基于组件的复杂能源转换系统多模态结构化建模与性能分析

Multi-operation-mode switch is a typical characteristic for the complicated energy conversion systems under variable working conditions, but it can hardly be expressed effectively and standardly by general modeling theories and methods. In this project, from the system level, a complicated energy conversion system is divided into components, according to their regions or functions. Taking components as the basic units, through the components' information interaction and functional reconstruction, a structured mode perception model is established to solve the problems of mode representation and identification. A coordinated strategy for mode switch is developed to determine the hierarchical level and the response way to mode switch instructions for each component, and to make the components spontaneously form the ordered mode switch structure. Through the mode switch timing analysis, combined with the technical index of the system operation under variable conditions, the mode switch performance is further analyzed and evaluated. Multi-operation-mode characteristic analysis experimental system for the complicated energy conversion systems is researched and developed to verify by experiments the established multi-operation-mode model and its performance. The above research will provide an innovative theory and technical support for the analysis and optimal operation under all of the operating conditions of the complicated energy conversion systems. And it has important theoretical significance and application value for the process industry, represented by the electric power production, to improve its operation management level.

多模态转换是复杂能源转换系统变工况运行的典型特征，但常规的建模理论和方法难以对其进行有效、规范的表达。本项目从系统层面，将复杂能源转换系统按区域或功能划分为组件；以组件为基本单元，通过组件之间的信息交互及功能重构，建立结构化的模态感知模型，解决模态表示与模态辨识问题；研究模态转换的协同策略，确定各组件的层级地位和对切换指令的响应方式，使各组件自发地形成模态转换所需的有序结构；通过对模态转换过程的时序分析，结合系统变工况运行的技术指标，对模态转换性能进行分析和评价；研发复杂能源转换系统多模态特性分析实验系统，对所建多模态模型及其性能进行实验验证。上述研究将为复杂能源转换系统的全工况特性分析与优化运行提供重要的理论和技术支撑，对提高以电力生产为代表的流程工业运行管理水平具有重要的理论意义和应用价值。

多模态转换是复杂能源系统变工况运行的典型特征，但常规的建模理论和方法难以对其进行有效、规范的表达。本项目从系统层面开展了基于组件的复杂能源系统多模态结构化建模与性能分析，并在此基础上进行了控制与优化问题研究。建立了多种复杂能源系统模态感知模型，并将这些模型用于系统性能分析、控制器设计及能量调度：如对可再生能源出力及用户冷热电负荷进行精细化建模，包括基于模型树、相似日、支持向量机、多元自适应样条的多尺度预测模型等；研究了锅炉汽轮机单元协调控制系统、循环流化床锅炉燃烧系统的大范围变工况运行特性描述及多模态控制问题；考虑可再生能源不确定性对新能源消纳和电力系统稳定性的影响，研究了多种运行模式下提高系统运行性能的优化求解策略；针对复杂能源系统模态切换过程中信号传递和组件间交互行为的描述需求，提出了时序分析方法；在所建立的复杂能源系统模态感知模型的基础上，对模态转换过程的特性、转换策略等进行了研究，如针对马尔可夫跳变神经网络模型以及T-S模糊模型的研究等。上述研究可为复杂能源转换系统的全工况特性分析与优化运行提供重要的理论和技术支撑，对提高以电力生产为代表的流程工业运行管理水平具有重要的理论意义和应用价值。