استفاده از اطلاعات اشتباه اندازه گیری شده جهت تخمین احتمالی ضعیف شدن ولتاژ

ABSTRACT Probabilistic estimation of voltage sags using erroneous measurement information

Voltage sags represent the most significant component of power quality (PQ) problems in distribution networks [1] both in terms of gross numbers of events and the high associated costs to end users [2–4]. An EU study estimated that voltage sags and short interruptions contributed to an annual cost of D 86bn [5] and in 2008/2009, voltage sags (and faults) caused an average of 0.73 interruptions per customer and contributed to an average of 76 customer minutes lost (CMLs) over the course of a one year period. These interruptions cause different impacts to different customers. For example, a momentary interruption for a large customer is estimated to cost £216k, whereas a 4 h interruption on a residential customer is estimated to cost only £4.78 [6]. Voltage sag performance estimation concerns both localizing the source of a voltage sag (most often a fault) and estimating the voltage sag magnitude in order to subsequently assess the impact on customers. DNOs are placing increasing emphasis on power quality monitoring to obtain greater visibility of power quality events, such as voltage sags, [7] beyond monitoring a limited group of large important customers. Through network wide monitoring, the level of power quality within a network can be quantified in terms of both events and costs and important current problems can be identified. Monitoring allows a DNO to perform reliability benchmarking, monitor power quality contracts and plan predictive maintenance [8]. Significant efforts in surveying power quality are also evidenced in the Benchmarking Report on the Quality of Electricity Supply of the Council of European Energy Regulators (CEER) [9]. A network operator’s monitoring investment decision is also driven by other factors such as new initiatives like the Smart Grid, changes in the regulatory environment, concerns about customer retention and new competition within the utility sector. The Electric Power Research Institute (EPRI) lists forecasting and short circuit analysis as the two main reasons for monitoring alongside permanent power quality monitoring [8] in future power networks. Finding the source of a voltage sag is closely related to the task of fault localization. This topic has been covered in existing research, recently in [10–14]. Fault localization in modern distribution networks is complicated in practice because information is only available from a limited number of variably accurate monitors. One of the most notable works which deals with measurement error is [13]. In [13] the authors developed a technique capable of locating faults using measurements taken from any two locations in the power network by utilizing an optimal estimation procedure based on the method of least squares. However, [13] does not consider how information from multiple monitoring devices could be combined to yield a distribution (rather than a point estimate) for the most likely fault location and does not consider situations where the number of monitors is fewer than required to obtain a single unique fault location estimate.