LS-SVM model based nonlinear predictive control for MCFC system

This paper describes a nonlinear model predictive controller for regulating a molten carbonate fuel cell (MCFC). In order to improve MCFC’s generating performance, prolong its life and guarantee safety, it must be controlled efficiently. First, the output voltage of an MCFC stack is identified by a least squares support vector machine (LS-SVM) method with radial basis function (RBF) kernel so as to implement nonlinear predictive control. And then, the optimal control sequences are obtained by applying genetic algorithm (GA). The model and controller have been realized in the MATLAB environment. Simulation results indicated that the proposed controller exhibits satisfying control effect.     

Integrated statistical and engineering process control based on smooth transition autoregressive model

Traditional studies on integrated statistical process control and engineering process control (SPC-EPC) are based on linear autoregressive integrated moving average (ARIMA) time series models to describe the dynamic noise of the system.However,linear models sometimes are unable to model complex nonlinear autocorrelation.To solve this problem,this paper presents an integrated SPC-EPC method based on smooth transition autoregressive (STAR) time series model,and builds a minimum mean squared error (MMSE) controller as well as an integrated SPC-EPC control system.The performance of this method for checking the trend and sustained shift is analyzed.The simulation results indicate that this integrated SPC-EPC control method based on STAR model is effective in controlling complex nonlinear systems.     

An iterative computation method for interpreting and extending an analytical battery model

Battery models are of great importance to develop portable computing systems, for whether the design of low power hardware architecture or the design of battery-aware scheduling policies. In this paper, we present a physically justified iterative computing method to illustrate the discharge, recovery and charge process of Li/Li-ion batteries. The discharge and recovery processes correspond well to an existing accurate analytical battery model： R-V-W＇s analytical model, and thus interpret this model algorithmically. Our method can also extend R-V-W＇s model easily to accommodate the charge process, The work will help the system designers to grasp the characteristics of R-V-W＇s battery model and also, enable to predict the battery behavior in the charge process in a uniform way as the discharge process and the recovery process. Experiments are performed to show the accuracy of the extended model by comparing the predicted charge times with those derived from the DUALFOIL simulations. Various profiles with different combinations of battery modes were tested, The experimental results show that the extended battery model preserves high accuracy in predicting the charge behavior.     

SVM for density estimation and application to medical image segmentation

A method of medical image segmentation based on support vector machine （SVM） for density estimation is presented. We used this estimator to construct a prior model of the image intensity and curvature profile of the structure from training images. When segmenting a novel image similar to the training images, the technique of narrow level set method is used. The higher dimensional surface evolution metric is defined by the prior model instead of by energy minimization function. This method offers several advantages. First, SVM for density estimation is consistent and its solution is sparse. Second, compared to the traditional level set methods, this method incorporates shape information on the object to be segmented into the segmentation process. Segmentation results are demonstrated on synthetic images, MR images and ultrasonic images.     

Optimal design of Stewart platforms based on expanding the control bandwidth while considering the hydraulic system design

We proposed an optimal design method to expand the bandwidth for the control of large hydraulic Stewart platform. The method is based on generalized natural frequency and takes hydraulic oil into consideration. A Lagrangian formulation which considers the whole leg inertia is presented to obtain the accurate equivalent mass matrix. Using the model, the effect of leg inertia and the influence of design parameters on the generalized natural frequency are investigated. Finally, numerical examples are presented to validate and confirm the efficiency of the mathematical model. The results show that the leg inertia, especially the piston part plays an important role in the dynamics. The optimum diameter ratio of the base to the moving platform is between 2 and 3, and the optimum joint angle ratio of the base to the moving platform is about 1. The smaller joint angles and a longer leg stroke are favorable for raising system frequencies. The system oil should be preprocessed for large platforms with a requirement for good dynamic performance.     

A minimax optimal control strategy for uncertain quasi-Hamiltonian systems

A minimax optimal control strategy for quasi-Hamiltonian systems with bounded parametric and/or external disturbances is proposed based on the stochastic averaging method and stochastic differential game. To conduct the system energy control, the partially averaged Ito stochastic differential equations for the energy processes are first derived by using the stochastic averaging method for quasi-Hamiltonian systems. Combining the above equations with an appropriate performance index, the proposed strategy is searching for an optimal worst-case controller by solving a stochastic differential game problem. The worst-case disturbances and the optimal controls are obtained by solving a Hamilton-Jacobi-Isaacs （HJI） equation. Numerical results for a controlled and stochastically excited DulTlng oscillator with uncertain disturbances exhibit the efficacy of the proposed control strategy.     

