复杂系统反向建模方法及偏最小二乘法建模应用研究

Research and Application of the Reversed Modeling Method and Partial Least-square Regression Modeling for the Complex Thermal System

针对常规建模方法在实际应用中所面临的建模难题,提出了一种新的建模思想--反向建模方法,即利用实际复杂热力系统运行时产生的大量实时数据,建立系统数学模型的学术思想。阐述了反向建模方法的思想、建模过程的模式。以烟气含氧量为例,利用华能福州电厂1号机所采集的大量现场运行数据,选用偏最小二乘算法(partial least-square regression,PLS),利用文中所提的反向建模思想尝试建立了一个局部系统的数学模型,验证结果初步表明该建模思想的有效性和实用性。文中所提方法可望提高复杂热力系统的建模效率和建模精度。

By using the real operation data of the complex thermal system, a new modeling method, called the reversed modeling method was presented, so that the system model can be established by the method. A step treatment of the reversed modeling method was introduced. As an example, based on the No.l unit real operation data of Huaneng Fuzhou Power Plant, the model procedures of 02 content in flue gas were listed. Based on partial least-square regression （PLS）, some models of O2 content in flue gas were given. Preliminary results show that the reversed modeling method is effective and can be used to improve modeling efficiency and accuracy.