If the failure and repair performances obey the exponential distribution, then you can use the RBD reduction method of the following publication

https://www.researchgate.net/publication/307585524_Wind_Energy_Conversion_Systems_Reliability_Perspective?_sg=wBLopoqa0rbYlo238eE_b5bHMaKYSwNIKwfXj1_oW22Xfp3XlvJD-YdgIiT_qgERi-5O-iIF_x8vmZhskCleeebRyoWcfas0J_-PdfX2.VShLrfqrBVqC7EEjAequg8__8utVT_lOqZYq-3IW_7go7iF_bUJO_D91YyJCtxRNwOb8A__zJQ_rIs7tu67IxQ

Within its list of references, you will find more focused publications by R. Billinton.

Hi,

In this case, you can calculate the reliability of series-parallel combination of the system similar to calculating the equivalent capacitor of a series-parallel combination of the capacitors.

for parallel systems.

1/Req=1/R1+1/R2+ . . .

for series systems;

Req=R1+R2+ . . .

for a series-parallel system the above-mentioned relations can be useful.

This book can be useful.

Reliability Evaluation of Engineering Systems: Concepts and Techniques

Author(s): Roy Billinton, Ronald N. Allan (auth.)

Publisher: Springer US, Year: 1992

the Markov models are known models for your studies.

I suggest you read the following

1.Reliability Evaluation of Engineering Systems: Concepts and Techniques

2. Reliability Evaluation of Power Systems | Roy Billinton | Springer

It really depends on the size of the system and what reliability indices you require. See https://www.sciencedirect.com/science/article/pii/S0951832016300163 for a summary of available approaches, with their merits and demerits.

Thank you all..