Life cycle resilience quantification and enhancement of power distribution systems: A risk-based approach

Highlights

• Current standards consider a strength-based maintenance program for pole replacement.

• A risk-based replacement index called Expected Outage Reduction (EOR) is proposed.

• EOR ranks poles based on reduction in risk of outages when a pole is replaced.

• EOR and NESC strategies are applied to a power distribution system with over 7000 poles.

• The cumulative life cycle resilience over 70 years is improved by 22.3% using EOR.

Abstract

Distribution systems in the US are commonly supported by wood utility poles. Since wood poles may experience substantial decay rates, current standards specify a strength-based maintenance program for pole replacement regardless of the poles’ vulnerability and importance in the system. While state-of-the-art methods have developed risk-based metrics to guide system hardening decisions, such metrics are analyzed for the current conditions of the system. In this context, the potential of a stochastic series of hazards over extended horizons and the subsequent effects on the resilience of systems have been largely neglected. To address these limitations, a risk-based methodology is proposed for quantifying the life cycle resilience of power distribution systems. To project pole vulnerability, a recursive approach is developed that captures the stochasticity of precedent failures and subsequent corrective actions over extended horizons. Furthermore, a risk-based replacement index called Expected Outage Reduction (EOR) is introduced that estimates the expected power outage reduction if an existing pole is replaced by a new pole. The application of the proposed method for life cycle resilience analysis and management of a realistic distribution system subjected to stochastic hurricanes indicates that EOR can improve the cumulative life cycle resilience by up to 22.3% over 70 years.

Introduction

The overhead distribution infrastructure commonly consists of a large number of aged poles that have experienced several years of decay and degradation [4], [5], [6], [7], [28]. According to Darestani and Shafieezadeh [7], poles with over 30 years of age may experience a high rate of strength degradation. Considering the fact that wood poles are commonly replaced when their age is around 70 years [4], [26], in a common distribution line, there are many aged poles that are vulnerable to severe wind-related hazards. In order to ensure the safety of distribution systems, utility companies continually perform inspection and replacement. However, since there are a large number of poles in a distribution system, inspecting the entire distribution system is not feasible and pole replacement is constrained by the maintenance budget of utilities.

Maintenance and replacement of utility poles is regulated by the National Electric Safety Code [17], in over 90% of US states [2]. The NESC specifies a strength-based replacement strategy, in which a pole is replaced only when its strength falls below 67% of its initial strength. This maintenance strategy may not be adequate for protecting communities against outages because it does not properly account for the vulnerability of poles. It should be noted that strength is only one of the factors that affect the probability of failure of poles, and strength should be integrated with wind-induced demand models using reliability methods to estimate the vulnerability of poles. Moreover, this maintenance approach does not take into account the importance of the pole in the system and thus the consequence of pole failure in terms of, for example, the number of customers that lose power if the pole fails. A pole serving only a few customers might be more vulnerable than a pole that serves a large number of customers. However, if the consequence (risk) of failure is taken into account, replacing the latter pole could be more beneficial in terms of reducing the number of power outages in case a hurricane hits the distribution system. Risk-based strategies have been suggested for improving the performance of infrastructure systems [14], [31]. However, risk-based maintenance methods for power distribution systems aimed at improving the system resilience over extended periods have not been developed and implemented.

Resilience is an index that has been used to characterize the performance of infrastructure systems against severe hazards such as hurricanes, earthquakes, and tornadoes, among others. Resilience is the ability of a system to absorb shocks and to recover as fast as possible [3], [19], [20]. Considering the high rate of decay and degradation in distribution lines supported by wood poles and taking into account the fact that distribution systems are commonly designed for a service life of 70 years, there is a high demand to accurately estimate and enhance the resilience of such systems for long decision horizons.

