Multi-model modeling is an important method for solving the modeling problem of complex non-linear systems. However, there is no such good classification principle of its sub-models' time span, and the division of sub-windows affects the accuracy of the model, as well as the cost of computing. Through analysis of the characteristics of the bed temperature object, the principal component analysis was introduced into the method of designing the time span division of the sub-windows. The dividing points of the time span of the sub-windows were determined by piecewise analysis, rolling merging, and cyclic validation. The problem with this is that the sub-models in different windows may be different. Aiming at the problems of modeling in the sub-windows, the contribution rates of variables to the principal components were analyzed, the independent variables that play a major role in the dependent variables in the sub-windows were determined with the results of PCA, and the regression models in the the sub-windows were identified by multivariate linear regression. Compared to the sub-window model established by the principal component regression method, the former method had higher accuracy and could better reflect the actual operation of the bed temperature object.

The accurate measurement of NOx concentration is the basis of accurate ammonia injection in selective catalytic reduction (SCR) system of thermal power unit. Excessive or too little ammonia injection will cause ammonia escape or exceeding the standard of NOx emission, and cause environmental pollution. There exist a large measurement lag and errors during the purging process in the measurement of NOx concentration in the SCR system of thermal power units. Therefore, to enable the intelligent control of the SCR system, it is necessary to carry out soft sensing of NOx emissions. This paper proposes a soft sensing algorithm for NOx emissions of thermal power units based on an adaptive Gaussian Mixture Model (GMM) two-step clustering algorithm and ensemble learning. Adaptive GMM two-step clustering (AGTSC) algorithm is used to softly divide the historical data into working conditions. Temporal Pattern Attention Long Short-Term Memory (TPA-LSTM) is also adopted to construct individual learners for each working condition. Here we use the parameter regression algorithm to train the combiner based on the output of the individual learner and the membership signal of the working condition as the input of the combiner. The clustering algorithm is tested with three datasets. The results show that this algorithm addresses the shortcomings of GMM including not being applicable to the clustering of non-convex datasets and manual selection of the number of clusters. The soft sensing algorithm is verified by using the historical data of the SCR system of 1000 MW thermal power units. The results show that the proposed method has higher accuracy and stronger generalization ability than the conventional method, hence providing an effective method for soft sensing of NOx emissions of thermal power units.

In view of the features of superheated steam temperature system in the power plant, with the models of controlled object under different circumstances, a multiple models modeling method based on mixed logical dynamical (MLD) model is given, and the GPC for the MLD is put forward, its stability is also discussed. The control algorithm is simple and more adaptable for the change of circumstance, also can avoid the disturbance of controllers' switch while using the common method. Considering the main steam temperature plant of a supercritical once-through 600MW boiler, the simulation is done to validate the control strategy. The result show the method has a strong robustness and better tracking performance, the satisfying control of the total operating conditions is also realized.

Under the conditions of high demand for energy saving and environmental protection, the thermal power unit is required to phase out the traditional extensive operation mode—a method of oxygen-enriched combustion in a furnace, considering safety first. Achieving efficient and economic operation with an optimal proportion of air distribution in these thermal power units is crucial. The high-precision simulation equipment could provide an experimental basis for optimal operation of field units. This paper starts by improving the accuracy of simulation equipment. In this work, the method of dividing nodes and branches in the boiler was based on signal flow graph theory. According to the flow characteristics of the working substance, the method for calculating the node and branch pressure drop was analyzed and set up. Subsequently, a fluid network model of the multi-dimensional flue gas system was constructed. With the help of our self-developed simulation model and data-driven platform, a modular simulation algorithm was designed. The simulation analysis of the boiler showed the accuracy of the model.