An active approach to space-reduced NCO tracking and output feedback optimizing control for batch processes with parametric uncertainty

Necessary conditions of optimality (NCO) tracking is a promising approach to run-to-run optimization of batch processes, by converting the optimization problem into a feedback control problem. Since batch processes often contain numerous decision variables that hamper input adaptation in a feedback control manner, the directional effect of uncertainty has been utilized to reduce adaptation directions. This paper proposes an active approach that can further simplify the design of NCO tracking controllers for run-ro-run optimization of batch processes. The idea is to actively restrict the plant inputs in an optimal subspace, prior to the separation of constraint- and sensitivity-seeking directions of plant inputs. For this purpose, an extended system is constructed and then the system is operated by the so-called surrogate variables. Depending on the dimensions of active constraints and uncertain parameters, two cases are distinguished and their NCO tracking controllers are designed respectively. In addition, when the number of parameters is greater than the constraints, the neighboring-extremal based output feedback is incorporated into the active approach, such that the time-consuming gradient evaluations are avoided hence convergence is accelerated. In both cases, the number of adapted directions equals to the number of uncertain parameters. A numerical example and a batch distillation column are investigated to show the new methodology.     

Robust end-point optimizing feedback for nonlinear dynamic processes

The paper suggests two novel approaches to the synthesis of robust end-point optimizing feedback for nonlinear dynamic processes. Classically, end-point optimization is performed only for the nominal process model using optimal control methods, and the question of performance robustness to disturbances and model-plant mismatch remains unaddressed. The present contribution addresses the end-point optimization problem for nonlinear affine systems with fixed final time through robust optimal feedback methods. In the first approach, a nonlinear state feedback is derived that robustly optimizes the final process state. This solution is obtained through series expansion of the Hamilton-Jacobi-Bellman PDE with an active opponent disturbance. As reliable measurements or estimates of all states may not always be available, the second approach also robustly optimizes the process end-point, but uses output rather than state information. This direct use of measurement information is preferred since the choice of a state estimator for robust state feedback is non-trivial even when the observability issue is addressed. A linear time-variant output corrector is obtained by feedback parametrization and numerical optimization of a nonlinear H _∞ cost functional. A number of possible variations and alternatives to both approaches are also discussed. As model-plant mismatch is particularly common with chemical batch processes, the suitability of the robust optimizing feedback is demonstrated on a semi-batch reactor simulation example, where robustness to several realistic mismatches is investigated and the results are compared against those for the optimal open-loop policy and the optimal feedback designed for the nominal model.     

Differential flatness based nonlinear predictive control of fed-batch bioreactors

A trajectory optimization scheme based on the property of differential flatness is proposed in this paper. A dynamic optimization problem is transformed into a lower dimensional nonlinear programming problem through the use of flat outputs. This optimization approach is demonstrated in the repeated optimization of nonlinear dynamic systems under feedback in an approach similar to nonlinear model predictive control. This approach is illustrated on two examples involving biomass optimization and product optimization. Optimization under feedback is studied for the nominal problem and the case where uncertainty is present. The proposed scheme is also used in conjunction with a nonlinear Luenberger observer to generate the optimal trajectories under parametric uncertainty.     

A probabilistic approach for predictive congestion control in wireless sensor networks

Any node in a wireless sensor network is a resource constrained device in terms of memory, bandwidth, and energy, which leads to a large number of packet drops, low throughput, and significant waste of energy due to retransmission. This paper presents a new approach for predicting congestion using a probabilistic method and controlling congestion using new rate control methods. The probabilistic approach used for prediction of the occurrence of congestion in a node is developed using data traffic and buffer occupancy. The rate control method uses a back-off selection scheme and also rate allocation schemes, namely rate regulation （RRG） and split protocol （SP）, to improve throughput and reduce packet drop. A back-off interval selection scheme is introduced in combination with rate reduction （RR） and RRG. The back-off interval selection scheme considers channel state and collision-flee transmission to prevent congestion. Simulations were conducted and the results were compared with those of decentralized predictive congestion control （DPCC） and adaptive duty-cycle based congestion control （ADCC）. The results showed that the proposed method reduces congestion and improves performance.     

