Reconsidering Value Engineering

The recent addition of critical library space at the Rhode Island School of Design afforded the opportunity to firmly establish the importance of value engineering in assisting the architect and the owner in significantly improving a facility’s function, in fully respecting budget constraints, and in effecting a library structure of the highest quality. By Matthew H. Johnson, P.E., A.M.ASCE  Author portrait by Skip Brown

Through the years my curiosity has led me to investigate the historical basis of value engineering. I wanted to understand, even simply, what the original intent of value engineering was and who developed the process and why.

As it turns out, the term “value engineering” is neither unique to the professions of structural engineering, architecture, and construction nor even most notably associated with these interrelated fields. Value engineering—originally termed value analysis or the value method—can readily be traced to its development by Lawrence D. Miles at the General Electric Company in December 1947. As was implied in opening this discussion, value engineering is viewed negatively in the realm of building construction—at least in my experience—by consultants. However, the definition as developed at General Electric provides greater insight into the basis of the idea and a preferable goal for the architectural and structural design team. As Miles noted in his book Techniques of Value Analysis and Engineering (New York: McGraw-Hill, 1961), “Inherent in the philosophy of value analysis is full retention for the customer of usefulness and esteem features of the product. Identifying and removing unnecessary cost, and thus improving value, must be done without reducing in the slightest degree quality, safety, life, reliability, dependability and the features and attractiveness that the customer wants.”

John Horner

“The library project at the Rhode Island School of Design provides a case study of value engineering as the ideal was defined sixty years ago at General Electric,” says author Matthew H. Johnson, P.E., A.M.ASCE, top. The main area of the school’s new library is the central banking hall, an ornate space designed in the style of the Italian High Renaissance that covers almost half an acre and rises 50 ft (15.2 m) in height.

In the eyes of many consultants, the “primary tenet” that “quality not be reduced” seems to be lost when value engineering is applied to a project by an owner or a construction manager. Moreover, it rarely seems that function is improved; conversely, it often seems that both function and cost are reduced simultaneously in the name of value engineering. If improving function and maintaining quality are truly a part of the value engineering process, the idea of value engineering would lose some of the negative connotation it carries for many architectural and structural designers. The idea of value engineering would become even more palatable—and, in fact, rewarding—if a design team could work together with the owner and the construction manager toward a common goal to achieve a functionally improved, cost-effective design project of high quality.

The library project at the Rhode Island School of Design (RISD), in Providence, Rhode Island, provides a case study of value engineering as the ideal was defined 60 years ago at General Electric. Over the course of this project Simpson Gumpertz & Heger, Inc., (SGH) followed the tenet of value engineering as defined by General Electric to apply the art and science of structural engineering and the tools of the profession, past and present, to help the architect and the owner increase the function significantly above the existing RISD facility, prevent escalation of the budgeted cost of the project, and provide a facility of exceptionally high quality.

In 2002 FleetBoston Financial, now part of Bank of America, donated the first and second floors of the venerable Rhode Island Hospital Trust Building to risd. The school purchased the balance of the building from the other owners and subsequently announced that a portion of the building, constructed in 1917, would house the university library, providing an increase of more than 400 percent in floor area above that of the existing library facility.

The central space in the new library is the almost 0.5 acre (0.2 ha), 50 ft (15.2 m) tall central banking hall, a supremely detailed and ornate hall in the style of the Italian High Renaissance. The library would ultimately occupy the lowest two and a half floors of the existing 13-story Beaux-Arts building, which was designed by the New York City architecture firm York and Sawyer. A half floor not directly integrated into the library project is the area of the basement that housed the building’s central mechanical facilities and other building support systems. The remaining levels of the building and the overall structure comprising the building and its base constitute a separate project and would not be a part of the library renovation.

John Horner

Both the Rhode Island School of Design and the architect—Boston-based Office DA—wanted to treat the library as a living room for the 500 students who are housed in the nine floors of dormitory space above it. This vast hall provides spatial arrangements for both group interaction and individual study.

