中压配电网线损计算新方法

长期以来，国内中压配电网线损计算困难，虽已做过大量研究，但仍未完全解决。文中针对这一难点以及遗传算法（GA）与人工神经网络（ANN）相结合的优点，提出了一种利用改进GA优化的ANN计算中压配电网线损的新方法。该方法首先利用负荷测试仪获得ANN模型所需的学习样本，然后利用改进的GA优化ANN的结构，最后应用优化的BP型ANN来拟合影响线损的特征参数与线损之间的复杂关系，从而建立了适合配电网线损计算的新模型。实例验证结果表明，该方法具有计算精度高、模型简单和经济实用等优点。

NEW METHOD FOR CALCULATING ENERGY LOSSES IN MEDIUM-VOLTAGE DISTRIBUTION SYSTEMS

For a long time, there have been lots of energy losses in medium voltage distribution networks in China, and it is very difficult to calculate them properly. To overcome this difficulty and taking the advantage of GA combined with ANN, a new method is proposed for calculating energy losses in medium voltage distribution systems based on ANN optimized by GA. The first step is to obtain learning samples of ANN by load survey instrument. The second step is to determine the structure of ANN by the improved GA. Then BP models are applied to map the complex relationship between energy losses and the related parameters. Based on the above steps, a model fit for the calculation of energy losses is established. Results show that it is a more accurate, simple, practical and economical method.