Contract electric quantity optimization decomposition method based on machine learning under new energy uncertainty

Abstract

The invention discloses a contract electric quantity optimization decomposition method based on machine learning under new energy uncertainty, and the method comprises the following steps: firstly considering uncertainty factors under new energy, and completing the extraction of uncertainty factors; considering uncertainty factor constraints under new energy, and establishing an uncertainty contract electric quantity optimization decomposition model; proposing a machine learning algorithm-based iterative solution optimization model to obtain a rolling correction contract electric quantity decomposition plan; and finally, analyzing and comparing according to an obtained result. The influence of uncertainty factors can be considered, the electric quantity decomposition result is more persuasive, the result can be more reasonable by using the Q learning method to solve, the influence of new energy uncertainty factors, electricity price and electric quantity fluctuation is considered, theelectric quantity decomposition result is more suitable for reality, and the machine learning algorithm can be used for solving the uncertainty problem more reasonably. The invention has more advantages in the aspect of ensuring economy and adaptability.

Description

technical field [0001] The invention relates to the technical field of electric power system electricity decomposition, in particular to a method for optimal decomposition of contract electricity based on machine learning under the uncertainty of new energy sources. Background technique [0002] Different from the mature economic growth models in foreign countries, medium and long-term transactions will still be the main transaction method in my country at present and in the future. On the one hand, this trading method can ensure the user's electricity demand, and on the other hand, the power generator can obtain the maximum profit. More importantly, the power generator can make a good power generation plan according to the market forecast, so as to maintain the stability of the market. For large electricity purchasers, the supply and demand of electricity exist in real time, which requires breaking down the medium and long-term contract electricity into smaller months, days...

AI Technical Summary

Problems solved by technology

The modern power system has the characteristics of high dimensionality, uncertainty, and time-varying nonlinearity, which makes it very difficult to solve such problems. Classical optimization analysis methods such as interior point method, gradient descent method, Newton method and alternating direction multiplier algorithm Although it has a fast solution speed and relies too much on the mathematical model, and when the model has characteristics such as nonlinearity, discrete variables, and multi-extreme values, the solution effect is not ideal, or even cannot be solved.

However, traditional intelligent algorithms such as particle swarm optimization, ant colony algorithm, and genetic algorithm, although relying less on mathematical models, are all based on the behavior of simple biological groups and do not have the ability to learn and remember. Optimizing Requirements for Scale Data Models

Images