基于PSONN的大坝监控实时预报模型

将粒子群算法（PSO）引入大坝监测领域，提出一种基于粒子群神经网络（PSONN）的大坝监控预报模型。该模型充分发挥PSO的全局寻优能力和BP神经网络局部细致搜索优势，给BP神经网络提供了良好的初始权值。对逐一粒子群（SPSONN）、整体粒子群（WPSONN）、逐一BP（SBPNN）及整体BP（WBPNN）4种预报模型的对比分析表明：逐一预报模型（SPSONN和SBPNN）的预报精度明显高于对应的整体预报模型（WPSONN和WBPNN）的预报精度；与BP神经网络模型相比，PSONN模型不仅收敛速度明显加快，而且预报精度也有较大提高，尤其是SPSONN模型，其高精度和短历时性完全满足实时预报的需要，可以准确、有效地应用于大坝监测量的实时预报。

Real-time monitoring and prediction model for dams based on particle swarm optimization neural network

Particle Swarm Optimization（PSO）is introduced in the field of dam monitoring, and then a new dam monitoring model based on PSO Neural Network（PSONN）is put forward. The model makes full use of the global optimization of PSO and local accurate searching of BP Neural Network（BPNN）to provide better initial weights for BPNN. Seriatim PSONN（SPSONN）model, Whole PSONN（WPSONN）model, Seriatim BPNN（SBPNN）model and Whole BPNN（WBPNN）model are compared here, the results show that the seriatim prediction models are superior to those whole models with higher forecast precision. Furthermore, the PSONN models exhibit faster convergence rate, less iterative number and better forecast precision than the BPNN ones. Especially, the SPSONN model is more efficient to satisfy the demand of real-time prediction of dam monitoring.