The building is located at 15 Westminster Street in Providence and is a short walk across the river from risd’s main campus. The combined goal for risd as well as for the architect, Boston-based Office dA, was to treat the library as a living room for the 500 students housed in the nine floors of dormitory space above it. Not only would the library offer unprecedented direct access to the extensive collection of books and visual media in the risd collection; it would also offer multiple spatial arrangements for both group interaction and individual study. Furthermore, risd wanted to create a space in which students could browse through the existing collection of more than 130,000 volumes and 685,000 images. However, even with the generous donation by FleetBoston Financial, the space allotted to the library was insufficient to house the entire program planned by Carol Terry, RISD’s director of library services.

What does all of this mean and why is structural engineering—let alone the idea of value engineering—central to a project that was not originally envisioned as a significant effort in structural engineering and structural construction? Despite budget limitations, SGH consulted reference works from the period of original construction and employed modern destructive and nondestructive investigative techniques and basic engineering principles that enabled the owner to realize substantial savings and obtain a library of higher quality and functionality. SGH’s work also enabled the architects to achieve their program goals without reinforcing, at significant cost, the existing, antiquated structural framing system. Early in its involvement Office dA recognized the potential limitations of the budget with respect to its ambitious goals for the architectural and programmatic design and its innovative concepts for providing risd with the majority of the desired program within a discrete space and budget. Office dA’s principals knew from experience that in order to successfully control costs and implement a complex program, a team of strong technical consultants would be required to work through many ideas—typical and atypical—to formulate a solution consistent with the budget but founded on improved function and quality.

The discussion for the project started out with a simple meeting between SGH and Office dA project participants. sgh was already working with Office dA principals Nader Tehrani and Monica Ponce de Leon on a small project in China, an interior design project in Boston, and the early design planning of the Macallen Building, a large, multifamily luxury residential project in Boston. Through a request for qualifications Office dA had been short-listed, interviewed, and chosen by risd to lead the design of the library project. Now Office dA was conceptualizing solutions to the problem of space and program. Even though the library area would be almost four times larger in plan square footage than risd’s existing facility, Terry expressed concern that the new library would not be large enough to house the entire program she believed was necessary to support one of the nation’s leading design schools.

John Horner

The larger of the two interior structures designed as colossal, full-story-tall pieces of furniture, the seating pavilion, opposite, unfolds from the upper level of the east side of the seating pavilion down to the center of the banking hall. The space beneath the seating pavilion houses rooms for individual and group study, above.

Following the design team’s first meeting at the building with Paul Mullen, risd’s director of construction planning and management, risd provided sgh access to the drawings in the bank’s flat files. At this time the bank was in the process of vacating the building and its files were still accessible. We were fortunate to find partial sets of original architectural drawings as well as structural framing design and erection drawings. There were hundreds of additional drawings stored in two 5 ft (1.5 m) tall drawers of flat files documenting renovations and various other types of work carried out during the building’s almost 100-year history. At the time many of these seemed insignificant and were not retrieved. However, they would later prove to be valuable as we began to understand the building, the sequence of renovations, and the original and new demands on the capacity of the existing structure.

The existing structural and erection drawings were dated 1906 and 1917, although it is believed that construction of the building did not begin until 1917 and that a second phase was initiated in the mid-1920s. The framing of the banking hall floor, which is just above the existing exterior grade, is a structural slab of cinder concrete with one-way reinforcement supported on structural steel that is monolithically encased in the reinforced cinder concrete. Vertical support is provided by riveted plate and angle sections built up to form I-shaped columns. All beam-to-girder and beam-to-column connections, as described by the drawings, are riveted double-angle shear clips and riveted angle-seated connections. Typical column bays measure 16 by 16 ft (4.9 by 4.9 m). Typical spans of the floor slab of cinder concrete with one-way reinforcement measure 8 ft (2.4 m). The foundations of the building are wood piles, which support concrete-encased steel grillages. The interior partition walls are clay tile blocks. The exterior cladding of the building is predominantly limestone, and less prominent areas are clad in a standard brick. The exterior backup wall construction generally appeared to be clay tile blocks constructed around the exterior steel skeleton.

