In this paper, the modelling and multi-objective optimal control of batch processes, using a recurrent neuro-fuzzy network, are presented. The recurrent neuro-fuzzy network, forms a ＂global＂ nonlinear long-range pre...In this paper, the modelling and multi-objective optimal control of batch processes, using a recurrent neuro-fuzzy network, are presented. The recurrent neuro-fuzzy network, forms a ＂global＂ nonlinear long-range prediction model through the fuzzy conjunction of a number of ＂local＂ linear dynamic models. Network output is fed back to network input through one or more time delay units, which ensure that predictions from the recurrent neuro-fuzzy network are long-range. In building a recurrent neural network model, process knowledge is used initially to partition the processes non-linear characteristics into several local operating regions, and to aid in the initialisation of corresponding network weights. Process operational data is then used to train the network. Membership functions of the local regimes are identified, and local models are discovered via network training. Based on a recurrent neuro-fuzzy network model, a multi-objective optimal control policy can be obtained. The proposed technique is applied to a fed-batch reactor.展开更多
This paper presents several neural network based modelling, reliable optimal control, and iterative learning control methods for batch processes. In order to overcome the lack of robustness of a single neural network,...This paper presents several neural network based modelling, reliable optimal control, and iterative learning control methods for batch processes. In order to overcome the lack of robustness of a single neural network, bootstrap aggregated neural networks are used to build reliable data based empirical models. Apart from improving the model generalisation capability, a bootstrap aggregated neural network can also provide model prediction confidence bounds. A reliable optimal control method by incorporating model prediction confidence bounds into the optimisation objective function is presented. A neural network based iterative learning control strategy is presented to overcome the problem due to unknown disturbances and model-plant mismatches. The proposed methods are demonstrated on a simulated batch polymerisation process.展开更多
Batch process is a typical multi-phase process. Due to the interaction between the phases of the batch process, high precision control in a single phase cannot guarantee high precision control of the whole batch proce...Batch process is a typical multi-phase process. Due to the interaction between the phases of the batch process, high precision control in a single phase cannot guarantee high precision control of the whole batch process. In order to solve this problem, the guaranteed cost iterative learning control(ILC) of multi-phase batch processes is studied in this paper. Firstly, through introducing the output error, the state error and the extended information, the multi-phase batch process is transformed into an equivalent 2D switched system which has different dimensions. In addition, under the measurable condition, the guaranteed cost iterative learning control law with extended information is designed. The proposed control law ensures not only the stability of the system but also the optimal control performance. Next, in order to study the stability of the system and the minimum running time under the condition of stable running, the multi-Lyapunov function method is used. By means of the average dwell time method, the sufficient conditions ensuring system to be exponentially stable are given in the form of linear matrix inequality(LMI). Finally, the injection molding process is taken as an example to make simulation, which shows the feasibility and effectiveness of the proposed method.展开更多
Batch processes have been increasingly used in the production of low volume and high value added products. Consequently, optimization control in batch processes is crucial in order to derive the maximum benefit. In th...Batch processes have been increasingly used in the production of low volume and high value added products. Consequently, optimization control in batch processes is crucial in order to derive the maximum benefit. In this paper, a run-to-run product quality control based on iterative learning optimization control is developed. Moreover, a rigorous theorem is proposed and proven in this paper, which states that the tracking error under the optimal iterative learning control （ILC） law can converge to zero. In this paper, a typical nonlinear batch continuous stirred tank reactor （CSTR） is considered, and the results show that the performance of trajectory tracking is gradually improved by the ILC.展开更多
The batch dyeing process is a typical nonlinear process with time-delay,where precise controlling of temperature plays a vital role on the dyeing quality.Because the accuracy and robustness of the commonly used propor...The batch dyeing process is a typical nonlinear process with time-delay,where precise controlling of temperature plays a vital role on the dyeing quality.Because the accuracy and robustness of the commonly used proportion integration differentiation(PID) algorithm had been limited,a novel method was developed to precisely control the heating and cooling stages for batch dyeing process based on predictive sliding mode control(SMC) algorithm.Firstly,a special predictive sliding mode model was constructed according to the principle of generalized predictive control(GPC);secondly,an appropriate reference trajectory for SMC was designed based on the improved approaching law;finally,the predictive sliding mode model and the Diophantine equation were used to predict the output and then the optimized control law was derived using the generalized predictive law.This method combined GPC and the SMC with their respective advantages,so it could be applied to time-delay process,making the control system more robust.Simulation experiments show that this algorithm can well track the temperature variation for the batch dyeing process.展开更多