A number of previous studies on risk, resilience, and life cycle cost assessment of distribution systems have investigated various hardening, inspection, strengthening and replacement strategies with the purpose of improving the performance of the system. For example, Onyewuchi et al. [18] developed a framework to prioritize county level pole inspection based on the total vulnerability of wood poles at each county. However, in this study, similar to the NESC standard, pole importance in the system is not considered. This vulnerability-based index for inspection prioritization may therefore not be an effective solution. Moreover, Salman et al. [26] investigated the effect of pole hardening on the life cycle cost of an imaginary distribution system. For this purpose, they used an index called Risk Achievement Worth (RAW) which takes into account the vulnerability and the consequence of failure for each line segment of a distribution system. However, RAW was calculated only once for each line, when all poles were assumed to be new. The realistic evolution of each pole’s risk with age was therefore not accounted for. Additionally, the effect of strengthening on an aged distribution system is not investigated in that study. In another study, Ma et al. [15] proposed a three-stage resilience enhancement optimization to minimize damage and the consequences of damage in a distribution system. The approach offers new capabilities in optimizing resilience, however the procedure assigns one decision variable to an entire distribution line segment rather than individual poles. It can be shown that with the same maintenance budget, targeting individual poles for replacement achieves greater resilience than replacing an entire distribution line segment. In addition, in this study, maintenance strategies and resilience assessment were only conducted for the current condition of the distribution system. However, as wood poles are usually expected to be in service for over 70 years, it is imperative to also estimate the future performance of the system to ensure that the chosen maintenance strategy can achieve and maintain high system resilience over extended periods. In another study, Ouyang and Dueñas-Osorio [20] proposed a three-stage resilience enhancement strategy for power networks, and explored effects of grid hardening, adding redundancy, and improving the recovery time on the annual resilience of a power distribution system. In addition, the study analyzed effects of multiple occurrences of hazards within a year on the annual resilience. However, such effects on the life cycle resilience of power distribution systems have not been investigated.

In all of the aforementioned studies where life cycle cost and resilience analysis of power distribution systems is extended into the future, the possibility of precedent failures are neglected. According to Fereshtehnejad and Shafieezadeh [10], [11], overlooking this effect could result in significant underestimation of the reliability and resilience of systems. To address this limitation, Ryan et al. [23], [24] performed service life reliability analysis of a set of wood poles in Brisbane, Australia. These studies applied a sequential Monte Carlo simulation method to account for failure scenarios due to multiple occurrences of hazards and annual replacements based on a strength-based replacement strategy. Although this approach can provide accurate estimates of the life cycle reliability of wood poles in distribution systems, if extended to analyze power network performance e.g., number of power outages or resilience, it may incur very high computational costs. The high costs stem from the fact that the approach will require network analysis for each individual Monte Carlo realization. Considering that for individual wood poles the number of required samples in the Monte Carlo simulation to capture their life cycle reliability is in the order of tens of thousands, resilience analysis of power distribution networks that include thousands of poles will become computationally very expensive using sequential Monte Carlo simulations.

Another often overlooked factor in resilience analysis of power distribution networks is the correlations in the failure events of overhead structures. In real distribution systems, considering that adjacent poles experience similar environmental conditions and hazard-induced stresses, the failure events of poles may be statistically correlated. Previous studies have explored the effect of correlation on the vulnerability of power systems by applying common cause failures [13], [25]. Some other studies investigated effects of correlation in vulnerability assessment of bridge networks and gas pipelines using Dichotomized Gaussian Model (DGM) [9], [22] and Matrix-based System Reliability (MSR) method [29], [30]. Nonetheless, the significance of these correlations for resilience estimation of power distribution systems is yet to be explored.

In order to address the aforementioned gaps, this study proposes a framework for life cycle resilience assessment of power distribution systems. The developed methodology accounts for (i) stochastic occurrences of hurricanes over extended time horizons, (ii) the subsequent compounding effects of aging and precedent failures for the future vulnerability of the grid, and (iii) effects of correlation in hurricane-induced failure of structures. Moreover, an annual risk-based replacement strategy is proposed for enhancing the life cycle resilience of power distribution systems. This risk-based strategy is centered on prioritizing pole replacement using a novel risk-based index called Expected Outage Reduction (EOR), which ranks poles based on the expected power outage reduction when a pole is replaced by a new pole. To evaluate this method, a realistic distribution system located in the southern US is considered in this study. This distribution system consists of a large number of decayed poles, which are in need of strengthening. Subsequently, the effectiveness of the proposed risk-based strengthening strategy is compared to existing methods for the service life of the distribution system, which is assumed to be 70 years. In addition, the significance of considering precedent failures and correlations in failure events of poles for the vulnerability and resilience estimates of power distribution systems are analyzed for the case study power network.