds on the top chord and is not practical to measure welds on the top chord. If the roof decking is down, resting on top of the joist, the welds cannot be measured - you can only look up at the welds. When conducting inspections, the only inspection Mr. Comeau made of puddle welds was to look to see if a weld was present and feel it with a finger. [115] Mr. Comeau did not know when he became aware of the paper “Shear Resistance of Flare Beveled and Puddled Welds in Open Web Steel Joist Applications” by Skarborn and Daneff. He did not think he was aware of it at the time of his visit to the Mayflower Mall. [116] John Richardson, a professional engineer and a partner in DMR Engineering testified. He was involved with inspecting ten to twenty buildings to determine if they contained Robb joists. If the building contained Robb joists, it would be inspected. Mr. Richardson would look at 5% of joists by doing a visual inspection. He said the top chord of joists was difficult to see because of the roof deck resting on it. He spent most of his time looking at the bottom chord. He used a prybar mainly in inspecting the top chord. If deficiencies in welds were found, he recommended a full inspection. Then he looked at every bottom chord weld and as many top chord welds as he could. [117] Referring to work he did at the North Sydney Mall, he stated welds were not measured as the welds varied along their length and the depth of the weld could not be measured in any way to give meaningful information. He reviewed the work of seven or eight consultants in New Brunswick dealing with Robb joists and agreed their approaches varied widely, as each consultant had his or her own approach. [118] The following facts have been established by the evidence. The Mayflower Mall was built in 1979. It was constructed with open web steel joists, fabricated by the defendant through its subsidiary, Robb Engineering, of Amherst, Nova Scotia. The open web steel joists were transported and erected by the defendant. The Mall’s corrugated metal roof sat on the top chord of the joists. [119] There were problems with the welding of open web steel joists at Robb Engineering. There was a lack of quality control. Welding of the joists was not in accordance with Standards CSA W59 and W47.1. As welds were covered with slag, they could not be inspected to determine their adequacy. There was no standard way to pile joists for inspection. Joists were piled so inspectors could see either the top or bottom chord, but not both. There was not adequate inspection of joists. No welding data sheets existed concerning open web steel joist welding. [120] Robb Engineering was aware of the problems with the welding. Samples of production welds were sectioned and examined. Weld metal cracks were evident in samples examined. The welding of Robb joists was inadequate. Considering the Robb Engineering welding practices, there was no non-destructive method to determine if a weld had strength and was sound. [121] At the Mayflower Mall all accessible panel points and all accessible joists were inspected. Of the 1,900 open web steel joists that could be inspected, a total of 199, or 10.5%, had at least one panel point with a broken or missing weld, 84 joists had one or more bent web members and 40 joists had one or more bent chord members. The quality of the welding at the Mayflower Mall was poor, with porosity, overlap and undercut present, slag not removed, and cracked and missing welds. I accept the evidence of Messrs. Pinhorn, Crocker and Follett and Dr. Southward as to the quality of the welding at the Mayflower Mall. The length and width of welds on the top chord of joists could not be measured. Where the metal roof deck rested on the top chord, welds could not be seen or inspected. Cracked or missing welds are a danger and must be repaired. [122] There were defects in the open web steel joists at the Mayflower Mall. The defects constituted a real and substantial danger to the occupants of the Mall. The danger was caused by the defendant’s negligence in the fabrication, transportation and erection of the open web steel joists. [123] I prefer the evidence of Mr. Follett and Dr. Southward to the evidence of the defendant’s experts. [124] I have difficulty with Dr. Kennedy’s evidence for the following reasons. [125] Firstly, Dr. Kennedy considered the remediation at the Mayflower Mall unnecessary as the capacity of welds in the joists could be determined by measuring their length and width. Determining the capacity of welds by outside measurements is premised on properly fabricated welds. To use outside measurements to determine capacity, Dr. Kennedy testified he depends on the efficacy of the welder. Dr. Kennedy also assumed welds