To assist Office dA in assessing the feasibility of its program, sgh had to make a rapid determination as to whether the existing structure could support the required library live load of 150 psf (7.18 kPa) for book stacks. A review of the existing drawings explicitly indicated the materials and design thicknesses as well as the original design live load. Furthermore, the existing shop drawings indicated the structural framing sizes; for example, typical beams were 12I31½ and 12I40 shapes spaced 8 ft (2.4 m) on center. The yield strength of these members was determined from reference works of the period to be 30 ksi (206,842 kPa). Initially, these shapes could be approximated by checking modern S-shaped beams similar in geometry to the existing sections. We used the ram Structural System, developed by ram International, of Carlsbad, California, to prepare a preliminary analysis model of the existing framing geometry as depicted in the existing drawings. This model included the uniform dead load based on the construction depicted in the drawings and the original design live load of 150 psf (7.18 kPa). Moreover, we developed line loads to represent each beam and girder self-weight, including the concrete encasement. At this early stage we approximated the analysis dead loads based on the existing plan information. In accordance with the governing building code for this project, the 2003 edition of the International Code Council’s International Building Code (IBC), the appropriate live load for a library stack room is 150 psf (7.18 kPa), and at that point our confidence was high that the existing floor would be adequate to carry this load because the existing drawings indicated that the floor was designed for a live load of 150 psf (7.18 kPa).

As expected, the results of the analysis indicated that the existing drawings were accurate: the bare steel sections could support the original design dead loads and a design live load of 150 psf (7.18 kPa). At this point, we analyzed the continuous floor slabs of cinder concrete over the beams. We treated these as concrete slabs with one-way reinforcement, modeling the beams as unyielding supports. The existing drawings indicated that the slabs of cinder concrete were reinforced with a smooth wire fabric with 0.255 sq in of steel area for each foot of slab width. We performed a terse analysis assuming concrete and reinforcing steel properties based on historical material strengths commensurate with the age of the building and the continuous beam approximate moments outlined in the American Concrete Institute’s standard 318-02 (Building Code Requirements for Structural Concrete). For historical information about materials and construction practices, we relied heavily on the Kidder-Parker Architects’ and Builders’ Handbook: Data for Architects, Structural Engineers, Contractors, and Draughtsmen, by Frank E. Kidder and Harry Parker (New York City: John Wiley & Sons). This text was published continuously from 1884 to the middle of the 20th century. Volumes between 1906 and 1946 are available in the SGH library, as are other reference works for the period in question. Our preliminary analysis indicated that the slab construction had the necessary shear and moment capacity to support the design dead and live loads.

We next considered the effects of the proposed loads of the two new structures on the existing building. These structures, the seating pavilion to the west and the circulation island to the east, serve as furniture—both literally and figuratively—on a massive scale. As previously noted, they provide a variety of functions within a small footprint at each end of the banking hall. The seating pavilion, the larger of the two structures, comprises two fully accessible stories. Its upper level, which serves as a seating lounge with individual desks integral to the structure’s perimeter millwork, is connected by a bridge to the existing mezzanine on the north side of the banking hall. This bridge serves as a means of egress and provides access to those incapable of using the stairs.

An egress stair/stadia seating grandstand unfolds from the upper level of the east side of the seating pavilion down to the center of the existing banking hall. The space under the seating pavilion houses rooms for individual and group study. The perimeter of the seating pavilion incorporates individual study cubicles with integral lighting and electrical outlets. The west side of the seating pavilion contains additional shelving integrally constructed with the other features of the structure. We designed both the upper level of the seating pavilion and the study rooms below for a live load of 60 psf (2.87 kPa), consistent with the IBC criteria for library reading rooms. The stair/stadia grandstand is designed for a live load of 100 psf (4.79 kPa), consistent with the IBC criteria for stairs. We considered designing the stadia seating areas for 60 psf (2.87 kPa), consistent with the IBC criteria for areas of assembly with fixed seating, but because the effect on the existing structure was small, we chose to provide the additional flexibility 100 psf (4.79 kPa) allows.

