A critical evaluation of deterministic methods in size optimisation of reliable and cost effective standalone hybrid renewable energy systems

Reliability of a hybrid renewable energy system (HRES) strongly depends on various uncertainties affecting the amount of power produced by the system. In the design of systems subject to uncertainties, both deterministic and nondeterministic design approaches can be adopted. In a deterministic design approach, the designer considers the presence of uncertainties and incorporates them indirectly into the design by applying safety factors. It is assumed that, by employing suitable safety factors and considering worst-case-scenarios, reliable systems can be designed. In fact, the multi-objective optimisation problem with two objectives of reliability and cost is reduced to a single-objective optimisation problem with the objective of cost only. In this paper the competence of deterministic design methods in size optimisation of reliable standalone wind-PV-battery, windPV-diesel and wind-PV-battery-diesel configurations is examined. For each configuration, first, using different values of safety factors, the optimal size of the system components which minimises the system cost is found deterministically. Then, for each case, using a Monte Carlo simulation, the effect of safety factors on the reliability and the cost are investigated. In performing reliability analysis, several reliability measures, namely, unmet load, blackout durations (total, maximum and average) and mean time between failures are considered. It is shown that the traditional methods of considering the effect of uncertainties in deterministic designs such as design for an autonomy period and employing safety factors have either little or unpredictable impact on the actual reliability of the designed wind-PV-battery configuration. In the case of wind-PV-diesel and windPV-battery-diesel configurations it is shown that, while using a high-enough margin of safety in sizing diesel generator leads to reliable systems, the optimum value for this margin of safety leading to a cost-effective system cannot be quantified without employing probabilistic methods of analysis. It is also shown that deterministic cost analysis yields inaccurate results for all of the investigated configurations.