Improved air conditioner load aggregation method based on KNN algorithm and k-means algorithm

Abstract

The invention relates to an improved air conditioner load aggregation method based on a KNN algorithm and a k-means algorithm, and the method comprises the steps: 1) building an equivalent thermal parameter model of an air conditioner load, and generating a training data set; 2) preprocessing data in the training data set; 3) selecting classified key parameters in the air conditioner load equivalent thermal parameter model from the preprocessed data set; 4) using the classification key parameters as classification standards to train a KNN classification model, and determining a k value; (5) aggregation key parameters are selected in different air conditioner load categories, the determined k value serves as the k value of a k-means algorithm, aggregation is conducted through the k-means algorithm according to the key parameters, and aggregation power is obtained through an equivalent substitution method; and 6) selecting regulation and control strategies for different types of aggregated air-conditioning loads according to characteristics, and further obtaining a combined regulation and control strategy of the aggregated air-conditioning loads. Compared with the prior art, the method has the advantages of improving the aggregation precision of the air conditioner load, improving the response potential of the aggregated air conditioner load and the like.

Description

technical field [0001] The invention relates to the technical field of power system load modeling, in particular to an improved air-conditioning load aggregation method based on a KNN algorithm and a k-means algorithm. Background technique [0002] Using the controllable air-conditioning load to actively participate in the active power balance control of the power system, under the accurate controllable load power consumption model, the use of distributed control methods or distributed control methods for loads such as air-conditioning has become a hot spot for domestic and foreign electric power researchers. question. In the prior art, the air-conditioning load aggregation methods include primary aggregation, secondary aggregation and aggregation based on the Fokker-Planck equation. The primary aggregation is to group the air-conditioning loads in the same area into a group, and the secondary aggregation is for the control range. The air-conditioning load parameters are wi...

AI Technical Summary

Problems solved by technology

[0003] However, the typical aggregation models in one aggregation include state queue model, state space model, etc., which require the air conditioning load to have a certain similarity, which does not conform to the distribution characteristics of the actual power system air conditioning load

The secondary aggregation also uses a clustering algorithm to aggregate the air-conditioning load, but it is not accurate enough to distinguish the proportion of the air-conditioning load only through the weight setting

The aggregation method based on the Fokker-Planck equation is based on the first-order thermodynamic model, the calculation amount is large, the calculation process is complicated, and the solution of the equation is a difficult problem. So far, no analytical solution of the equation has been found.

To sum up, none of the existing air-conditioning load aggregation technologies takes into account the differences in air-conditioning loads. It is often assumed that the same type of air-conditioning loads are aggregated, which may cause large-scale loads to respond at the same time and cause resonance in the power system, affecting the safe operation of the power grid.

Images