Against the grain
implementing design/build projects with municipal agencies
The Milwaukee Metropolitan Sewerage District (MMSD) provides sewerage services to nearly one million people in a 420- square-mile area. A public agency, the MMSD is governed by 11 commissioners and operates with an annual budget of approximately $270 million. Besides providing services for wastewater conveyance and treatment, the MMSD plays a leading role in flood management within the six watersheds located in its service area.
A subwatershed within one of the six watersheds, Lincoln Creek, has been particularly problematic for the MMSD. Draining much of the northern half of the City of Milwaukee and a portion of the City of Glendale, Lincoln Creek is an almost entirely urban watershed, draining a 21-square-mile area, with a long history of flooding. More frequent flooding in the past decade is attributed to increased run-off and recent storms of a larger magnitude than expected. The Lincoln Creek regulatory floodplain contains approximately 1,600 structures, most of which are single-family residences.
A community effort led by the MMSD is now investing approximately $100 million in a project to manage the flooding. Like many urban creeks and rivers, Lincoln Creek has experienced significant environmental and ecological degradation in recent years. Partially contained within an open concrete channel, the existing creek provides little aquatic habitat and therefore supports little aquatic life. The goal of the Lincoln Creek improvement project is to provide environmentally sensitive flood protection for a one percent probability flood along the nine-mile creek. Creek improvements include removing the existing concrete channel, widening its cross-section to allow more flow, creating a natural streambed, adding wet and dry flood storage, creating new wetlands, adding aquatic habitat features and enhancing green space along the banks.
The flood management project began by assembling a diverse steering committee comprised of elected and regulatory officials, community groups and MMSD personnel. This group helped establish overall objectives for flood management and ecological restoration. Studies and hydraulic modeling were then conducted for the overall Lincoln Creek watershed to develop a comprehensive plan for flood management. The creek was then divided into 10 reaches—each about one mile in length—to better manage more detailed design and construction activities.
One of the 10 reaches, Reach 6, is the focus of this paper. Reach 6 is:
• Approximately 6,500 feet long
• 18-foot, concrete lined channel within a 75- to 90-foot drainage easement
• Exists entirely within a residential neighborhood
• Conveys one percent probability flood flows from 1,960 to 4,900 cubic feet per second with velocities ranging from 4.2 to 11.8 feet per second.
Exhibit 1 represents Reach 6 at the start of construction.
Fast Project Delivery Needed for Flood Management
In early 2000, the MMSD either began construction or initiated the final design of all 10 Lincoln Creek reaches. All of the work at that time followed a typical design/bid/build process used by the MMSD and most public agencies. In this process, engineers are selected based on qualifications and price to provide preliminary and final design services. Construction projects are then bid out and awarded based on the lowest responsible bid. In early March, the MMSD assessed the projected completion schedule for Reach 6. The feasibility assessment included the time required to complete the design, bid the project, obtain permits and construct the improvements. As shown in Exhibit 2, a summer 2002 completion was expected.
When the MMSD updated its plan and determined that all Lincoln Creek improvement work must be completed by fall 2001 to provide Milwaukee and Glendale residents with significant flood protection before the 2002 spring rains, the schedule review illustrated a major shortfall. After assessing the potential for future flood-related damages, the MMSD declared Lincoln Creek an emergency project on March 29, 2000. Doing so enabled the District to utilize alternative procurement methods and to implement other, faster methodologies to complete the work. A new schedule was then developed to assess the potential time-savings of a design/build project delivery. This schedule is also shown in Exhibit 2.
As with most design/build projects, the actual time for construction was not shorter compared to design/bid/build. However, the design/build process significantly shortened the time needed for design, permitting and construction mobilization. The overall schedule was shortened in the following ways:
• Design/build contractors were asked to propose on 60% completed design plans. The plans served as a guide for basic performance criteria, but did not constrain the contractors in specific design and construction details. Each contractor submitted guaranteed maximum price proposals to the MMSD, as well as basic information that allowed the MMSD to initiate environmental permitting with the Wisconsin Department of Natural Resources (WDNR) early in the process. The established process to obtain WDNR permitting and approval would not suffice in a design/build scenario. This led to the WDNR's unprecedented cooperation on this project both in becoming a member of the project implementation team and in establishing a new process for approval.
• The selected design/build contractor took the design to a 90% level three weeks after the contract was awarded. The 90% submittal allowed the MMSD to complete its technical review and the WDNR to finalize environmental permitting. It was also sufficient for the contractor to mobilize for construction. The 100% design was completed after the contractor was in the field. Not only did the process allow the contractor to mobilize more quickly, the design/build team was able to incorporate construction methods better suited to actual site conditions in the final design. The design/build process allowed the design to proceed simultaneously with the construction planning. The 90% design plans were submitted to the WDNR at the same time the contractor submitted its construction plans. If the traditional design/bid/build process had been utilized, there would have been significant time between the design plans being finalized and the contractor being hired, plus time needed for the contractor to complete its construction plans.
• As shown in Exhibit 2, the parallel efforts of design completion, environmental permitting and mobilization ultimately enabled the contractor to start construction a full six months before planned construction based on a projected design/bid/build schedule.
