This paper presents the impact of different explanatory variables such as remote control availability and conducted preventive maintenance, among others, on failure statistics of a disconnector population in Sweden using the proportional hazard model. To do so, 2191 work orders were analysed which included 1626 disconnectors and 278 major failures. Here, the results show that the remote control availability for disconnectors - an example of such Smart Grid technology - has a negative effect on the failure rate, whereas preventive maintenance has a positive impact. It is also shown that the disconnector age is not significant and that certain disconnector types have a significant and positive correlation towards failures when compared to other disconnector types. The results increase the understanding of disconnector failures to improve asset management.

This paper utilises the proportional hazard model to understand and quantify the impact of explanatory variables on the failure rate of circuit breakers (CB). Particularly, 4496 work orders with 2622 high voltage CBs are investigated with an occurrence of 281 major failures. Different explanatory variables such as CB type, manufacturer, preventive maintenance (PM), and others are gathered to quantify their significance and magnitude of their effect. The results present that PM has a positive impact, the number of operations within the last year a negative impact, and age has a small but negative impact on the failure rate. The CB type is not significant in all analyses which can be explained by examining the PM and age of these CB types. This paper contributes to the understanding of how explantatory variables impact the failure rate which is essential for power system asset management.

High voltage circuit breaker (CB) are of fundamental importance to protect and operate the power system. To improve their performance and to better predict failures, it is necessary to understand the effect of covariates such as preventive maintenance, age, voltage level, and the CB type. A straightforward approach is to investigate recurrent failures with regression models for count data. In this paper, several regression models are developed to estimate the impact of the aforementioned covariates to predict the recurrence of failures. The results show that age has a significant and negative impact, preventive maintenance before the first failure has a positive impact, and that the voltage level has a negative impact. Moreover, the Poisson, Negative Binomial, and zero-inflated models are compared. The comparison shows that the Negative Binomial model has the best fit to the studied recurrent failure data.

The failure rate is a reliability measure which isused for planning and operation of the power system. Thus far, average or experience based failure rates were applied to power system equipment due to their straightforward implementation. However, this approach limits the accuracy of the gained resultsand neglects the important differentiation between populationand individual failure rates. Hence, this paper discusses and demonstrates the necessity to distinguish between populationand individual failure rates and reviews the existing literature offailure rate estimation within the power system domain. The literature is categorized into statistical data driven approaches and failure rate modelling with focus on different criteria whichcan be used to describe the heterogeneity within populations. The review reveals that the environmental impact was modelled predominantly.

The high monetary value of a transformer has placed the transformer life-time optimization into the focus of asset management. The average failure rate has created reasonable results within reliability modeling, however, it cannot reflect the probability of failure for an individual transformer. In this paper, a method is introduced to calculate individual failure rates for a transformer population based on failure statistics and diagnostic measurements such as dissolved gas, and 2-furfuraldehyde analysis. The method is applicable to all types of components and the comprehensibility makes it effective for practical implementation. The results are evaluated against two health indices based on a weight factor and fuzzy logic. It can be observed that the presented individual failure rates are plausible representatives of the transformer's probability of failure. Therefore, the results can also be utilized for asset management decision-making.

The failure rate is essential in power system reliability assessment and thus far it has been commonly assumed as constant. This is a basic approach that delivers reasonable results. However, this approach neglects the heterogeneity in component populations which has a negative impact on the accuracy of the failure rate. This paper proposes a method based on risk functions, which describes the risk behaviour of condition measurements over time, to compute individual failure rates within populations. The method is applied to a population of 12 power transformers on transmission level. The computed individual failure rates depict the impact of maintenance and that power transformers with long operation times have a higher failure rate. Moreover, the paper presents a procedure based on the proposed approach to forecast failure rates. Finally, the individual failure rates are calculated over a specified prediction horizon and depicted with a 95\% confidence interval.