基于一维卷积神经网络的电力系统暂态稳定在线评估

传统电力系统暂态稳定评估基于时域仿真计算,计算复杂度高,难以在线应用。提出一种基于一维卷积神经网络的电力系统暂态稳定在线评估,可极大提升暂态稳定在线评估速度。通过马尔可夫链蒙特卡洛抽样算法进行电力系统运行状态模拟,生成大规模运行数据。通过电力系统时域仿真计算确定发电机最大功角差。将电力系统运行数据作为一维卷积神经网络的输入,发电机最大功角差作为输出,训练一维卷积神经网络。在线应用场景下,一维卷积神经网络可基于当前运行数据快速计算发电机最大功角差,实现暂态稳定性在线评估。新英格兰39节点系统验证了所提在线评估算法的可行性。

Online Transient Stability Assessment of Power System Based on One-dimensional Convolutional Neural Network

Traditional transient stability assessment of power system is based on time domain simulation calculation, which has high computational complexity and is difficult to be applied online. An online transient stability assessment method based on one-dimensional convolutional neural network is proposed, which can greatly improve the speed of online transient stability assessment. Markov chain Monte Carlo sampling algorithm is used to simulate power system operation state and generate large-scale operation data. The maximum power angle difference of generator is determined by time domain simulation of power system. The operation data of power system is taken as the input of one-dimensional convolutional neural network, and the maximum power angle difference of generator is taken as the output to train the one-dimensional convolution neural network. In the online application scenario, one-dimensional convolutional neural network can quickly calculate the maximum power angle difference of generator based on the current operation data to realize online transient stability assessment. The New England 39 bus system verifies the feasibility of the proposed online evaluation algorithm.