Guidelines for Transportation Management Systems
Maintenance Concept and Plans

4. Maintenance Considerations for the Life-Cycle of a TMS

4.1. Introduction

The Systems Engineering process is more than just steps in systems design and implementation; it is a life-cycle process. It recognizes that most systems are built incrementally and expand over time. The basic steps in the process do not change. There is an even stronger need to provide feedback and assessment with each incremental deployment phase so that future phases build on and expand the system, rather than simply replace elements of the earlier phases. The "V" diagram used to depict the systems engineering model in Chapter 3 can be used to illustrate this evolutionary deployment process. Figure 4-1 shows how successive "V" diagrams illustrate the multiple deployment phases with each phase following the systems engineering process of definition, decomposition, implementation, recomposition, integration, and testing. Not shown in this graphic, but implied, is the feed-back and assessment crosscutting activities that both validate older requirements and generate new requirements in each subsequent version.

As defined in Chapter 1, responsive maintenance is the repair or replacement of failed equipment and its restoration to safe, normal operation. Preventive maintenance is the activity performed at regularly scheduled intervals for the upkeep of equipment.

Figure 4-1 The System Engineering Life-Cycle Applied to an Evolutionary System Deployment Model D

It is important to stress that as each phase or version is deployed, the maintenance concept and requirements are going to change. The concept of operations is also changing as each set of additional requirements is implemented.

This chapter is the logical follow-on to the previous chapter, which introduced the maintenance concept and its context within the systems engineering process. This chapter also describes a systems approach and structure to support multi-year maintenance program planning, as presented in Chapter 6. Another way to view the life-cycle of a multi-phase evolutionary deployment project is to lay out the activities on a time line. Figure 4-2 shows a typical time line with overlapping sets of activities.

The following sections address the key phases in the life-cycle of a typical TMS and show how the maintenance concept and requirements should be considered and adapted for each phase.

Figure 4-2 Overlapping Activities D

4.2. Key Phases in the Life-Cycle of a Transportation Management System

There are several dimensions to the life-cycle of a TMS, moving through the process of planning, design, implementing, and operating a multi-phase deployment. As part of the systems engineering process, the initial vision will generate a more-or-less complete list of requirements for what the final system will do. These requirements must then be prioritized, either by need, cost, or technological feasibility. These priorities are use d to determine what requirements will be met in each phase.

The ability to defer, or even eliminate, requirements based on a quantitative assessment and priority ranking is a key element in the life-cycle planning of a system. The prioritization of requirements provides the flexibility needed to adjust schedule and budget over the life of the project.

Just as there are different priorities for functional and operational requirements, there are different priorities for the corresponding maintenance requirements. Whatever method or exercise is used to prioritize functional requirements can be used to rank maintenance concepts and requirements. Depending on the number and diversity of stakeholders, it can be difficult to achieve consensus on the ranking of priorities. Some examples of techniques include the "utility weighting" approach, which assigns a utility or weight to each requirement, and then each stakeholder "scores" the relative importance of that requirement compared to others. The product of weights times the score is summed across all stakeholder to develop a composite score that determines an overall ranking. There are also Delphi techniques for soliciting input from a very diverse group of stakeholders.

The decomposition steps of a TMS design should each consider the maintenance aspects and the implications for future recurring costs. These steps are described in the section below.

Operations and Maintenance Concept

At this early stage the maintenance issues that need to be included are such issues as:

• Location of property,
• Institutional responsibilities,
• Staffing levels and reporting requirements, and
• Whether or not to use contractors.

Strategic placement of property can assist in the later maintenance activities. Often the maintenance and operations centers are co-located. This has significant efficiencies in that building and facilities can be shared and communications is enhanced.

A definition of the institutional activities should be considered early in the planning process. Not just the responsibilities within the Agency should be included, but other Agencies in the same institutions and adjacent jurisdictions need to be included. Adjacent jurisdictions and other Agencies may either provide support or revenue in exchange for maintenance coverage beyond items that are owned by the Agency.

There are many significant issues associated with whether or not to contract out all or part of the maintenance activities. Options for contracting are described in more detail in Chapter 7.

Functional and Performance Requirements

Functional and performance requirements need to be considered right at the beginning of the life-cycle. These two effectively are inseparable in that when the function of a specific element is defined its performance should also be defined. Without a performance definition acceptance testing and performance monitoring has no standards with which to judge acceptability. Hence the developer of plan needs to not just define that a device meets some standard but also how long this can be maintained.