on the top chord were the same as welds on the bottom chord. This assumption is not supported by the evidence. It is obvious from the evidence, including the great number of cracked and missing welds, and lack of quality control, welds were not properly fabricated at Robb Engineering. Welds on the top chord could not be measured - Dr. Kennedy said the length could be estimated - that is not sufficient when dealing with public safety. [126] Secondly, Dr. Kennedy placed great importance on the results of the testing contained in the Newfoundland report. He considered only gross defects had to be repaired, as the joists performed well during testing. However, as he agreed on cross-examination, a fair amount of repairs had been made to the joists in accordance with W59 and the welds, when tested, were no longer Robb welds. [127] Thirdly, while stating it was important for an engineer to review all available information when undertaking an assignment, when preparing his report Dr. Kennedy did not review the available Robb Engineering documentation concerning the welding process or Mr. Follett’s and Dr. Southward’s expert reports. [128] Fourthly, the calculations set out in Table IV of his report were wrong. When the calculations were corrected, the resistance of three of four welds was less than the maximum potential shear load. Table IV does not tell anything about the top chord welds as they were not tested, nor does it indicate what is going on inside the weld. On redirect examination, Dr. Kennedy, after his calculations were corrected, described the welds referred to in Table IV as “marginally inadequate”. [129] Finally, in his report, Dr. Kennedy concluded 43% of the broken welds at the Mayflower Mall were broken by excessive prying forces during the inspection by CBCL and fga Consulting. I do not accept that conclusion. Mr. Crocker testified he used the prybar on both top and bottom chords, but sometimes he could not apply enough pressure on the top chord as you might break a spot weld which attached the roof deck to the top chord or bend the chord. He added, he had never broken a good weld with a prybar. Both Mr. Follett and Dr. Southward testified a person could not apply enough pressure to crack a sound weld. I accept that a person standing on a stepladder holding a flashlight and prybar could not apply enough force to crack a sound weld. [130] I am not prepared to attach any weight to Dr. Kennedy’s opinion that the remedial work carried out by the plaintiff was excessive. [131] I have difficulty with Mr. Comeau’s opinions for the following reasons. [132] Firstly, Mr. Comeau’s opinion was, broken or missing welds and bent web members must be repaired. The reinforcement of 25% of each end of joists thirty feet or more in length was unnecessary if a joist had properly fabricated welds, and missing or cracked welds repaired. His opinion is based on there being properly fabricated welds. That is where his opinion fails. Nobody knew the capacity of the welds. They were fabricated without following recognized Standards. There was lack of quality control. Mr. Comeau testified, with poor workmanship or defects of various types, including interior defects, can be expected. Mr. Comeau assumed compliance with all Standards. [133] Secondly, Mr. Comeau examined only a very small number of joists, only a dozen. He did not inspect all panel points on any one joist. [134] Thirdly, in his report, Mr. Comeau comments on the snow load data for Sydney and says the Sydney area has received snow loads up to the design values for the Mall since the Mall was constructed. On cross-examination, he stated he did not know how much snow was on the roof of the Mayflower Mall. [135] I am not prepared to attach any weight to Mr. Comeau’s opinion that the remedial work carried out by the plaintiff was excessive. [136] I have difficulty with Mr. Skarborn’s opinion for the following reasons. [137] Firstly, Mr. Skarborn did not visit the Mayflower Mall and could not and did not give evidence of the condition of the welding at the Mall. [138] Secondly, Mr. Skarborn was not prepared to agree that his method of determining capacity of welds by outside measurement required the welder to follow approved procedures. He assumed high enough amperage was used in the welding to ensure an adequate weld. He agreed an outside measurement will not show partial fusion of web and chord, porosity inside a weld or a crack inside the weld. He was unaware of the internal correspondence concerning welding problems at Robb Engineering. Considering the problems with Robb Engineering’s welding, there is not sufficient consistency in the welding to use his method. [139] Thirdly, Mr. Skarborn measured bottom chord of