Dakuo He

Real-time product quality control for batch processes based on stacked least-squares support vector regression models

Abstract A novel real-time final product quality control method for batch operations based on stacked least-squares support vector regression models (stacked LSSVR) is proposed. It combines midcourse correction (MCC) and batch-to-batch control. To enhance the model prediction accuracy and generalization capability, a stacked LSSVR approach is presented. Quality control is achieved by predicting the final product quality using stacked LSSVR models and adjusting process variables at some pre-specified decision points. Then a decision is made on whether or not control action is taken at every decision point. Once the control action is expected, the manipulated variable values are calculated and the control action is taken to bring the off-spec product quality back to the target. Then a batch-to-batch control is used to overcome the model plant mismatches and unmeasured disturbances. At last, the proposed modeling and quality control strategy is illustrated on a simulated batch reactor.

Soft sensor for cobalt oxalate synthesis process in cobalt hydrometallurgy based on hybrid model

In the cobalt oxalate synthesis process in cobalt hydrometallurgy, the key end-product quality index, average particle size of cobalt oxalate, needs to be monitored and controlled. It is difficult to measure such particle size online by existing hardware sensors. Soft sensor technique has been widely used for estimating product quality or other important variables when online instruments and sensors are not available. In this paper, a hybrid modeling approach for cobalt oxalate synthesis process in cobalt hydrometallurgy is proposed by combining simplified first principle model with stacked LSSVR model. The former based on population balance equations and mass conservation equation with some assumptions is used for description and analysis of synthesis process in general; and the latter is developed to compensate the unmodeled characteristic and to enhance model generalization capability. Furthermore, a model output offset compensation strategy is also employed to increase the model prediction accuracy. Applications to a cobalt hydrometallurgy pilot plant demonstrate that the proposed approach is more precise and effective than the other conventional models.

Recursive parameter estimation for Hammerstein-Wiener systems using modified EKF algorithm

This paper focuses on the recursive parameter estimation for the single input single output Hammerstein-Wiener system model, and the study is then extended to a rarely mentioned multiple input single output Hammerstein-Wiener system. Inspired by the extended Kalman filter algorithm, two basic recursive algorithms are derived from the first and the second order Taylor approximation. Based on the form of the first order approximation algorithm, a modified algorithm with larger parameter convergence domain is proposed to cope with the problem of small parameter convergence domain of the first order one and the application limit of the second order one. The validity of the modification on the expansion of convergence domain is shown from the convergence analysis and is demonstrated with two simulation cases.

A comparative study on four modified artificial bee swarm algorithms

Artificial Colony Algorithm (ABC) is a random optimization algorithm based on swarm search behavior, which is widely used in recent years. Considering the slow convergence and ease of falling into the local optimum of basic ABC, researchers try various modified methods to overcome its shortcomings. Different modified ABC algorithms have different characteristics and application scope. In order to improve the performance of ABC, and provide some guidance for the future improved development and application of the algorithm, four modified ABC algorithms are studied and compared in this paper. These four modified algorithms include improved search equation artificial bee colony algorithm, artificial bee colony algorithm based on control parameters, artificial bee colony algorithm based on block perturbation strategy and artificial bee colony algorithm based on chaos mapping strategy. To show the performance and characteristics in all aspects, 13 standard test functions are selected for simulation experiments. And the performance of four selected algorithms and basic artificial bee colony algorithm are analyzed from the average value, the optimal value, the worst value, the variance and the convergence ability. This paper summarizes the advantages and disadvantages of each algorithm and its applicable scope, and provides some guidance for the improvement and application of ABC.

