Studies On Reliability Assessment Models Of Power Systems Considering The Additive Correlation For Load And Solar Radiation

Power system reliability evaluation is to evaluate the ability of meeting the load demand for various random failure events. Therefore, establishing accurate and effective probabilistic model for variety of random phenomena is the basic problem in power system reliability assessment. The study is carried out for probabilistic modeling method considering the randomness/spatial corrlation/additive correlation of bus loads and for probabilistic modeling method considering randomness/chronological correlation/additive correlation of solar radiation. And finally they are applied into the power system reliability evaluation.Existing probabilistic load models a re not always accurate, because they do not consider the additive correlation between system load and bus loads. The additive correlation means that system load equals to the sum of all bus loads. This paper proposes a nonparametric disaggregation load model for power system reliability evaluation. The proposed model does not need any assumption of the probability distribution for system load and bus loads, or the correlation form between bus loads. Considering the additive correlation between system load and bus loads, the proposed load model can successively disaggregate system load to individual bus loads, and provide more accurate simulation of load variation in power system reliability studies. The modified RBTS and IEEE-RTS79 system are analyzed to show that the proposed model is correct and effective.In present studies, the additive correlation of solar radiation can be established based on the fixed function between daily and hourly solar radiation, and cannot consider the effects of the randomness involved. A new two-step conditional kernel density estimation(CKDE) model of the solar radiation is proposed based on conditional kernel density estimation, where the first step consists of the CKDE of hourly solar radiation, and the second step consists of the CKDE of daily solar radiation CKDE. In addition to the additive correlation between daily and hourly solar radiation, the randomness and chronological correlation of the solar radiation can be represented. Due to the known product kernel expression of the CKDE having disadvantages, an improved CKDE is also developed through the detailed mathematical derivation using the full elements in the bandwidth matrix. The measured radiation data of photovoltaic generation in USA and the RBTS system are used to demonstrate the correctness/ effectiveness and applicability of the proposed probabilistic model, and analyze the impacts of the photovoltaic probabilistic model on the reliability indexes of the generation system.