Demand response user selection method based on perceptron

Abstract

The invention discloses a demand response user selection method based on a perceptron, and the method comprises the steps: protecting and obtaining historical feature data of each user, and carrying out the normalization processing; expanding the feature vector of each user to obtain a generalized feature vector, acquiring historical response conditions of the users, and forming a sample training set by the generalized feature vector and the historical response conditions; obtaining a separation hyperplane of each user; the given load regulation target of the system at the moment t is as follows: acquiring generalized feature vectors of all users at the moment t, and predicting response conditions according to the corresponding separation hyperplanes; selecting k users to send load adjusting instructions; and acquiring a historical corresponding condition of the user, judging whether the historical corresponding condition is consistent with a user response result predicted according to the perceptron model, and updating a separation hyperplane of the user. According to the invention, accurate power adjustment can be realized, and the potential of user resources at the demand side can be fully exerted.

Description

technical field [0001] The invention relates to the field of power system demand response, in particular to a perceptron-based demand response user selection method. Background technique [0002] With the widespread access of renewable energy sources such as photovoltaics and wind power in the power system, their inherent intermittent and randomness have brought great challenges to the safe operation of the power grid and seriously threatened the stability of the power grid. At the same time, for the purpose of carbon reduction, high-polluting power generation resources such as traditional coal-fired units are gradually being retired, which weakens the flexibility of grid adjustment and increases the demand for flexible and adjustable resources. In this context, user-side load resources have gradually been paid attention to, and demand response technology has emerged as the times require, which can be used to stabilize the fluctuation of renewable energy in the power grid, c...

AI Technical Summary

Problems solved by technology

[0003]However, in demand response projects, users are often affected by their own factors and the external environment, and have complex uncertainties. The specific performance is that users may choose Exit a demand response event

For load aggregators, if they cannot correctly evaluate the user's response behavior and blindly send adjustment instructions, the actual load adjustment will not meet the dispatching needs of the power system.

[0004]Currently, most studies on demand response focus on mechanisms, economics, physical models, etc. Research on demand response uncertainty is simply from the perspective of probability To model the adjustment capacity, there are few studies that consider the influence of the user's actual self and external factors on the user's uncertainty, and it is difficult to solve this problem essentially

Images