An end-use approach to reliability investment analysis Clinton J. Andrews Reliability investments occurring on the customer side of the meter are becoming increasingly important, and warrant consideration when planning utility eflorts to improve electricity reliability. Thispaperpresents an end-use approach to the analysis of reliability investments, and suggests that .from an economic eficiency perspective, society might prefer a greater emphasis on customer-side reliability investments. The jeasibility zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA qf this approach is demonstrated using utility-supplied end-use data. zyxwvutsrqpo Keywords: Electricity; Reliability, End-use The target reliability level for most US utilities is quite high (one day of outages in 10 years). This standard, which was developed by engineers, has come under the scrutiny of economists during the past 15 years, as they seek to balance investments in electricity infrastructure with the needs of the rest of the economy. There has been a growing consensus in the economic literature that current reliability targets are above the social optimum Cl, 11,12,13,14]. However, recent experience suggests that electricity producers cannot reduce reliability levels without raising storms of protest among their customers. Following episodes of increased outages during recent peak periods, regu- lators and customers have suggested increasing the current reliability standard [lo]. The intense objec- tions to reduced reliability recommend re-evaluation and refinement of the cost-benefit analysis techniques that have been developed to date. This paper briefly reviews the theory for determining the socially optimal reliability level, and then suggests directions for improvement of the techniques outlined in the existing literature. These include a focus on customer-side reliability investments, on the supply- side relationship between system and subsystem reliability, and on the importance of disaggregation across customer classes, locations, time and types of reliability problems. Empirical evidence from a typical The author is with the Woodrow Wilson School, Princeton University, Princeton, NJ 08544-1013, USA. Final manuscript received 9 March 1992. US utility is used throughout, to illustrate the conceptual points I seek to make. Defining the socially optimal reliability level The basic definition of the socially optimal reliability level has not changed since Shipley, Patton and Denison first explicitly proposed it at the IEEE Power System Engineering Society Winter Meeting in 1972. In evaluating annual electric service availability, they stated that: ‘Presumably, the total electric power- related cost to the nation would be minimized for that value of service availability which minimizes the sum of: (i) utility system annual owning and operating costs, and (ii) the total annual cost of service interruptions’ [ 131. Supply-side, system investments in reliability must actually be applied to a complex series of inter- connected subsystems. The four main categories of subsystems are generation, transmission, substations and distribution. Investment projects typically apply to only one of these levels, and therefore represent subordinate contributions to the overall system optim- ization. There is wide agreement on the range of costs associated with supply-side reliability investments, based on a century of design, construction and operating experience with these technologies. The cost of service interruptions is much more difficult to assess; therefore it has been the primary focus of the economic literature on reliability planning. A variety of ‘production-lost’ and ‘willingness-to-pay’ approaches have been proposed by various authors 248 0140/9883/92/040248807 CC> 1992 Butterworth-Heinemann Ltd