Study On Some Contol Problems In Multiple Long Time-delay System For Alumina Continuous Carbonation Decomposition Process

Alumina continuous carbonation decomposition process (ACCDP) is composed of six series decomposers, in which the aeration amount of the dioxide carbon is controlled to ensure the process index of last resolution ratio and the quality of aluminum hydroxide. Because of large geographical span of the six decomposers, the distances between each decomposer control point and last resolution ratio check point are different, which causes the time between control signale output and last resolution ratio change reach from tens of minutes to several hours. So the delay times are long and multiple, it is difficult to realize production targets closed-loop control because the feedback information of last resolution ratio is not real-time. Multiple control points form muti-loops, in which correlative coupling is existed. When the steady state operating point is changed causing by ore source change or external interference, a long time is needed to adjust the process to a new state operating point. The production index fluctuation caused by unstable condition leads to poor product quality. Therefore, some researches on ACCDP decoupling control, delay time identification, closed-loop stable control strategy have an important and practical significance for improving product quality and industrial economic profits.On the basis of our research findings into the carbonation decomposition process and reaction mechanism, the multiple time-delays dynamic reaction model is established. The decoupling Smith control method is presented for multi-input-multi-output system with time delays. Multiple time-delays identification methods based on improved cross-correlation function and time-correlation analysis matrix are proposed. Time correspondence and parameters self-adjustment (TCPSA) control strategy for multiple long time-delay system is derived. ACCDP decentralized control strategy based on process index decomposition is researched. All these methods are applied to the ACCDP, and some good performances are achieved, the main content and some innovative achievements are depicted as follows:(1) The multiple time-delays dynamic reaction model of ACCDP is established. Based on the analysis for operational mechanism and production technology of ACCDP, dynamic differential equations derived from material balance principle are constructed according to the modeling method of continuous stirred-tank reactor. The model reflects ACCDP characteristics of association, multiple time-delays, and nonlinear.(2) The decoupling Smith control method is presented for multi-input-multi-output system with time delays often encountered in practical engineering. A new design method of decoupler based on the adjoint matrix of the multivariable system model with time delays is proposed. By analyzing the amplitude-frequency and phase-frequency characteristics, the models decoupled are reduced to first-order plus time delay models. According to the closed-loop characteristic equation of Smith predictor structure, PI controllers are obtained using the principle of pole assignment for Butterworth filter. At the same time, sufficient and necessary conditions for robust stability are analyzed with adaptive and multiplicative uncertainties which encountered frequently in practice. Simulation example results show the superiority of the proposed method. At last, the feasibility of applying the proposed method to ACCDP is analyzed by simulation in the condition of model mismatch.(3) The improved cross-correlation function method and time-correlation analysis matrix method are proposed for the problem of multiple time-delay parameters identification of ACCDP. For all the process variables effected by control signals, the reference variable is selected to be considered the correlation with the other variables respectively. For the considered variable, a set of data in a continuous time segment is selected as identification object, and the cross-correlation matrix of the data sets are calculated. By comparing the singular values of cross-correlation matrix, the delay corresponding to the maximun singular value is the required delay. On the other hand, from the perspective of multiple time-delay sequence, time-correlation analysis matrix of the data matrix related to different time-delay sequence is defined. The delay sequence corresponding to the maximun H infinity norm of time-correlation analysis matrix is the required value. The proposed methods are applied to identify the multiple time-delays of ACCDP using the field data. And the comparative analysis of the two methods is given to show the superiority of time-correlation analysis matrix method. At last, the ACCDP model parameters are calculated, and the accuracy of identified time-delays is verified.(4) Aiming at the problem of the stable closed-loop control for long time-delay system, TCPSA control strategy is proposed. The control effect of traditional time delay control methods for the system which delay time up to tens of minutes is discussed. The basic reason for control difficulties result from long time-delay is analyzed. On this basis, the pulse response equivalent system is introduced to coordinate the time corresponding relationship of control variable and output feedback variable. The parameters of pulse response equivalent system and PID controller are adjusted based on the deviation of pulse response equivalent system output and set value. Simulation results comparing the control effect of proposed method and other time delay methods show the superiority of TCPSA control strategy.(5) In view of the control problem of ACCDP multiple long time-delay system, decentralized control strategy based on process index decomposition is proposed. Decompose the last resolution ratio to optimal resolution ratio gradient of front five decomposers by using genetic algorithm to solve dynamic constrain optimization of aeration amount of the dioxide carbon. So the multiple long time-delay system control problem of ACCDP is decomposed to decentralized control problem of each decomposer. Using the TCPSA method for long time-delay system of each decomposer, simulation and applictioan results show the effectiveness of TCPSA method for ACCDP.