Turkish Journal of Electrical
Power and Energy Systems
Research Article

Apply Bayesian Inference with Normal–Normal Conjugate to Forecast Renewable Energy Generation: A Case Study of Waste-to-Energy in Taiwan

1.

Department of Electrical Engineering, I-Shou University, Kaohsiung City, Taiwan, R.O.C.

Turkish Journal of Electrical Power and Energy Systems 2024; 4: 50-56
DOI: 10.5152/tepes.2024.24004
Read: 394 Downloads: 273 Published: 27 June 2024

This paper applies Bayesian inference with normal–normal conjugate to forecast renewable energy generation. The generation forecasts a probability distribution rather than a quantitative value. An assumed normal distribution is initialized for renewable energy generation. This assumed normal distribution’s parameters, the mean μ, and the standard deviation σ, are inferred by Bayesian inference afterward. However, applying Bayesian inference barely shall encounter an intractable integral. To circumvent the intractable integral, this paper considers the normal-normal conjugate method. This method fixes the assumed normal distribution’s σ and characterizes μ as another normal distribution and then infers the latter normal distribution parameters. A case study of waste-to-energy generation forecast in Taiwan is investigated in this paper. It has been found from the investigation that the Bayesian inferred probability distribution outperforms the assumed one.

Cite this article as: Y. Lin, “Apply Bayesian inference with normal–normal conjugate to forecast renewable energy generation: A case study of waste-toenergy in Taiwan”, Turk J Electr Power Energy Syst., 2024; 4(2), 50-56.

Files
EISSN 2791-6049