Estimating reliability of a telecommunications energy network

Abstract

The paper presents a novel Telepower (−48V DC) reliability method, directly utilising site asset data in simulation software, to determine the effectiveness of battery replacement programs. Reliability is often estimated by a simple count of the number of outages per annum. This approach has many shortcomings because a count of outages can change dramatically even when the reliability of the physical network remains unchanged. Since infrastructure expenditure, such as battery lifecycle replacement programs only affects the physical network, basing that expenditure on a count of historical outages can lead to significant over or under-spending. This paper presents a way to estimate reliability that overcomes these difficulties. The paper will address how to estimate and track reliability without the attendant “noise” that accompanies more traditional methods. The “noise” includes external factors such as seasonality, singular environmental disturbances and changing standards, just to name a few. A new method will be demonstrated by an assessment of the reliability of approximately 25,000 48V DC power systems. The paper will show how the new method can effectively utilise “big data” inherent in large networks to arrive at a reliability estimate consistent with the physical network. The predicted change in reliability before and after a capital expenditure program will be demonstrated. The impact on Telepower reliability from changing battery reserves and rectifier redundancy will be demonstrated. This new method has wide application in solving a range of difficult cost-reliability problems.