No-cooperative games for multiple emergency locations in resource scheduling

When an emergency happens, the scheduling of relief resources to multiple emergency locations is a realistic and intricate problem, especially when the available resources are limited. A non-cooperative games model and an algorithm for scheduling of relief resources are presented. In the model, the players correspond to the multiple emergency locations, strategies correspond to all resources scheduling and the payoff of each emergency location corresponds to the reciprocal of its scheduling cost. Thus, the optimal results are determined by the Nash equilibrium point of this game. Then the iterative algorithm is introduced to seek the Nash equilibrium point. Simulation and analysis are given to demonstrate the feasibility and availability of the model.     

How a learner self-regulates reading comprehension： A case study for graduate level reading

The purpose of this study is to examine how a learner self-regulates learning while reading an academic text. In particular, the aim is not to generalize self-regulatory processes for any learning task, but to have an overall idea about how a learner self-regulates. In particular, Pintrich＇s SRL （self-regulated learning） model is used to find out whether the model was apparent in the learner＇s reading comprehension process. In this model, self-regulatory processes are categorized into 4 phases （forethought, monitoring, control and reflection）, and each phase is divided into 4 areas of self-regulation （cognitive, motivational, behavioral and contextual）. The data were collected through observation, videotaping and semi-structured interview. Purposeful sampling was used to obtain an in-depth understanding about how an experienced learner self-regulates and uses different kinds of strategies while reading an academic text. The result of the study revealed that all the phases in the model were apparent in the participant＇s reading comprehension task. However, it was difficult to decide on which strategies were belonging to monitoring or control phases. Actually, much of the empirical work also does not find much separation on these phases. This might be because these phases are reflecting the learner＇s thinking process.     

Nonlinear multifunctional sensor signal reconstruction based on least squares support vector machines and total least squares algorithm

Least squares support vector machines （LS-SVMs） are modified support vector machines （SVMs） that involve equality constraints and work with a least squares cost function, which simplifies the optimization procedure. In this paper, a novel training algorithm based on total least squares （TLS） for an LS-SVM is presented and applied to multifunctional sensor signal reconstruction. For three different nonlinearities of a multifunctional sensor model, the reconstruction accuracies of input signals are 0.00136%, 0.03184% and 0.504 80%, respectively. The experimental results demonstrate the higher reliability and accuracy of the proposed method for multifunctional sensor signal reconstruction than the original LS-SVM training algorithm, and verify the feasibility and stability of the proposed method.     

Shortcut Algorithm for Simulation of Batch Extractive Distillation

The batch extractive distillation (BED) process has the advantages of both batch and extractive distillation. It is one of the most promising means for the separation of azeotropic and close-boiling point systems. However, so far this process has not been applied in industry due to its over-complexity. A new shortcut model was proposed to simulate the operation of the batch extractive distillation operations. This algorithm is based on the assumption that the batch extractive distillation column can be considered as a continuous extractive distillation column with changing feed at any time. Namely, the whole batch process is simulated as a succession of a finite number of steady states of short duration, in which holdup is considered as constant mole. For each period of time the batch extractive distillation process is solved through the algorithm for continuous extractive distillation. Finally, the practical implementation of the shortcut model is discussed and data from the laboratory and literature are presented. It is found that this model has better adaptability, more satisfactory accuracy and less calculative load than previous rigorous model. Hence the algorithm for simulating BED is verified.     

Run-to-run product quality control of batch processes

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. Project supported by the Science Foundation of Shanghai Municipal Education Commission (Grant No.09Y208), and the Science Foundation of Science and Technology Commission of Shanghai Municipality (Grant Nos.08DZ2272400, 09DZ2273400) and the “11th Five-Year Plan” 211 Construction Project of Shanghai University     

IMPROVING OUTCOMES IN MATHEMATICS IN NEW ZEALAND: A DYNAMIC APPROACH TO THE POLICY PROCESS

Intervention at scale with the aim of improving student participation, engagement and outcomes in mathematics education is a challenge for educational policy makers and reformers. This article argues that an iterative annual cycle of policy formulation, implementation and evaluation enabled ongoing adjustments to the strategic focus, the professional development model and the system infrastructure as the New Zealand Numeracy Development Project was taken to scale. The analysis draws on the project’s evaluation data over a 6-year period to demonstrate how adjustments were made over time to the pedagogical tools and to the professional development processes. The ongoing development of knowledge supported the management of strategic risks in taking the project to scale: the ongoing appropriation of adequate levels of resourcing to support the school-based professional development model and the availability of system-wide expertise for effective implementation. The analysis suggests that conceptualising implementation as an interdependent and interrelated component of an iterative policy process and as an opportunity for knowledge building ensured a continuing focus on student outcomes. The dynamic approach to the policy process appeared central to building this intervention’s effectiveness and feasibility at scale.     

