원문정보
초록
영어
In this paper, we propose a hybrid optimization algorithm based on Improved Differential Evolution (IDE) algorithm and Gaussian Process (GP). Firstly, the paper constructs the assessment index system using Fault Tree Analysis (FTA) based on the summary and classification of the factors that could affect the power system security. Secondly, establish the risk assessment model of power system security based on the hybrid optimization GP algorithm. Hyper-parameter of GP has a great influence on construction of evaluation model, while conjugate gradient method which is usually used has strong dependence on initial values and is easy to fall into local optimal solution. So the paper uses the IDE algorithm for the traditional Hyper-parameter optimization, then the optimal Hyper-parameter is used to construct evaluation model for power grid security risk assessment. In the process of improvement, this paper adds the local search (Bees accelerated evolution operation) and global search (Bees scout operation) thought of ABC algorithm into the DE algorithm to reduce the population size required by the algorithm. After that, do the risk assessment of power system by using the established assessment model. Finally, do the simulation experiments using the standard data IEEE-39 and IEEE-118 bus example, and besides compare the IDE-GP with other optimization model like ABC-GP, DE-GP, MA-GP, GA-GP, and the experimental results show that hybrid optimization algorithm has better performance in accuracy while the time-consuming difference is minor. The validity of the proposed method is also demonstrated.
목차
1. Introduction
2. Establishment of Index System for Power System Security Risk Assessment
2.1. Security Risk Assessment Index System
2.2. Calculation of Security Risk Assessment Index
3. Power Grid Security Risk Assessment based on Hybrid Optimization Gaussian Process
3.1. Gaussian Process Regression
3.2. Hyper-Parameters Optimization of Gaussian Process based on IDE
3.3. Power System Security Risk Assessment based on Improved Gaussian Process Model
4. Experiment Results and Analysis
4.1. Experiment Analysis
4.2. Assessment Index
4.3. Assessment Result Analysis
5. Conclusions
References