Non-intrusive electric appliance load identification method based on hidden Markov chain

Abstract

The invention relates to a non-intrusive electric appliance load identification method based on a hidden Markov chain, and discloses a non-intrusive user load identification algorithm based on a hidden Markov model. The method comprises the steps of constructing the number of hidden states of the electric appliance load model, and counting the gear states of all electric appliances and the numberof the gear states of each electric appliance; constructing a model observation sequence value range, and calculating the maximum sum power of the electric appliance; determining a transition state matrix, and counting the transition probability between every two hidden states to form a transition matrix; counting the power consumption of the hidden state to obtain an emission matrix from the hidden state to the observation sequence; counting the occurrence probability of each hidden state in the power utilization process of the electric appliance to form an initial probability vector of eachhidden state; and obtaining the reading of the current electric meter as the current observation sequence, processing to obtain the optimal hidden state, and obtaining the current use condition of each electric appliance. The method is simple and effective, the electricity consumption condition of the user is obtained, too many sensors do not need to be deployed, high-frequency sampling is not needed either, early warning is conducted when the load is too high, and the electricity consumption risk is reduced.

Description

technical field [0001] The invention belongs to a power processing method in the technical field of non-invasive load identification, and relates to a non-invasive electrical load identification method based on a hidden Markov chain model. Background technique [0002] The grid load technology is a key component of the smart grid. It identifies the electrical load, and through data analysis, rationally allocates resources for peak shaving and valley filling to achieve the role of grid intelligence. At present, many smart home appliances can provide users with information such as power consumption of appliances, but these appliances do not have common data interfaces and standards. In addition, there are still a large number of home appliances that cannot provide data. Therefore, the research direction of load identification is to find an effective The algorithm identifies the user's current electricity consumption. There are two commonly used methods for current user load i...

AI Technical Summary

Problems solved by technology

[0003] Intrusive load identification, that is, using a large number of sensors at electrical sockets to measure the power of each electrical load, and then summarizing it to obtain the user's load status. This method is characterized by simplicity, efficiency and accurate results. The disadvantage is that a large number of sensors need to be deployed, and the investment is huge

[0004] Non-intrusive load identification, that is, it is not necessary to add sensors to each electrical appliance, but the load can be identified by reading the readings of the ammeter. This identification method does not need to purchase too many additional sensors, but the requirements for the algorithm Very high, the identification result is not accurate enough

[0005] Non-intrusive load identification generally adopts the load signature identification method, because each device will cause changes in the waveform when switching or shifting gears, and the electrical appliances can be identified by detecting the pattern of these changes, but the time of these changes is very short, so the sampling frequency is required High enough, which has certain requirements for the reading device on the user side

Images