Issues for Public Agencies Procuring Design/Build Services
Implementing a design/build project poses significant challenges for a public agency, particularly in coordinating with other agencies and local governments. To successfully deliver a project, the following issues must be considered.
Involvement of Other Regulatory Authorities
Prior to starting any construction in Wisconsin waterways, a permit under Chapter 30 of the Wisconsin Statutes is required from the WDNR. Various components of the Lincoln Creek project design had to be included as part of the Chapter 30 permit application, including final plans and specifications. The WDNR also reviews hydraulic and hydrologic impacts to verify flood-plain zoning requirement compliance and make sure there are no increases in flood stages downstream because of the changes and no adverse affects to downstream properties. In the design/bid/build process, plans are completed before obtaining contractor bids. Final plans in the design/build process, however, are completed after construction begins. Therefore, it was necessary to bring the WDNR into the Lincoln Creek project early. In addition to getting information to the WDNR earlier than in the traditional process, the WDNR became a vital member of the project implementation team. The team consisted of members from both the WDNR and the MMSD, and a consultant in the role of Owner's Representative who was hired by the MMSD. The Owner's Representative brought vital experience to the team relative to the design/build process. The project implementation team established the criteria under which the project was designed, constructed, monitored and maintained.
Design and performance criteria and measurement, testing and monitoring requirements were established in the Reach 6 Request for Proposal (RFP) to allow the MMSD and WDNR to establish project requirements without finalizing the design plans. These criteria set the baseline project expectations to be met by the design/builder.
The WDNR also required the selected contractor to provide plans detailing its proposed dewatering, erosion control, material management and construction sequencing. Because the contractor prepares these plans, they are typically not submitted until after final selection. In the design/build process, however, these plans were required as part of the Reach 6 proposal.
The performance criteria identified all the components required by the WDNR to issue a Chapter 30 permit. The WDNR also agreed to accept substantially complete plans and specifications, rather than final plans and specifications, as the basis for the permit.
As a partner in the project, the WDNR reviewed all design/build proposals, offered input on the submitted design alternatives and approved project components that fell within its jurisdiction. This enabled the WDNR to gain confidence in the process and in the selected design/builder.
As the project construction progressed, the design/builder modified its construction sequencing plan. Because the WDNR was a partner in the project from the beginning, it understood both the project and the need to modify processes from time to time to enable a successful project delivery. Thus, permit modifications were timely.
Involvement of City and County Governments
Because Lincoln Creek is in an urban setting, many of the components overlapped with both city and county governments. As such, two permits were required from Milwaukee County—one from the Department of Public Works (DPW) and one from the Department of Parks, Recreation and Culture (DPRC). The County was willing to accept substantially complete plans, rather than final plans, to issue the permits. Because construction occurred in the creek underneath a county-owned bridge, DPW required that it meet County specifications and approval. The DPRC permit was needed because a portion of the construction took place on land it owned, and an access road had been planned at that location.
A number of city-owned utilities cross the creek, including sanitary sewers and water mains, and the city had many storm sewer outlets that required modifications as part of the construction. As part of the project, a memorandum of understanding between the MMSD and the City of Milwaukee was negotiated to address the impact on the utilities and each party's roles and responsibilities.
Utilizing the design/build process means that some of the final design features are developed just prior to construction. This requires a fast turnaround process for the review of the infrastructure owners. Because both Milwaukee County and the City of Milwaukee were committed to the project, they instituted processes that allowed for quick responses to keep it moving forward. The project was a model of intergovernmental cooperation.
Developing Design/Build Contractor Selection Documents
This was the first design/build project undertaken by the MMSD, which meant developing new processes, procedures and implementation plans. For example, although the MMSD included preliminary plans and specifications with the original Reach 6 RFP, the design/build proposers were encouraged to submit alternative designs for consideration. These alternatives had to incorporate the information provided with the RFP, including the hydraulic and hydrologic modeling, geotechnical data, performance criteria and measurement and testing criteria.
The performance criteria established the design requirements and criteria needed to address the project's technical, environmental and aesthetic components. These included flood conveyance, channel and bank stability, habitat creation, water quality, property ownership and right-of-way constraints, de-watering, erosion control, materials management, construction sequencing and operation and maintenance. The components identified in the performance criteria served more than one purpose. They ensured the final project met the MMSD's requirements, and they confirmed all information needed by the WDNR to permit and approve the project was provided.
The design/builder was also required to prepare a monitoring plan as part of the performance criteria. This plan described all relevant monitoring parameters related to the reconstructed flood conveyance channel components that needed measuring and monitoring to evaluate project performance. The monitoring activities are expected to continue for 24 months following substantial completion of the construction work. In addition to the maintenance requirement, the design/builder was required to develop a plan establishing maintenance requirements for all components of the constructed project.
Measurement and testing established a quality control system that produced an end product in compliance with contract requirements, including specified tolerances. Responsible for quality control, including establishing and maintaining an effective system, the design/builder was required to employ a QC system manager, who also served as the Environmental Compliance Officer. This individual was responsible for making sure all aspects of the project were constructed to meet both the owner's performance criteria and criteria required for the WDNR permit. This role went beyond construction oversight because the QC manager had the duty to halt construction if the performance criteria or permit requirements were not being met.