A three-step synthesis strategy for the automatic synthesis of operating procedures in batch processoperations is presented in this paper.The first step is pre-operation which consists of a set of checks that mustbe c...A three-step synthesis strategy for the automatic synthesis of operating procedures in batch processoperations is presented in this paper.The first step is pre-operation which consists of a set of checks that mustbe considered before starting the basic process operation.The second step is operation which is described by a setof basic control instructions.The third step is post-operation described by a set of instructions that must becarried out when the basic process operation is terminated.The three-step synthesis method has been imple-mented as part of a prototype recipe management system.展开更多
Control of the fed-batch ethanol fermentation processes to produce maximum product ethanol is one of the key issues in the bioreactor system.However,ethanol fermentation processes exhibit complex behavior and nonlinea...Control of the fed-batch ethanol fermentation processes to produce maximum product ethanol is one of the key issues in the bioreactor system.However,ethanol fermentation processes exhibit complex behavior and nonlinear dynamics with respect to the cell mass,substrate,feed-rate,etc.An improved dual heuristic programming algorithm based on the least squares temporal difference with gradient correction(LSTDC) algorithm(LSTDC-DHP) is proposed to solve the learning control problem of a fed-batch ethanol fermentation process.As a new algorithm of adaptive critic designs,LSTDC-DHP is used to realize online learning control of chemical dynamical plants,where LSTDC is commonly employed to approximate the value functions.Application of the LSTDC-DHP algorithm to ethanol fermentation process can realize efficient online learning control in continuous spaces.Simulation results demonstrate the effectiveness of LSTDC-DHP,and show that LSTDC-DHP can obtain the near-optimal feed rate trajectory faster than other-based algorithms.展开更多
The aim of this study was to establish a control system for saccharification process using quality control charts. To achieve this goal, temperature, pH and brix were measured at 12 minutes intervals for 15 consecutiv...The aim of this study was to establish a control system for saccharification process using quality control charts. To achieve this goal, temperature, pH and brix were measured at 12 minutes intervals for 15 consecutive batches which took 2 hours each. The time variations for three process parameters were assessed to establish a good understanding of the saccharification process. The temperature varied between 58℃ and 62℃ while the pH decreased slowly due to oxidation, values of which varied between 5.7 and 5.0. Brix values increased linearly with time. The initial and final values of the three parameters varied from one batch to another. Of the three parameters, brix was not well represented on the quality control charts due to wide difference between initial and final values during saccharification. The final brix values varied between batches, from 10.6% to 11.6%. The control charts used in this study were X-bar and Range charts. The rules for interpreting control charts were implemented for both X-bar and R charts, results of which showed that the process was out of control, although some rules were not violated due to little number of batches studied. The values of for temperature and pH data (2.27℃ and 0.35, respectively) were lower compared to brix data (11.2%). The corresponding values of span between control limits, SP<sub>x</sub> and SP<sub>R</sub> for temperature and pH were also comparatively lower than those established from brix data. Due to larger values of for brix measurements, the corresponding control charts for brix were insensitive in identifying out-of-control points during saccharification process.展开更多
An optimal iterative learning control （ILC） strategy of improving endpoint products in semi-batch processes is presented by combining a neural network model. Control affine feed-forward neural network （CAFNN） is p...An optimal iterative learning control （ILC） strategy of improving endpoint products in semi-batch processes is presented by combining a neural network model. Control affine feed-forward neural network （CAFNN） is proposed to build a model of semi-batch process. The main advantage of CAFNN is to obtain analytically its gradient of endpoint products with respect to input. Therefore, an optimal ILC law with direct error feedback is obtained explicitly, and the convergence of tracking error can be analyzed theoretically. It has been proved that the tracking errors may converge to small values. The proposed modeling and control strategy is illustrated on a simulated isothermal semi-batch reactor, and the results show that the endpoint products can be improved gradually from batch to batch.展开更多
基金This work was supported by the UK EPSRC (GR/N13319, GR/R10875).