Modeling of a Fermentation Process with Multiple Support Vector Machines

A fermentation process suffers from a lack of good mathematical models, because of its complexity, since it is highly non-linear, has time varying parameters. It also has many important variables such as concentrations that cannot be measured on-line, which makes this set of variables of limited use for control purposes. However, these variables can be on-line estimated using soft sensing. This paper describes the identification of fermentation process with multiple least square support vector machine models. And the parameters in least square support vector machine were optimized by Genetic Arithmetic. The soft sensing model was trained on data coming form Nosiheptide fermentation process, and evaluated the production. The estimated results obtained were in good agreement with those found with the off-line laboratory analysis. It is also shown that software based state estimation was a powerful technique in biochemical products

Adaptive Differential Evolution Based on Simulated Annealing for Large-Scale Dynamic Economic Dispatch with Valve-Point Effects

Dynamic economic dispatch (DED) that considers valve-point effects is a complex nonconvex and nonsmooth optimization problem in power systems. Over the past few decades, multiple approaches have been developed to solve this problem. In this paper, an adaptive differential evolution based on simulated annealing algorithm is proposed to solve the DED problem with valve-point effects. Simulated annealing (SA) algorithm is employed to carry out an adaptive selection mechanism in which the mutation operators of differential evolution (DE) are selected adaptively based on their historical performance. A mutation operator pool consisting of five operators is built to make each operator show its strength at different stages of the evolutionary process. Moreover, a heuristic strategy is introduced to transform infeasible solutions towards feasible ones to enhance the convergence rate of the proposed algorithm. The effectiveness of the proposed methods is demonstrated first on 10 popular benchmark functions with 100 dimensions, in comparison with the classic DE and five variants. Then, it is used to solve four DED problems with 10, 15, 30, and 54 units, which consider the valve-point effects, transmission loss, and prohibited operating zones. The simulation results are compared with those of state-of-the-art algorithms to clarify the significance of the proposed method and verify its performance. Three systems with 100-500 generators are also tested to confirm the advantages of the proposed method on large-scale DED problem.

A knowledge based intelligent control method for dehydration and mixing process

In the production process of hydrometallurgy, the concentration and flow rate of the slurry in the dehydration and mixing process have an important influence on the leaching process. However, due to the lack of online hardware analyzers in the dehydration and mixing process, it is difficult to realize the automatic control of the concentration and flow rate of the slurry and the artificial control method still widely used. Thus, it has become the bottleneck of the whole production process of hydrometallurgy. Therefore, an intelligent control method based on the knowledge was proposed to control the concentration and flow rate of the slurry in the dehydration and mixing process. The knowledge was extracted from the industrial data, process mechanism and expert experience. In addition, this method is successfully applied to a certain dehydration and mixing process in a hydrometallurgical plant, and achieved a satisfactory control effect.

Particle Swarm Optimization based on clustering in searching process

The population distribution of Particle Swarm Optimization (PSO) directly concerns global convergence and searching efficiency of PSO. The reasonable setting of population distribution and operational parameters is an important problems in the application of PSO to perform optimization calculation. Based on the study on how to set the population distribution, such conclusion can be drawn that the population distribution must reflect the information on solution space scientifically. The PSO based on the population distribution of clustering is proposed. The population distribution was analyzed according to the discrepancy in the solution space and objective function space. The integrated clustering index, which combines the fitness value and space location, was applied to design the population distribution. Simulation results show that the method is feasible and effective.

Chaotic differential evolution algorithm based on simplex method for large-scale industrial processes of fuzzy model

In the large-scale industrial processes, there are more slow perturbations. So the mathematical model of an actual system is difficult to be accurate. When optimizing large-scale industrial processes, the mathematical model and the actual system does not match, that is model-reality difference. In order to deal with this problem, the structure of decomposition and coordination is used in this paper. The whole large-scale industrial processes can be decomposed into several sub processes that are interactive, and the simplex method including open-loop simplex method and the simplex method with global feedback is the coordinate strategy. The main idea of the differential evolution algorithm with simple method is that find out the best individual in every generation with the standard differential evolution algorithm method, then around the best individual do local search for some times and compared the former best one with the ones from the local search, if the latter is better, substitute the best one with the latter. A classical example of large-scale industrial processes is applied and the simulation results show the validity of the method.

An improved multi-objective differential evolution algorithm

Based on multi-objective differential evolution algorithm, adaptive chaotic multi-objective differential evolution algorithm (AC-DEMO) is proposed, combining with adaptive and chaotic principles. In AC-DEMO, chaotic initialization and adaptive mutation operator are introduced to improve the efficiency of the algorithm. Numerical experiment results of commonly used test functions show that the algorithm has a good approximation and uniformity index and is suitable to solve complex multi-objective optimization problems.