joists. He testified it was impossible to measure the length of welds on top chords were the roof deck obstructed them. [140] Fourthly, although critical of the inspection methodology used in Nova Scotia when acting as a sub-agent of CBCL Limited in conducting inspections in Nova Scotia, Mr. Skarborn used the Nova Scotia methodology. [141] I am not prepared to accept Mr. Skarborn’s opinion that the remediation carried out by the plaintiff was excessive. [142] Burnac’s claim is for economic loss for the cost of repairs and remediation to the Robb joists at the Mayflower Mall. The test to recover for such an economic loss was set out by La Forest, J. in giving the Court’s judgment in Winnipeg Condominium Corporation No. 36 v. Bird Construction Co. Ltd., [1995] 1 S.C.R. 85 at p. 121 as follows: I conclude that the law in Canada has now progressed to the point where it can be said that contractors (as well as subcontractors, architects and engineers) who take part in the design and construction of a building will owe a duty in tort to subsequent purchasers of the building if it can be shown that it was foreseeable that a failure to take reasonable care in constructing the building would create defects that pose a substantial danger to the health and safety of the occupants. Where negligence is established and such defects manifest themselves before any damage to persons or property occurs, they should, in my view, be liable for the reasonable cost of repairing the defects and putting the building back into a non-dangerous state. [143] In describing the burden of proof on a plaintiff, La Forest, J. stated at p. 125: ... The burden of proof will always fall on the plaintiff to demonstrate that there is a serious risk to safety, that the risk was caused by the contractor’s negligence, and that the repairs are required to alleviate the risk. [144] United Dominion Industries Corporation, through its subsidiary, Robb Engineering, fabricated, transported and erected the open web steel joists in the Mayflower Mall. [145] The evidence clearly shows there was a lack of control at Robb Engineering over how welds on open web steel joists were made. There was a protocol established through CSA Standards W47 and W59 to be followed in the fabrication of welds to ensure confidence the welds are properly fabricated. Robb Engineering did not comply with the Standards. Robb welds were covered with slag, which made them virtually impossible to inspect. There was no comprehensive inspection of joists at Robb Engineering. Inspections took place in a “hit or miss” fashion. When joists were inspected, only one chord, top or bottom, depending on how the joists were piled, was inspected. Mr. Smith testified, if he was busy, he had no time to inspect joists. He agreed there could be joists not inspected. [146] Through his involvement with the Open Web Steel Joist Review Committee for the Province of Newfoundland, Mr. Follett was aware of the problems with Robb open web steel joists, including the lack of quality control. This lack of quality control was obvious, considering the number of missing or cracked welds found during inspections. Dr. Southward was also aware of the problem with Robb open web steel joists. [147] The inspection at the Mayflower Mall showed there were many defects in the welds in the joists, many more than the average number of defects found in buildings containing Robb joists in Nova Scotia. The evidence shows, and I find, there were defects in the open web steel joists at the Mayflower Mall. [148] All experts who testified at the trial agreed missing or cracked welds had to be repaired. The evidence clearly shows that the defects created a real and substantial danger to the occupants of the Mall. Joists should not leave fabricators’ facilities with cracked or missing welds. From the evidence, the danger was caused by the defendant’s negligence. [149] The defendant agrees missing and cracked welds, as well as damaged chord and web members, have to be repaired, but says the reinforcing of the joists and the addition of spacers to the top chord was not necessary to mitigate the damage. Through inspections, as described by witnesses called by the defendant, not all welds could be inspected and their capacity determined. Once the repairs of missing or cracked welds, and damaged web or chord members were made, the joists were capable of supporting the load on them. There being no danger once the repairs were made, there was no longer any real and substantial danger needing repairs. [150] The defendant says the reinforcing of the joists was dealing with a “perceived” danger and not an actual “real and substantial” danger. In support of its position, the defendant cites M. Hasegawa & Co. v. Pepsi Bottling Group (Canada) Co., [2002] 213 D.L.R. (4th) 663 (B.C.C.A.), where Finch, C.J.B.C. stated in giving the Court’s judgment at p. 676: With respect, a test of perceived as opposed to actual danger is, in my view vague, overbroad, and impractical. The plaintiff does not suggest whose perception should govern, or how a trier of fact could, on any reasoned basis, choose between evidence of differing perceptions of risk. I would accept as sound the premise that the policy of the law should encourage the production and distribution of food products that are wholesome, and not a danger to health. But whether a food product is an actual danger is a matter upon which scientific opinion, however uncertain it may sometimes be, can be offered, tested and weighed. A test of “perceived” danger is, however, no test at all. Perception is not a matter susceptible of proof, or disproof, by evidence. [151] The plaintiff takes the position remediation was necessary. No confidence can be had in the capacity of Robb joists. Proper welding procedures were not followed. There was no quality control at Robb Engineering. Inspections at the Robb facility were haphazard and incomplete. There were internal problems with welds from Robb Engineering which could not be discovered by non-destructive inspection. The lack of proper welding procedures and quality control issues prevented one from having confidence in Robb welds, at least on external observations. Welds on the top chord could not be inspected to determine if a problem existed. There being a real and substantial danger, the extent of which could not be determined, the remediation, including reinforcement and addition of spacers, was necessary to alleviate the danger. [152] I find the remediation program carried out at the Mayflower Mall was necessary to alleviate the real and substantial danger existing at the Mall. The fabricator, Robb Engineering, did not employ proper welding procedures and practices in welding the joists. There was no quality control. Given what was known about the practices at Robb Engineering, and observations made in the field, a reasonable person could not have confidence in the capacity of welds used by Robb Engineering. Welds on the top chord of joists could not be inspected. The evidence was the defects in the welding occurred randomly and, what was found on the bottom chord, could not be taken as what existed on the top chord. [153] The placing of spacers on the top chord was part of the alleviation of the danger. The state of the welds on the top chords was unknown. If there was a cracked or missing weld on the top chord, the lateral buckling length of the chord would be lengthened and, therefore, the possibility of lateral buckling increased. The placing of spacers reduced the possibility of lateral buckling. [154] The remediation was required to alleviate the real and substantial danger existing at the Mayflower Mall caused by the defendant’s negligence. [155] The parties agreed, subject to determination of liability, the following are the hard construction costs incurred by the plaintiff: A. Weld Repair $ 10,119.00 B. Web Repair 22,252.00 C. Web Reinforcement 1,324,210.00 D. Spacers 103,020.00 E. Chord Repair 10,428.00 F. 1998 Repairs 13,021.00 G. Design Update (C.O. No. 1) 74,668.00 TOTAL $1,557,718.00 [156] Agreement was also reached that soft construction costs were in the amount of $845,925.00 and the plaintiff will recover the soft construction costs in the same percentage it recovers hard construction costs. Inspection costs were agreed to be $50,895.00. The parties also agreed Harmonized Sales Tax (HST) is to be included in the calculation of damages. [157] The plaintiff abandoned its claim for the cost of design upgrade (C.O. No. 1) in the amount of $74,668.00, as not being needed to alleviate the danger. [158] Having found the remediation was required, the defendant will pay to the plaintiff the sum of $1,483,050.00 in hard construction costs as follows: A. Weld Repair $10,119.00 B. Web Repair 22,252.00 C. Web Reinforcement 1,324,210.00 D. Spacers 103,020.00 E. Chord Repair 10,428.00 F. 1998 Repairs 13,021.00 Total $ 1,483,050.00 [159] The plaintiff recovered 95.2066 per cent of its hard construction costs and will recover 95.2066 per cent of its soft construction costs, which results in recovery of soft construction costs of $805,376.43. [160] The defendant will pay to the plaintiff soft construction costs of $805,376.43 and inspection costs of $50,895.00. The plaintiff will also recover from the defendant HST due on the above amounts. [161] If the parties are unable to agree, I will hear them as to prejudgment interest and costs. __________________________________ Coughlan, J.