The circulation island, as noted above, is located at the east end of the existing banking hall. The front end of this canopy structure, which features open scissors trusses of varying depths, houses the necessary equipment to conduct such library activities as material checkout, while the rear of the island houses staff offices. The perimeter of the island includes shelving and additional individual study cubicles around the three closed sides.

The exterior treatment of each interior structure surrounding the seating pavilion and the circulation island—indeed, the entire structure of the circulation island—is part of a massive custom millwork package. Office dA is resolute about investigating the potential of materials and spaces and typically does so through the exploration of traditional and digital techniques of design and assembly. The entire millwork package of desks, shelves, stairs, walls, et cetera—including the entire circulation island—was digitally fabricated by means of a CNC (computer numerically controlled) routing process. Each piece was modeled and then converted to fabrication drawings by Office dA. The final product was assembled by on-site labor. This technique enabled Office dA, through iterative digital modeling, to maximize the use of every segment of space to provide the greatest number of possible uses within the larger space. This resulted in greater functionality within a fixed square footage. In essence Office dA increased space to increase function through the creative use of a millwork budget.

SGH initially conceived the seating pavilion structure as a structural steel frame positioned off the existing one-story columns supporting the floor framing beneath the proposed islands. This would provide maximum flexibility with little effect on the existing beams and girders and the slab construction of the floor. However, the construction manager believed the cost of such a scheme would have been prohibitive. Furthermore, review with Office dA indicated that such a system was, in fact, inflexible owing to the height restrictions imposed on the bridge linking the existing mezzanine level on the north side of the banking hall with the upper level of the seating pavilion.

Reviewing the original 150 psf (7.18 kPa) design live-load capacity of the floor and the new live-load capacity of the seating pavilion—120 psf (5.75 kPa) combined for both levels—it seemed reasonable to construct a structure featuring bearing walls with cold-formed steel studs and a plywood-covered metal deck to accommodate the net 30 psf (1.44 kPa) of dead load available. This idea emerged from a discussion with Office dA about strategies not only for the structure but also for the extensive and complex millwork package that was proposed to surround and be integral with the seating pavilion structure. We agreed it would be more consistent to use the light framing of bearing walls with cold-formed steel studs for what was essentially a large-scale piece of millwork, albeit a complex one with many uses. We proposed a metal floor deck to maximize the spacing of the cold-formed steel joists and support the dead and live loads of the second-level seating lounge. The plywood over the metal deck would provide a continuous flat surface and a subfloor for the cork flooring that adheres to it.

The results of our analysis of the existing floor structure—including the additional loads from the cold-formed steel bearing walls—indicated areas of local member stresses in excess of the allowable stress. We met with Office dA to review framing strategies that would reduce the demand on these existing members with only subtle modification to the geometry of the seating pavilion. We provided Office dA with a plan of proposed bearing wall locations and limits, along with a matrix that outlined a progression that showed which walls could work together without creating overstresses in the existing framing. Office dA used this to slightly revise the configuration of the seating pavilion. The circulation island at the east end was light enough that it did not overstress the existing structural framing members.

John Horner

The upper level of the seating pavilion serves as a seating lounge and is connected to the existing mezzanine on the north side of the banking hall by a bridge, above.