Using the traditional design/bid/build delivery method, the MMSD would have assigned a resident engineer and resident inspector to assist the project manager in the construction phase. In the design/build delivery, however, the MMSD hired an owner's representative to assist the project manager. This was done because the designer has a relationship with the builder, rather than the owner, and responsibility and liability for the project performance lies with the design/builder.
Because this was the first design/build effort undertaken by the MMSD, considerable time and effort was spent in preparation for the project. This includes creating the performance criteria, measurement and testing criteria, quality control system, required processes and procedures, and continuing oversight of the project.
The MMSD used a two-step process to select the design/builder for the Lincoln Creek project. The first step was requesting Statements of Qualification from interested design/build teams. After reviewing the submittals, three teams were invited to participate in the second step, which required a proposal based on the criteria established in the RFP.
Because the WDNR had to modify its processes and procedures to accommodate design/build, it was part of the selection team. This enabled the WDNR to check design alternative components for compliance with its regulations and review the erosion control, dewatering, construction sequencing and materials management plans required to issue the Chapter 30 permit.
Once the three proposals were reviewed, the MMSD interviewed two design/build teams. These interviews clarified portions of the proposals and provided insight regarding how well each design/build team worked together.
The contract was awarded on July 24, 2000 and substantially complete plans and specifications were submitted in August 2000. Notice to proceed was issued on August 29, 2000 with substantially complete construction scheduled for September 30,2001.
Project Implementation and Evaluation
By using the design/build process, the MMSD realized substantial savings through faster design completion and mobilization, a more robust and cost effective design, and the timely management of changing field conditions. These savings were realized because the design/build caused the designer to work directly with the contractor. This team effort led to better understanding by both the designer and the contractor concerning the other's role, and provided better communication between the two, which led to consideration of the materials and construction means and methods when establishing the design. Exhibit 1 provides a look at the newly reconstructed Lincoln Creek resulting from the design/build team's efforts.
While concurrent design, permitting and mobilization efforts saved six months on the overall project schedule, the design/build process ultimately produced an aesthetically improved design that was both easier and cheaper to construct than what was originally proposed. Recreating natural streams and environments involves complex designs that integrate bioengineering, geotechnical considerations, hydraulic criteria and aesthetic goals. The constructability and final construction costs are primarily a function of both the materials selected to reconstruct the stream and the difficulty of handling and placing those materials. The MMSD, WDNR and design/build team went through multiple iterations of design options, enabling the team to meet overall design and regulatory criteria, minimize overall construction costs and simplify field operations.
For example, one key design element was large, angular, limestone blocks used to provide support for the creek's relatively steep side slopes. The 60% design assumed three rows of rocks. The particular arrangement and orientation of those blocks went through multiple iterations until an acceptable design was found using only two rows of rocks while still meeting hydraulic, geotechnical, material handling and availability and aesthetic objectives. This optimization could not have been accomplished in a traditional design/bid/build process.
At the same time, engineers do not typically specify a contractor's means and methods. Instead, they develop detailed plans and specifications that are considered sufficiently conservative to meet design criteria while allowing a variety of contractor-selected materials and construction methods. Engineers typically expend a great deal of effort developing these detailed drawings and specifications in the final design process. However, it is much more efficient to have the designers work with the contractors to mutually agree on materials and design details that are adequate for the project. The final design then documents these collaborative design decisions.
On the Lincoln Creek project, for example, it was assumed through much of the design that angular rip rap would be required for the creek bottom. However, the contractor noted that rounded riprap would be much easier to place and cheaper to purchase. The design/build team then agreed to change the design to incorporate rounded riprap because it still met the initial design objectives, but provided the MMSD with a cost and schedule advantage.
The design/build process also allowed more flexibility in actual construction techniques. Some initial design ideas would have added significant complexity to the construction efforts. As an example, the contractor requested early on that the bottom of the existing concrete channel be left in place as a road to construct the channel's sidewalls. The bottom would then be removed when the walls were done. This methodology ensured the contractor's equipment would always have a strong base to work from—especially important in a wet creek bottom. The design/build team then developed a design and construction sequence to accommodate this idea.
Many public agencies are looking closely at methods to deliver better projects faster and cheaper, and design/build is one being increasingly considered by local governments. As shown in this paper, the design/build process has distinct schedule advantages over design/bid/build. For unusual and technically complex projects, design/build has the ability not only to compress project timetables, but also to deliver more creative and effective solutions.
At the same time, local governments and public agencies must consider the drawbacks of design/build. Specifically, managing a design/build project is far more labor-intensive, it redefines the traditional relationship between the agency and the engineer, it requires additional coordination time with other public agencies, and it redefines the process public agencies typically use for quality assurance and quality control.
Because every project is different, public agencies can determine case-by-case which method provides the greatest return on investment. More and more often, they are finding design/build brings projects online faster, satisfying their constituencies with effective solutions at a competitive price.
Proceedings of the Project Management Institute Annual Seminars & Symposium
November 1–10, 2001 • Nashville,Tenn., USA