Abstract—

Reliability of the solar power plant depends on its performance and economics factor compared to the conventional fueled power plants. In this paper, reliability performance assessment of grid connected roof top solar photovoltaic power plant (GCRTSPP) are presented at site location 12.0950° N, 75.5451° E) by considering various operating factors of subcomponents of solar photovoltaic panels, diode, capacitor and controller using Markov model approach. Critical stress factors are identified and are improved upon to enhance the system reliability. The performance of the solar photovoltaic array and the subcomponent of GCRTSPP have been examined with the sensitivity results. Also, sensitivity analysis of failure rate of these components with respect to stress factors is performed and critical stress factors are identified. Pareto analysis as a tool, reliability studies of grid connected solar photovoltaic system have been carried out to compute the highest failure rate of component. It is found that electronic controller and diode is more sensitive item compared to solar photovoltaic panels and capacitor. These components failure rate sensitivity analysis with respect to stress factors is performed. The reliability on standalone and grid connected photovoltaic system also compared. The critical stress factors of the components are identified. Artificial neural network (ANN) and machine learning programme (MLP) proposed for further study.

中文翻译：

---

基于马尔可夫模型的并网屋顶太阳能光伏电站可靠性评估

摘要—

与传统的燃料发电厂相比，太阳能发电厂的可靠性取决于其性能和经济因素。在本文中，通过考虑太阳能光伏板，二极管，电容器和使用控制器的子组件的各种操作因素，在站点位置12.0950°N，75.5451°E上提出了并网屋顶太阳能光伏电站（GCRTSPP）的可靠性性能评估。马尔可夫模型方法。确定并改善了关键应力因素，以提高系统可靠性。通过灵敏度结果检查了太阳能光伏阵列和GCRTSPP子组件的性能。另外，对这些组件的失效率相对于应力因子进行敏感性分析，并确定临界应力因子。帕累托分析作为工具，已经进行了并网太阳能光伏系统的可靠性研究，以计算组件的最高故障率。发现与太阳能光伏板和电容器相比，电子控制器和二极管更为敏感。进行了有关应力因素的这些组件故障率敏感性分析。还比较了独立和并网光伏系统的可靠性。确定了组件的临界应力因子。提出了人工神经网络（ANN）和机器学习程序（MLP）进行进一步研究。进行了有关应力因素的这些组件故障率敏感性分析。还比较了独立和并网光伏系统的可靠性。确定了组件的临界应力因子。提出了人工神经网络（ANN）和机器学习程序（MLP）进行进一步研究。进行了有关应力因素的这些组件故障率敏感性分析。还比较了独立和并网光伏系统的可靠性。确定了组件的临界应力因子。提出了人工神经网络（ANN）和机器学习程序（MLP）进行进一步研究。