During this review, we found a single architectural drawing from a midcentury renovation that indicated not a depression in the middle of the banking hall but rather a built-up floor to each side of center. We immediately went back to the bank’s flat files and began a search for more than just original architectural and structural drawings. Several hours of searching unearthed a set of electrical drawings that indicated as much as 6 in. (15.2 cm) of additional concrete had been placed on each side of the banking hall to encase a Walkerduct system (produced by Wiremold/Legrand) in the late 1960s. The Walkerduct is an electrical raceway system and it was used to organize and conceal within a concrete slab system the “new” wires that in the 1960s were part and parcel of the emerging electronic banking age. The wires were required by the development of IBM's System/360 mainframe computers, which were then becoming prevalent in banking. In an existing building with a surfeit of live-load capacity and limited access in the basement to the underside of the structure, such a product was probably seen as an inexpensive, quick solution to improve the banking operations.

Initially quite pleased to have unearthed this information, sgh needed to understand the elevation differences and further its analysis. My satisfaction quickly dissolved, however, as I realized this meant that the original design live load of 150 psf (7.18 kPa) had just been reduced to as little as 75 psf (3.59 kPa) by the addition of a concrete topping slab. We were currently engaged in a project that proposed a nonstructural renovation to almost half an acre of century-old structural steel framing that appeared grossly inadequate for the required library stack loads as well as the interior structures so cleverly conceived by Office dA to add quality and function within the cost confines of the project—the true objective of value engineering as originally defined in the 1940s.

At this point in time the analysis became more complex and more precise. In the time spent thus far, we had been performing more general analyses as the architectural concepts were being developed. Our confidence that the existing structure would be adequate had initially been strong given the live loads published in the original design drawings. Now, however, with the project’s budget under serious pressure from the potential structural renovations, we had to sharpen our pencils and delve more deeply into the literal nuts and bolts (or, more precisely, rivets) of the design. This required us to hit the books—no pun intended—consulting sources from both today and the time the building was originally constructed as a way to refine our analyses. We also had to engage the services of engineers from our testing laboratory to assist us in delving into the details of the existing construction.

The first step was to review the Specification for Structural Steel Buildings, published by the American Institute of Steel Construction (AISC), to verify the requirements of a composite analysis accounting for encasement by reinforced concrete. On the basis of our reviews to date, the beams and girders met the basic criteria. The result, as indicated by the AISC publication, is the ability to assume that the allowable stress in the steel is 76 percent of the yield stress. In our preliminary analyses, we had been using a yield stress of 30 ksi (206,842 kPa) and an allowable stress (0.6 . Fy) of 18 ksi (124,106 kPa). We had verified this through many sources, including the work by Kidder and Parker cited previously, the AISC’s Iron and Steel Beams, 1873–1952, the Carnegie Steel Company’s Carnegie Pocket Companion, and other works in our library from the steel producers in the early part of the 20th century. The AISC’s first Manual of Steel Construction was not published until the mid-1920s, almost 20 years after the original building project design was initiated, but we used this as a resource as well. We also used subsequent versions of that manual to determine the history and development of composite beam design. In numerous places on-site where the bare steel had previously been exposed, we observed that the reinforcement in the cinder concrete was continuous not only in the slab but also in the encasement around the beams and girders. The dimensions appeared always to meet the minimum cover requirements outlined in the aisc’s Specification for Structural Steel Buildings. Details in Kidder and Parker pertaining to the original floor structure supported this conclusion as well.

We updated the steel yield stress obtained from the RAM Structural System to a hypothetical value so that when the RAM system multiplied this fictitious steel yield stress by 0.66 for a continuously braced steel beam the allowable stress would equate to 0.76 multiplied by the yield stress of 30 ksi (206,842 kPa) for the original structural steel sections. Furthermore, we updated the analysis model to reflect the additional topping of lightweight concrete that was added to encase the Walkerduct system. A more thorough review of the drawings indicated that the topping slab was of lightweight concrete. Our allowable live load was still below the required 150 psf (7.18 kPa) but was an improvement on our original assumption that the concrete was of normal weight. We considered the topping slab as dead load only and did not consider it to be composite with the original construction. The majority of this concrete, sadly, was placed over the original marble floor. There was a brief discussion with the construction manager, Shawmut Design and Construction, of Boston, and risd about demolishing the topping of lightweight concrete, but the expense was considered too great and a marble surface would not have been conducive to a quiet library environment even if it could have been restored.

