Determining the interval between scheduled line clearance vegetation activities is one of the fundamental challenges utility vegetation managers face. In theory, the optimal cycle period is the point at which the cost of prevention equals the cost of deferral. In other words, it is the point of balance between proactive preventive maintenance and reactive corrective maintenance. This relationship between the cost of preventive and corrective maintenance often is depicted as a classic bathtub curve, where the optimal interval is the point of lowest combined cost. If only it was that simple.
The problem with this approach is determining costs. The cost of preventive maintenance would be relatively straightforward if not for the uncertainty as to what is being maintained. One simple approach is to define it as clearance between trees and energized conductors. The assumption is trees grow, so clearance loss is a function of time. That may be so, but regrowth rates vary over time post-pruning. Most interruptions are a result of tree and branch failures rather than incidental tree-conductor contacts, which further complicates the analyses. And, what assumption should be made as to the efficacy of the preventive maintenance action? Interruptions still occur on circuits recently maintained.
Determining the cost consequences of tree-caused incidents is even more problematic. The cost of unscheduled vegetation maintenance work, or hot spotting, can be estimated; the demand for this work is assumed to increase over time, but it, too, varies over time. What about other costs such as the repair of tree-related damages, loss of energy sales and the cost of outages to customers?
These are some of the questions being addressed in a project funded by the newly established Utility Arborist Research Fund. The Utility Arborist Association (UAA) has identified the need to develop a means of assessing preventive maintenance cycle periods as a top research priority.
Maintenance and Risk Management
Utility vegetation management (UVM) programs exist and maintenance practices are performed in an effort to reduce risk to a system. As risk can never be removed completely, it is important to understand the relationship between risk reduction, maintenance efficacy and economic outlay. An important factor of the maintenance optimization study, as defined by UAA, is the relationship between the relative costs of preventive versus corrective maintenance and the relative efficacy of each strategy in reducing risk to acceptable levels.
Vegetation managers face a challenge common to maintenance management in general, that is, how to maximize performance with limited resources. As financial resources become scarce, an optimal balance can be achieved by optimizing cost, system performance and levels of service. Risk management is an important topic that affects most industries. One of the primary management objectives of utility and urban foresters is to reduce the risk associated with large populations of trees in urban and utility forests. Vegetation management activities typically represent a significant expense to both municipalities and utilities interested in reducing risks associated with tree failures.
The intent of the UAA’s project was to support development of a framework to help utilities and other managers of tree populations identify cost-effective maintenance resource allocations and cycle times.
Vegetation Management Resource Allocation
The project began with a literature review that provided foundational reference information establishing the current state of practice. Of the more than 100 articles reviewed, 50 were deemed sufficiently important to warrant development of project-specific abstracts and include in a narrative summary of findings.
The literature review and discussion with industry thought leaders led to the identification of five different approaches being used to determine vegetation maintenance cycle periods and preventive maintenance funding requirements:
1. Clearance model. This approach considers three factors in determining an appropriate cycle:
- Amount of line clearance achieved at the time of preventive maintenance
- Regrowth response rate of the tree being maintained
- Tolerance for incidental tree-conductor contact.
Once the frequency of maintenance is established, the budgetary resource requirements are calculated. Environmental Consultants Inc. has applied this approach widely as a core element of vegetation maintenance assessments performed for nearly 100 utilities.
2. Cost of deferral model. This approach to determining an appropriate preventive maintenance cycle period is based on the work of D. Mark Browning and Harry V. Wiant (“The Economic Impacts of Deferring Electric Utility Tree Maintenance,” April 1997). This oft-cited work, funded by the ISA Research Trust (now TREE Fund), focuses on the cost premium incurred when the preventive maintenance interval extends to the point where conductors are fully enveloped within the tree canopy. The difference between this and the clearance model is this approach is based on determining the inflection point where the cost of work increases dramatically due to an increase in the amount of time it takes to prune trees as crews are working in close proximity to conductors and also due to increasing biomass requiring pruning, removal and disposal.
3. Reliability model. This approach is based on the relationship between the preventive maintenance cycle period and the frequency of tree-caused outages. There are at least two variations of this model. One approach has been to identify an inflection point in the reliability-over-time curve and to establish a preventive maintenance interval just short of the point where tree-caused interruptions increase markedly. Another approach considers economics using a metric such as customer minutes interrupted (CMI). In this approach, a relationship between preventive maintenance (cost) and CMI is defined. Once the relationship has been established, the utility determines how much to invest in preventive maintenance to buy down CMI.