基于遗传支持向量机的生物质发电气化研究

提出了一种基于遗传算法和支持向量机的生物质气化过程研究的新方法。采用竹子气化数据建立GA-LS-SVM模型,并验证该模型方法在生物质气化过程建模中的适用性。GA-LS-SVM模型预测方法精度较高,效果也比较理想,是一种可行有效的建模方法。     

Predictive control of a class of bilinear systems based on global off-line models

INTRODUCTION Bilinear systems are a kind of important nonlinear systems with relatively simple structure, and many industrial processes can be described as a bilinear system. Thus research on the control of this kind of systems is very important. On the other hand, model predictive control (MPC) (Clarke et al., 1987) has been widely used in industrial applications and many predictive control methods focusing on bilinear systems are emerging (Bloemen et al., 2001; Fontes et al., 2004; He…     

Multiobjective optimization and multivariable control of the beer fermentation process with the use of evolutionary algorithms

This paper describes empirical research on the model, optimization and supervisory control of beer fermentation. Conditions in the laboratory were made as similar as possible to brewery industry conditions. Since mathematical models that consider realistic industrial conditions were not available, a new mathematical model design involving industrial conditions was first developed. Batch fermentations are multiobjective dynamic processes that must be guided along optimal paths to obtain good results. The paper describes a direct way to apply a Pareto set approach with multiobjective evolutionary algorithms (MOEAs). Successful finding of optimal ways to drive these processes were reported. Once obtained, the mathematical fermentation model was used to optimize the fermentation process by using an intelligent control based on certain rules.     

Multiobjective optimization and multivariable control of the beer fermentation process with the use of evolutionary algorithms

INTRODUCTIONThefermentationprocesshasgreatscientificandpracticalimportance.Forseveralreasons,qua-litycontrol,forexample,itisimportanttoincludeallkindsoftechnologicaladvancesintheindustrialprocesses.Thisisthemainaimofourresearch,whichisverycomplexduetothenon-linearityofthestudiedphenomena.Themathematicalmodelsfoundinliteratureincludeonlylaboratoyconditions.Amathematicalmodelhasbeendevelopedforthebeerfermentationprocessbecauseoftheadvantagesthatitprovides.Itallowssimulatingwhatcanhappenwhe…     

Updated Bayesian detection of foundation parameter with Jeeves pattern search theory

Updated Bayesian detection of foundation parameters in the specific foundation mechanical model was studied based on Jeeves pattern search theory. Firstly, the updated Bayesian objective function for general foundation parameters was derived which could synchronously take the stochastic property of systematic parameters and systematic responses into account. Then the governing differential equations for the Winkler foundation model were gained with elastic Mindlin plate theory and the Fourier close form solution of the foundation model was achieved with the Fourier transform method. After the step length of pattern movement was determined with the quadratic parabolic interpolation method, the updated Bayesian detection of stochastic foundation parameters was resolved with Jeeves pattern search theory and then the corresponding detection procedure was completed. Through particular example analysis, the updated Bayesian detection of stochastic foundation parameters has excellent numerical stability and convergence during iterative processes. Jeeves pattern search theory is unconcerned with the partial derivatives of systematic responses to foundation parameters, and undoubtedly has satisfactory iterative efficiency compared with the available Kalman filtering or conjugate gradient detections of the significant foundation parameters. If the iterative processes are efficiently convergent, it is an important prerequisite that the systematic response assignment should be accurate enough. The derived Jeeves pattern search method with updated Bayesian theory can be applied in other kinds of foundation parameters.     

基于贝叶斯推断的多层软测量建模在丁醇发酵中的应用

针对丁醇生产过程中发酵产物品质参量难以实时测量,现有测量方法精度不高、测量结果受不确定因素影响较大的问题,提出一种基于贝叶斯推断和支持向量回归（Support vector machine regression,SVR）的多层软测量建模方法。首先应用贝叶斯推断计算后验概率、筛选偏置数据,并对偏置数据校准,建立第一层SVR模型;然后利用贝叶斯推断进行二次校准,建立第二层SVR模型,对第一层SVR模型输出进行修正,得到最终预测结果,克服干扰和偏差引起的模型不准确问题。将基于贝叶斯推断的多层支持向量回归（Bi-SVR）预测模型应用于丁醇发酵过程,仿真及实验结果表明,相较于传统SVR预测模型,系统在低干扰的情况下预测精度提高了4.52%,在高干扰时预测精度提高了5.37%。