文摘In this paper, the modelling and multi-objective optimal control of batch processes, using a recurrent neuro-fuzzy network, are presented. The recurrent neuro-fuzzy network, forms a ＂global＂ nonlinear long-range prediction model through the fuzzy conjunction of a number of ＂local＂ linear dynamic models. Network output is fed back to network input through one or more time delay units, which ensure that predictions from the recurrent neuro-fuzzy network are long-range. In building a recurrent neural network model, process knowledge is used initially to partition the processes non-linear characteristics into several local operating regions, and to aid in the initialisation of corresponding network weights. Process operational data is then used to train the network. Membership functions of the local regimes are identified, and local models are discovered via network training. Based on a recurrent neuro-fuzzy network model, a multi-objective optimal control policy can be obtained. The proposed technique is applied to a fed-batch reactor.
基金Supported by UK EPSRC (grants GR/N13319 and GR/R 10875)
文摘This paper presents several neural network based modelling, reliable optimal control, and iterative learning control methods for batch processes. In order to overcome the lack of robustness of a single neural network, bootstrap aggregated neural networks are used to build reliable data based empirical models. Apart from improving the model generalisation capability, a bootstrap aggregated neural network can also provide model prediction confidence bounds. A reliable optimal control method by incorporating model prediction confidence bounds into the optimisation objective function is presented. A neural network based iterative learning control strategy is presented to overcome the problem due to unknown disturbances and model-plant mismatches. The proposed methods are demonstrated on a simulated batch polymerisation process.
基金the National Natural Science Foundation of China(Nos.61773190 and 61433005)the Guangdong Innovative and Entrepreneurial Research Team Program(No.2013G076)
文摘Batch process is a typical multi-phase process. Due to the interaction between the phases of the batch process, high precision control in a single phase cannot guarantee high precision control of the whole batch process. In order to solve this problem, the guaranteed cost iterative learning control(ILC) of multi-phase batch processes is studied in this paper. Firstly, through introducing the output error, the state error and the extended information, the multi-phase batch process is transformed into an equivalent 2D switched system which has different dimensions. In addition, under the measurable condition, the guaranteed cost iterative learning control law with extended information is designed. The proposed control law ensures not only the stability of the system but also the optimal control performance. Next, in order to study the stability of the system and the minimum running time under the condition of stable running, the multi-Lyapunov function method is used. By means of the average dwell time method, the sufficient conditions ensuring system to be exponentially stable are given in the form of linear matrix inequality(LMI). Finally, the injection molding process is taken as an example to make simulation, which shows the feasibility and effectiveness of the proposed method.
基金supported by the Science Foundation of Shanghai Municipal Education Commission (Grant No.09Y208)the Science Foundation of Science and Technology Commission of Shanghai Municipality (Grant Nos.08DZ2272400, 09DZ2273400)the "11th Five-Year Plan" 211 Construction Project of Shanghai University
文摘Batch processes have been increasingly used in the production of low volume and high value added products. Consequently, optimization control in batch processes is crucial in order to derive the maximum benefit. In this paper, a run-to-run product quality control based on iterative learning optimization control is developed. Moreover, a rigorous theorem is proposed and proven in this paper, which states that the tracking error under the optimal iterative learning control （ILC） law can converge to zero. In this paper, a typical nonlinear batch continuous stirred tank reactor （CSTR） is considered, and the results show that the performance of trajectory tracking is gradually improved by the ILC.