For example if the function being specified in the light output from a pixel on a Dynamic Message Sign (DMS) that includes bunches of Light Emitting Diodes (LED) then the initial function may be that each LED need to emit some number of lumens within a specified angle. This is a defined function that can be tested and accepted. However, the performance pixel also needs definition because, some LED's may fail. How many failed LED's constitute a failed pixel? How many failed pixels constitute a failed sign? How long should the pixel continue to exceed the light output requirement? Adding temporal criteria to the LED specification will ensure that the sign will perform as required for some period. For example stating that a percentage of the LED must maintain a light output over a number of years and less than some percentage of the LED must not fail in the time period is one way this can be done. However these types of performance requirements require a testing routine to be scheduled into the life-cycle of that particular device.

Thus the functional activities are defined at the beginning of the devices life-cycle but the performance requirements will occur regularly over time.

Acceptance Testing

Acceptance testing occurs at the beginning of the procurement cycle for a specific device. It is typically linked to a payment item. The maintenance plan can use the acceptance test procedures to repeat testing at later stages in the life-cycle to verify continuing conformance.

Configuration and Asset Management

Configuration management and asset management are continuous applied throughout the life-cycle of a TMS. IEEE Std-729-1983 states "Configuration is the process of identifying and defining the items in the system, controlling the change of these items throughout their lifecycle, recording and reporting the status of items and change requests, and verifying the completeness and correctness of items".

Asset management is used more with physical infrastructure rather than systems and is concerned that the infrastructure be considered and managed to the end of its useful life. Both these operations can be a key component of the lifecycle of a TMS. More detains of configuration management can be found in Chapter 3 and information on asset management is contained in Chapter 7.

Preliminary and Final Design

Within the life-cycle, the maintenance input to the preliminary and final designs is critical. Preliminary designs tend to be more conceptual and it is at this stage that the maintenance planning needs to consider its requirements carefully. Coordination and cooperation between departments is essential. Specifications need to include facilities to ease maintenance. Gantries on variable message signs facilitate access. Co-locating devices can both reduce the costs of installing communications as well as enhance maintenance. For example, many Agencies now specify that traffic detectors must be above ground. However, there are Agencies that are still installing inductive loops. Loops are a significant maintenance burden and, in most cases, are unreliable, expensive to repair, and cause major disruptions to traffic. Having maintenance input in the early design stages can significantly aid the maintenance work.

Final designs tend to get more involved with locations, detailed functions, and connectivity. Maintenance groups should get involved with the final design to ensure that roadside devices:

• Can be accessed. Examples abound of:
  • Cameras that cannot be accessed by a cherry picker,
  • Communication links that are lost,
  • Signs that cannot be seen,
  • Communication equipment where every cable was black and unlabeled,
  • Controller cabinets that are buried by snow plows, and
  • Signs installed by the operations department and then removed by a construction project from another department.
• Can be accessed safely. The shoulders of freeways are dangerous locations and ensuring that maintenance crews can park behind the guardrails is advisable.
• Are protected from the environment. Nearby bridges are usually good locations for equipment cabinets.
• Are easily visible. Foliage that threatens to block signs, etc. must be routinely trimmed.
• Are properly marked and labeled. Contractors should mark signs according to maintenance group formats.
• Have available power sockets and in-cabinet lighting.

The maintenance staff needs to be involved with all levels of design from the conceptual picture of location, institutional, and contractual options down to the level of specifying the additional ports on the board in the controller that will assist in fault-finding.

Implementation

Maintenance group activities during implementation should include inspection and acceptance testing. This could be the group's first opportunity to see the equipment, determine access, and ensure correct operations, labeling, and documentation. Even if the responsibility for the implementation is with the contractor, this is an opportunity for the maintenance group to get their hands dirty. In addition, oversight from the maintenance group will help ensure correct installation.

Management and Operations

Maintenance needs support from the management and operations on an ongoing basis. The coordination with the operations functions is described in Chapter 5.

Maintenance Procedures

Maintenance procedures for all common ITS devices are contained in Appendix B. This appendix includes the frequency of maintenance for each device. These increments need to be scheduled on a regular basis and become part of the ongoing TMS life-cycle.

Monitoring and Evaluation

Monitoring and evaluation of the system for maintenance consists of two components that need to be incorporated into the life-cycle. First is the preventive maintenance that consists of the actions for each device referred to earlier and defined in the Appendix. Second is the responsive maintenance that results when the system or its operators initiate trouble reports.

Both of these need to be considered ongoing activities. Although the preventive actions can be planned, these are often a background task for most Agencies that concentrate on the response tasks. When allocating manpower and resources, Agencies need to be aware of the relationship between preventive and responsive maintenance and do their best to spread their limited resources appropriately.

Another dimension of the TMS life-cycle includes the degradation and obsolescence of system components as they age. Ultimately, the TMS may either be completely replaced or even decommissioned. Many of the technology components in a modern TMS become obsolete rather quickly. Desktop computers can typically be replaced after only three years by a new model that is twice as powerful and costs the same or less as the old one did originally. Other components, such as communications gear (modems, multiplexors, etc.), vehicle detectors, DMS controllers and display components, have somewhat longer life spans. Infrastructure items, like DMS structures and cabinets, have much longer life spans.

