基于CNGWO-LSSVM的汽轮机热耗率预测模型

为了准确计算汽轮机热耗率，提出一种改进灰狼优化算法优化最小二乘支持向量机（LSSVM）的热耗率软测量方法。首先针对灰狼算法收敛精度低的缺点提出一种混沌非线性灰狼优化算法（CNGWO），通过Kent混沌搜索策略和非线性动态递减权值策略来改善灰狼优化算法的性能。然后利用CNGWO算法预先选择LSSVM模型参数，并建立CNGWO-LSSVM的软测量模型。以某600 MW超临界汽轮机组实时运行数据仿真实验，对具有复杂非线性的热耗率变量进行预测，预测结果表明，经过CNGWO算法优化的LSSVM模型取得了较好的预测效果，为汽轮机热耗率的精确计算提供了一种有效方法。

Prediction Model of Steam Turbine Heat Rate Based on CNGWO-LSSVM

In order to accurately calculate the heat rate of steam turbines, an improved grey wolf optimization (GWO) algorithm is proposed to optimize the soft-sensing modeling method using least squares support vector machine (LSSVM). Firstly, aiming at the issue of low convergence precision of GWO algorithm, the chaotic nonlinear grey wolf optimization algorithm (CNGWO) is proposed to improve the performance of GWO algorithm by virtue of Kent chaotic search strategy and nonlinear dynamic decline strategy. Then the model parameters of LSSVM is optimized by chaotic nonlinear grey wolf algorithm, and the soft sensor model of CNGWO-LSSVM is established. Based on the real-time operation data of a 600 MW supercritical steam turbine unit in a power plant, the soft sensing technology based on CNGWO-LSSVM is applied to predicate the heat rate of the 600 MW supercritical steam turbine unit. The prediction result indicates that the LSSVM optimized by CNGWO algorithm has achieved satisfactory prediction effect, which can provide an effective way for accurate heat rate calculation.