基金National Natural Science Foundation of China(No.61074154)
文摘The batch dyeing process is a typical nonlinear process with time-delay,where precise controlling of temperature plays a vital role on the dyeing quality.Because the accuracy and robustness of the commonly used proportion integration differentiation(PID) algorithm had been limited,a novel method was developed to precisely control the heating and cooling stages for batch dyeing process based on predictive sliding mode control(SMC) algorithm.Firstly,a special predictive sliding mode model was constructed according to the principle of generalized predictive control(GPC);secondly,an appropriate reference trajectory for SMC was designed based on the improved approaching law;finally,the predictive sliding mode model and the Diophantine equation were used to predict the output and then the optimized control law was derived using the generalized predictive law.This method combined GPC and the SMC with their respective advantages,so it could be applied to time-delay process,making the control system more robust.Simulation experiments show that this algorithm can well track the temperature variation for the batch dyeing process.
文摘A three-step synthesis strategy for the automatic synthesis of operating procedures in batch processoperations is presented in this paper.The first step is pre-operation which consists of a set of checks that mustbe considered before starting the basic process operation.The second step is operation which is described by a setof basic control instructions.The third step is post-operation described by a set of instructions that must becarried out when the basic process operation is terminated.The three-step synthesis method has been imple-mented as part of a prototype recipe management system.
基金Supported by the National Natural Science Foundation of China(61573052)
文摘Control of the fed-batch ethanol fermentation processes to produce maximum product ethanol is one of the key issues in the bioreactor system.However,ethanol fermentation processes exhibit complex behavior and nonlinear dynamics with respect to the cell mass,substrate,feed-rate,etc.An improved dual heuristic programming algorithm based on the least squares temporal difference with gradient correction(LSTDC) algorithm(LSTDC-DHP) is proposed to solve the learning control problem of a fed-batch ethanol fermentation process.As a new algorithm of adaptive critic designs,LSTDC-DHP is used to realize online learning control of chemical dynamical plants,where LSTDC is commonly employed to approximate the value functions.Application of the LSTDC-DHP algorithm to ethanol fermentation process can realize efficient online learning control in continuous spaces.Simulation results demonstrate the effectiveness of LSTDC-DHP,and show that LSTDC-DHP can obtain the near-optimal feed rate trajectory faster than other-based algorithms.
文摘The aim of this study was to establish a control system for saccharification process using quality control charts. To achieve this goal, temperature, pH and brix were measured at 12 minutes intervals for 15 consecutive batches which took 2 hours each. The time variations for three process parameters were assessed to establish a good understanding of the saccharification process. The temperature varied between 58℃ and 62℃ while the pH decreased slowly due to oxidation, values of which varied between 5.7 and 5.0. Brix values increased linearly with time. The initial and final values of the three parameters varied from one batch to another. Of the three parameters, brix was not well represented on the quality control charts due to wide difference between initial and final values during saccharification. The final brix values varied between batches, from 10.6% to 11.6%. The control charts used in this study were X-bar and Range charts. The rules for interpreting control charts were implemented for both X-bar and R charts, results of which showed that the process was out of control, although some rules were not violated due to little number of batches studied. The values of for temperature and pH data (2.27℃ and 0.35, respectively) were lower compared to brix data (11.2%). The corresponding values of span between control limits, SP<sub>x</sub> and SP<sub>R</sub> for temperature and pH were also comparatively lower than those established from brix data. Due to larger values of for brix measurements, the corresponding control charts for brix were insensitive in identifying out-of-control points during saccharification process.
基金Supported by the National Natural Science Foundation of China (Grant Nos. 60404012, 60874049)the National High-Tech Research & Development Program of China (Grant No. 2007AA041402)+1 种基金the New Star of Science and Technology of Beijing City (Grant No. 2006A62)the IBM China Research Lab 2008 UR-Program
文摘An optimal iterative learning control （ILC） strategy of improving endpoint products in semi-batch processes is presented by combining a neural network model. Control affine feed-forward neural network （CAFNN） is proposed to build a model of semi-batch process. The main advantage of CAFNN is to obtain analytically its gradient of endpoint products with respect to input. Therefore, an optimal ILC law with direct error feedback is obtained explicitly, and the convergence of tracking error can be analyzed theoretically. It has been proved that the tracking errors may converge to small values. The proposed modeling and control strategy is illustrated on a simulated isothermal semi-batch reactor, and the results show that the endpoint products can be improved gradually from batch to batch.
