锅炉-汽轮机系统非线性预测控制与多目标节能优化

The optimization control of the boiler-turbine system in the thermal power plant plays important role in improving the production efficiency and energy conservation. This project focuses on three key problems in the energy conservation optimization and high-accuracy control of the boiler-turbine system: (1) Because of the multivariable and nonlinearity of the boiler-turbine system, it is difficult to obtain the accurate mathmatic model. To solve this problem, an algorithm based on data analytics is designed to obtain the model based on production data. An adaptive kernel is constructed and it is optimized by improved differential evolution algorithm to increase the accuracy of the model. (2) Since the variety and complexity of the control targets, it is difficult to construct the optimization model. To solve this problem, a multi-objective nonlinear optimization model considering actual production constraints is constructed. In the optimization model, a mathmatic model of energy conservation targets and production variables is constructed. Then a tailed differential evolution algorithm is developed to solve the optimization model in order to obtain the optimized control variables.（3） Because of the complexity and nonlinearity of the production process, it can hardly control the process with high accuracy. To solve this problem, a nonlinear general predictive control algorithm is proposed based on the accurate prediction model. In the meantime, a dual-optimization strategy is designed to control with high accuracy online.

锅炉-汽轮机系统的优化控制是提高火力发电厂生产效率和节能减排的重要手段。本项目针对锅炉-汽轮机系统高精度控制和节能优化中存在的三个关键问题进行研究，主要包括：(1)针对系统多变量、非线性导致难以建立准确数学模型的问题，设计数据解析算法根据实际生产数据构建非线性自适应数据解析模型，构造自适应核函数并开发差分进化算法改进模型精度；(2)针对系统控制目标多样、目标之间关系复杂导致难以准确进行优化控制的问题，在考虑实际生产约束的基础上，建立节能减排指标与生产变量之间关系的数据解析模型，以最大化生产效益和环境效益为目标，构建多目标非线性优化模型，开发高效差分进化算法优化控制变量设定值，实现节能优化；(3)针对系统生产过程复杂、非线性过程模型难建导致难以高精度在线控制的问题，在建立非线性高精度预测模型的基础上，设计非线性广义预测控制算法，采用双重优化策略，实现系统的高精度在线控制。

为了深入了解锅炉-汽轮机系统的生产过程原理，提高系统的优化与控制水平，实现降低火电生产企业污染排放和提高企业生产效率的目标，本项目针对锅炉-汽轮机系统高精度控制和节能优化中存在的模型精度低、操作优化目标单一、控制精度低的问题进行了研究。主要内容包括：（1）开发了基于大数据的锅炉-汽轮机系统生产过程主要参数建模方法，有效解决了锅炉-汽轮机系统预测模型不准的技术难题，提取了各参数的相关模型输入信息，构建了适用于锅炉-汽轮机系统问题的浅层神经网络算法和深度置信网络算法；（2）开发了基于深度学习和智能优化算法的锅炉-汽轮机系统生产过程智能优化方法，有效解决了操作优化目标单一的问题，以提高生产效率、降低污染物排放、保障生产稳定为目标，构建了多目标非线性优化模型，开发了高效智能算法优化控制变量设定值，实现生产过程节能优化；（3）开发了适用于锅炉-汽轮机系统的非线性广义预测控制算法，有效解决了锅炉生产控制精度低的技术难题，建立了节能减排指标与生产变量之间关系的数据解析模型，设计了非线性广义预测控制算法，开发了双重优化策略，提高了系统控制精度，实现了系统的高精度在线控制。