4. Annual increment model. This approach borrows from traditional forestry. This emerging concept considers two actors: the annual increase in biomass as trees grow and annual mortality rates. Maintenance optimization is achieved when the amount of vegetation maintenance work being funded and completed is in balance with both the annual increase in biomass due to growth and the increase in hazard tree population due to mortality.
5. Regulatory mandate. An alternative approach to a preventive maintenance cycle period is based on regulatory requirements. Obviously this is not a “model” but a mandate. Several states have adopted or are considering adopting mandatory vegetation maintenance cycle periods. In these cases, since a fixed-interval cycle is a requirement, attention turns to the amount of work to be performed when preventive maintenance is required. As a result, the decision as to funding level is based on the relationship between work intensity and risk of tree-caused interruptions. A variation on this mandate is to establish minimum tree-conductor clearance requirements.
None of these models involve a direct application of the previously described bathtub curve, nor do they consider costs beyond those directly experienced by the utility. In fact, an industry survey of readers of Transmission & Distribution World indicates a large percentage of industry practitioners believe utilities are doing enough with regard to reliability and have been able to optimize benefits and costs in general. In contrast, T&D World’s Paul Mauldin suggests the industry needs to determine what customers would be willing to pay in terms of higher rates for improved reliability.
The maintenance objective is to reduce the risk of poor outcomes. In the context of this project, the purpose of vegetation maintenance is to reduce the risks trees pose to overhead electric lines. There are two components to the risk equation of interest: the frequency or likelihood of tree failure having an adverse effect on overhead electric lines, and the consequence(s) of such a failure.
The bow-tie analysis technique was selected as an appropriate method for evaluating the trade-offs between causes and consequences of risk. The bow-tie method provides a relatively simple means of characterizing the risk equation using a diagram resembling a bow tie. Preventive measures in the cause-to-incident relationship can be thought of in terms of influencing the likelihood of an incident occurring. Mitigative measures can be thought of as barriers to or modifiers of consequence. They describe the relationship between incident and consequence. When all the potential cause-to-incident and incident-to-consequence relationships are mapped, they can be joined in a diagram that resembles a bow tie. Note that all pathways between cause and consequence are paved through the incident of interest.
Proposed Analytical Model
Probability bow-tie analysis makes it possible to describe mathematically an overall level of risk exposure for a particular incident. This project included development of algorithms intended to describe the relationship between an overhead power system’s exposure to tree-related risks and the consequences of exposure.
The analytical model being proposed considers a wide range of variables selected on the basis of two attributes: their expected significance in influencing outcomes and the availably of data. The genera of variables include those that define the system being maintained; the incident of interest; hazards that expose the system to an incident, including their probabilities; risk-mitigation methods, costs and their efficacy; and the consequence of an incident, both in terms of damage done and costs.
The consequence of a tree-caused incident is of particular interest. The approach being proposed considers three types of cost: those directly related to restoration and repair, other indirect cost to the utility and the societal cost of outages. The cost of an outage to customers can be much greater than the utilities’ direct cost. Recent work by the U.S. Department of Energy’s Lawrence Berkeley Nation Laboratory quantifying the cost of power outages to a variety of customer classifications makes it possible to consider consequences external to the utility.
The proposed analytical model is conceptual in nature, it has not been tested under operational conditions. The logical next step is validation. The first step being considered involves using existing industry data sets to develop a logical straw man representing a typical UVM program. The straw man is intended as a hypothetical yet relatively accurate representative surrogate for the industry. The resulting hybrid virtual UVM program would be used in initial validation testing. Final validation would be completed by conducting in a pilot project hosted by a willing utility.
John W. Goodfellow ([email protected]) is a principal consultant for BioCompliance Consulting Inc. and a vegetation management researcher with 33 years of experience in the electric utility industry. He has held positions for vegetation management, engineering and field services at investor-owned electric and gas utilities, and has managed T&D services for a major contracting organization. He has bachelor’s degrees in forestry and natural resources management from Syracuse University and the SUNY College of Environmental Science and Forestry.
Arkan Kayihan is an internal management consultant at the University of Washington Medical Center. He holds a bachelor’s degree in chemical engineering from the University of Washington, a master’s degree in chemical engineering from Purdue University and a MBA degree from the University of Washington.