At this stage the models also included the updated loading from the bearing walls of the west side seating pavilion. They also included the loads from the built-up posts of cold-formed steel required for the continually developing architectural details of the seating pavilion and bridge structure. The resulting analyses indicated that the increased allowable stresses based on a composite concrete encasement of the steel beams would be able to impart the necessary strength to support the required live load of 150 psf (7.18 kPa) in addition to the concrete topping slab of the Walkerduct system. In overview, this was logical since the net total load represented an approximate increase of 20 percent above the original net total design load, and the allowable steel stress in the composite section represented an approximate 26 percent increase over the noncomposite bare steel section. The working stresses were generally below the limit of the allowable composite steel stress, and in the few cases in which they were above the allowable stress, the amount was usually less than 10 percent. However, the working load stresses and allowable steel stresses were sufficiently close that we believed it was prudent to refine the analysis and perform additional investigations to verify the load and the allowable stresses in the steel sections. Our concern at this stage was to be sure we were not underestimating the existing dead loads.

The use of commercially available software as flexible as the ram Structural System was invaluable in supporting the many iterations of the design. Up to this point we had been using standard S beams from the latest aisc shapes database. We decided to modify the ram software tables to reflect the actual beam properties published in tables at the time of construction. In the early 20th century there were many mills producing similar shapes, but the shapes were not as standardized as they are today. Similar shapes varied slightly in cross-sectional area and, therefore, in section properties. Using the aisc’s Iron and Steel Beams, 1873–1952, we developed a matrix of all of the producers within the range of years around this project and the steel shapes used in the banking hall floor. Using a simple spreadsheet, we could mark each size available from a producer to see which producer, if any, provided all the available shapes. In the end, we determined that Jones & Laughlin Steel Company was the only firm that produced all of the shapes used in the original design. This seemed an acceptable result as that company was headquartered in Pittsburgh, relatively close to the site, and was one of the largest steel producers in the United States at the time. We modified the ram software tables using the properties of each steel section by this producer as outlined in Iron and Steel Beams, 1873–1952.

At the same time, we requested that the contractor remove through-thickness cores of the concrete slab, the finish floor, and the additional concrete topping in 12 locations around the banking hall floor slab. The contractor removed cores where Office dA planned to locate shelving stacks, the interior structures, and other areas of concern for our office. We retrieved these cores and brought them back to the office. In this case, we used the staff in the sgh testing lab to separate the cores into their constituent materials and determine the thickness and density of each material. We used this information to understand the variation in loads across the slab and the material densities so that we could make an accurate assessment of the correct dead loads to be assumed. Thus far, we had based our information about the existing dead loads on the various building drawings and on modern reference works and reference works from the period in question. Using the information from the cores, we verified the average and upper bound loading in the slab system. While some material densities or thicknesses were slightly different than we had assumed, the net effect was a slight reduction in the dead loads we had been working with and an overall corroboration of our work. Furthermore, we determined from the 5 in. (12.7 cm) diameter cores that the reinforcement fabric spacing and diameter equated to 0.295 sq in. per foot (624 mm² per meter) of steel reinforcement. This reinforcement is approximately 15 percent more than the original design drawings indicated.

We made an additional visit to the existing risd library to review the current book stacks, photo storage files, and other heavy items that would be reused within the new library space. The IBC provides explicit requirements pertaining to the use of a stack live load of 150 psf (7.18 kPa). We measured the typical metal shelf system, the number of shelves in use, and the average amount of unused shelf space. We used this information and a density of 65 lb/cu ft (1,041.3 kg/m³) as published by the IBC to determine the maximum and average stack loads. The proposed stack layout complied with the IBC requirements, and on the basis of our review and analysis it did not, on average, exceed the live load of 150 psf (7.18 kPa).

