Power system state estimation method of noise reduction self-coding and depth support vector machine

Abstract

The invention provides a power system state estimation method of noise reduction self-coding and depth support vector machine, which comprises the following steps: obtaining historical data of the measured value of voltage amplitude and the measured value of voltage phase angle in a power network node; obtaining the measured value of voltage amplitude and the measured value of voltage phase anglein a power network node. Performing Standardization; inputting the above into one-dimensional noise reduction self-encoding for noise reduction processing; inputting the data after noise reduction toa depth support vector machine for state estimation. Judging whether the error of state estimation meets the requirements; Outputting parameters of state estimation model based on 1-D noise reductionself-coding and depth support vector machine; Decomposing a large power network into p subnetworks; enabling A GPU to calculate a subnet, and performing summarization by the CPU and outputting astateestimation result of the whole network. The invention adopts the hybrid structure of GPU and CPU to estimate the state of the electric power system, which shortens the calculation time. The noise reduction self-coding and depth support vector machine model are used to improve the accuracy of state estimation, and the gradient descent method with variable learning rate is used in the training process to effectively find the optimal parameters and shorten the training time.

Description

technical field [0001] The invention belongs to the field of power system state estimation, and in particular relates to a power system state estimation method of noise reduction self-encoding and deep support vector machine. Background technique [0002] With the rapid development of the power system, the continuous upgrading of the grid structure and the continuous expansion of the scale, the operation mode of the grid is becoming increasingly complex. In order to be able to quickly deal with various problems that occur and ensure the safe and stable operation of the power system, this requires the power system dispatching center to be able to quickly and accurately grasp the operating status of the power system. In view of this, we must quickly and accurately estimate the state of the power system. [0003] In the 1970s, many scholars began to study the state estimation method of power system. The state estimation methods at that time were: fast decomposition method (P-...

AI Technical Summary

Problems solved by technology

SVM has also been applied in state estimation, and the estimation accuracy and calculation rate have been improved, but SVM also has certain disadvantages: first, the performance of the model depends on the kernel function selected a priori; second, it has a single-layer adjustable network parameters, the representational power of the model is limited

Images