An example of life expectancy for ITS and TMS components is shown in Table 4-1. This table is summarized from the ITS Deployment Analysis System (IDAS) and MitreTek/JPO Cost and Benefits Database (March 2002).

4.3. Considerations for Maintenance Throughout the TMS Life-Cycle

The same crosscutting activities that take place during the systems development life-cycle should also encompass the maintenance concept and maintenance management activities throughout a TMS life-cycle. Table 4-2 compares the crosscutting activities in the systems engineering process with the corresponding maintenance concept and requirements analyses.

Figure 4-3 PM Versus Annualized Maintenance Costs D

As mentioned above, a key decision is whether to replace aging or obsolete components. While it is recognized that preventive maintenance will allow most types of equipment to operate longer, it is also generally understood that there will be a point of diminishing returns. That is, at some point, there is little more that can be done to keep a worn out or obsolete device functional and the cost of preventive maintenance exceeds the annualized cost of replacement. Figure 4-3 presents a generic representation of preventive maintenance cost per year versus the cost of replacement, annualized over the expected life span of the device. The point where the preventive maintenance costs per year are the same as the annualized replacement costs is considered break-even and it would be wasteful to continue to keep the old equipment in service.

Significant amounts of empirical data are needed to calculate or forecast this break-even point. The only feasible way to collect these data and perform this type of analysis is via a computerized maintenance management system (CMMS). The minimum data required for this type of analysis includes the original purchase price, replacement cost, PM and repair cost history, MTBF, and MTTR. Chapter 7 discusses the characteristics of CMMS and other maintenance analysis tools. The use of a CMMS tool is highly recommended because it provides a way to allocate maintenance, repair, and replacement budgets more efficiently. The CMMS will also provide much of the history, documentation, and justification for the maintenance budget.

A life-cycle analysis of communications alternatives is critical to every TMS. There are both policy and technical issues that should be considered for the operation and maintenance of this communications infrastructure.

• Agency-Owned Versus Leased: This is often more of a policy decision than a technical one. Many of the TMS's referenced in the research, especially those built prior to 1996, have a communications infrastructure that is owned entirely by the Agency. Since 1996, the year the Telecommunications Act was passed, competition in the commercial data communication market has (somewhat) improved both the cost and reliability of leased communications service. Tennessee DOT evaluated the life-cycle cost of leased service versus owned and concluded, based largely on the cost of maintaining a fiber-optic communications plant, that leased services were more cost effective. This decision was also influenced by the fact that the State of Tennessee had a favorable existing contract for leased services that included a service level agreement for on-going maintenance.
• Design Alternatives and Their Impact on Maintenance: The evaluation the life-cycle cost of TMS design should be considered when developing alternatives. This is particularly important for the communications infrastructure, since it is typically the most expensive element of a TMS in terms of both the capital and O&M costs.
  

4.4. Staffing and Training Considerations

The availability of properly skilled and trained staff throughout the life-cycle of the TMS is critically important to getting the most out of any system and assuring that it meets its intended concept of operations throughout the system's intended life-cycle. There were several practices related to this critical need that were noted in the literature research and interviews. These include:

• Pre-Qualification of Contractors: Often a general pre-qualification process exists for many Agencies. However, as the skills change and complexity increases in a typical TMS, the pre-qualification process must be upgraded to match.
  
• Training by Vendors: Procurement contracts should include a requirement for on-site training of Agency staff in maintenance and operations of the equipment, preferably conducted by the vendor.
  
• Training by Contractors: Procurement contracts should also include a requirement for on-site training of Agency staff in the maintenance and operation of the assembled systems, including software, hardware, and devices.
  
• Training Library: The operating Agency should maintain a library of system documentation and, if available, a videotape library of training.
  
• Staff Retention: This can be difficult in a high-tech environment, but there are ways to improve retention, such as providing for additional training, allowing travel to technical conferences, and workshops and other non-salary related perquisites for Agency staff.
  
• Staffing Qualifications: If procuring contract maintenance staff, specify minimum qualifications by position and require that contractor personnel submit resumes and proof of required certifications. Periodically check on contractor personnel and require pre-approval of new staff prior to approving invoices for their work.
  
• Estimating Staffing Numbers: Staffing guidelines are often Agency-specific, due to the different organizational structures across the country. However, there were several references found in the literature that provide methodologies for calculating workload and using these workload estimates to generate full-time-equivalent (FTE) staff position descriptions.

Keep in mind that the only constant is change. Not only will the system change over time, but also the concepts of operation and requirements will also likely change, as will the stakeholders who generated the requirements. The key to success in this environment is flexibility and a good understanding of priorities for both operational and maintenance concepts and requirements.