With this information, we reviewed the slab capacity based on the updated dead loads and design live loads. The slab, which typically spans 8 ft (2.4 m) centerline to centerline between beams, also needed to support the additional concrete load from the Walkerduct system as well as certain live loads supporting parts of the seating pavilion, seating pavilion bridge, and circulation island. We also looked at the local effect of the book stacks. The IBC library stack room loads are based on an average load for book stacks with a particular ratio of height to width, a particular book density, and maximum aisle size. The load under the stack itself, however, may be much higher than the 150 psf (7.18 kPa) required as a uniform live load. As such, we investigated this also, both as a uniform live load and as point loads from the shelf feet. The slab’s shear and moment capacity exceeded the demands in all cases. For this analysis, we used the recently verified steel area in the slab, formulas defined in Kidder and Parker for similar slab systems from that era, and, as a modern source, the American Concrete Institute’s standard 318-02.

As an additional measure, we used ground-penetrating radar (GPR) to analyze the existing slabs to determine the extent of the Walkerduct system placed above the banking hall’s existing slab construction. The gpr enabled us to determine the extent of the ducts, their width within the topping slab of lightweight concrete, and the approximate height of the void created by the duct. We used this information as a possible way to further reduce the dead loads of the system in local areas. In the end, this refinement was not necessary. We understood that while we were sufficiently close to the maximum capacities of many members, the reduction in dead load afforded by these voids would provide an additional margin of safety.

In a project in which we are continually refining assumptions on the basis of new evidence and applying modern methods of engineering to older structural systems, we needed to be careful not to be overly aggressive in reducing loads or increasing allowable stresses. While an aggressive design based on sound logic is one definition of good engineering practice, we needed to make sure that the accumulation of approximate assumptions did not continually err on the side of unconservatism so as to end up with an inaccurate analysis. During the course of the project we made sure through consultations with other engineers in the office and independent interoffice reviews to corroborate our assumptions and approach to the project.

Following a year of diligent review, investigation, and iterative analysis, design, and coordination, the project seemed to be coming together. The beams, girders, and slabs all appeared to have the shear and moment capacity necessary to support the original design loads, the additional topping of lightweight concrete, the seating pavilion, the bridge, and the circulation island. The last step was the investigation and analysis of the connections of the beams and girders.

Shawmut Design and Construction initiated destructive work to expose the connections we requested access to, and SGH made another trip to the site to carry out a photographic survey of the connections and obtain information about their dimensions. Fortunately, the connections were consistent in type and geometry for each location reviewed. Again, using the reference works from the period in question mentioned previously as well as the Guide to Design Criteria for Bolted and Riveted Joints, by John W. Fisher and John H. Struik (New York: John Wiley & Sons, 1987), we combined our field observations with material property and construction practice data from that period to determine that the capacity of each connection was sufficient to support the additional loads.

In the end the project took in excess of two full years to design and construct. Office dA and SGH were able to help RISD achieve a new library on budget with significantly improved functionality and exceptional quality: true value engineering.

The library opened to the students in June 2006.

Project Credits
Owner: Rhode Island School of Design, Providence, Rhode Island (Paul Mullen, director of construction planning and management, and Carol Terry, director of library services)
Design architect: Office dA, Boston (Nader Tehrani, principal; Monica Ponce de Leon, principal; and Arthur Chang, project manager)
Construction manager: Shawmut Design and Construction, Boston
(Matthew Dempsey, project manager; Ludger Bain, assistant project manager; and Timothy Barges, project superintendent)
Structural engineer: Simpson, Gumpertz & Heger, Inc., Waltham, Massachusetts (James C. Parker, P.E., M.ASCE, senior principal; Matthew H. Johnson, P.E., A.M.ASCE, senior project manager; Michael A. Peddie, A.M.ASCE, senior engineer; and Peter M. Babaian, P.E., A.M.ASCE, staff engineer)