Profitable and dynamically feasible operating point selection for constrained processes

The operating point of a typical chemical process is determined by solving a non-linear optimization problem where the objective is to minimize an economic cost subject to constraints. Often, some or all of the constraints at the optimal solution are active, i.e., the solution is constrained. Though it is profitable to operate at the constrained optimal point, it might lead to infeasible operation due to uncertainties. Hence, industries try to operate the plant close to the optimal point by “backing-off” to achieve the desired economic benefits. Therefore, the primary focus of this paper is to present an optimization formulation for solving the dynamic back-off problem based on an economic cost function. In this regard, we work within a stochastic framework that ensures feasible dynamic operating region within the prescribed confidence limit. In this work, we aim to reduce the economic loss due to the back-off by simultaneously solving for the operating point and a compatible controller that ensures feasibility. Since the resulting formulation is non-linear and non-convex, we propose a novel two-stage iterative solution procedure such that a convex problem is solved at each step in the iteration. Finally, the proposed approach is demonstrated using case studies.     

A hybrid swarm intelligence algorithm for multiuser scheduling in HSDPA

Multiuser scheduling is an important aspect in the performance optimization of a wireless network since it allows multiple users to access a shared channel efficiently by exploiting multiuser diversity. To perform efficient scheduling, channel state information (CSI) for users is required, and is obtained via their respective feedback channels. In this paper, a more realistic imperfect CSI feedback, in the form of a finite set of Channel Quality Indicator (CQI) values, is assumed as specified in the HSDPA standard. A mathematical model of the problem is developed for use in the optimization process. A hybrid heuristic approach based on particle swarm optimization and simulated annealing is used to solve the problem. Simulation results indicate that the hybrid approach outperforms individual implementations of both simulated annealing and particle swarm optimization.     

动态输出反馈鲁棒模型预测控制

研究了具有多包不确定性和有界噪声的系统的动态输出反馈鲁棒模型预测控制的综合方法.系统为准线性参数时变的,即系统时变参数在当前时刻精确已知,但在将来时刻未知.采用二次有界概念刻画扩展闭环系统的稳定性.本文主要创新在于提出了一种辅助优化方法,从而为下个采样时刻的主优化问题提供可靠的真实状态的界.通过连续搅拌釜式反应器控制系统验证了该方法的有效性.     

A Multi-step Output Feedback Robust MPC Approach for LPV Systems with Bounded Parameter Changes and Disturbance

This paper considers a multi-step output feedback robust model predictive control (OFRMPC) approach for the linear parameter varying (LPV) systems with bounded changes of scheduling parameters and bounded disturbance. Less conservative bounds of future estimation error sets and system parametric uncertain sets are predicted by considering bounded changes of scheduling parameters in LPV systems. In the multi-step OFRMPC approach, an optimization problem is solved to obtain a sequence of controller gains, which considers predictions of future bounds of estimation error sets and system parametric uncertain sets. The optimized sequence of controller gains corresponding to a sequence of Lyaponov matrices have less constraint conditions and also introduce more degree of freedom for the optimization. The proposed multi-step OFRMPC guarantees robust uniform ultimately bounded of the estimation error and robust stability of the observer system. A numerical example is given to demonstrate the effectiveness of the approach.     

引入相关反馈的基于内容音频检索研究

提出两种基于内容的音频检索的相关反馈算法,一种算法是在对正反馈图像检索技术的改进基础上提出的;另一种算法的提出是基于强制优化概念思想.实验表明,后一种算法比前一种算法具有更好的性能,并且它以一种统一的机制,充分利用了负反馈和正反馈的特性.     

Theoretical analysis of unconstrained nonlinear model predictive control

Model predictive control (MPC) is a well-established controller design strategy for linear process models. Because many chemical and biological processes exhibit significant nonlinear behaviour, several MPC techniques based on nonlinear process models have recently been proposed. The most significant difference between these techniques is the computational approach used to solve the nonlinear model predictive control (NMPC) optimization problem. Consequently, analysis of NMPC techniques is often connected to the computational approach employed. In this paper, a theoretical analysis of unconstrained NMPC is presented that is independent of the computational approach. A nonlinear discrete-time, state-space model is used to predict the effects of future inputs on future process outputs. It is shown that model inverse, pole-placement, and steady-state controllers can be obtained by suitable selection of the control and prediction horizons. Moreover, the NMPC optimization problem can be modified to yield nonlinear internal model control (NIMC). The computational requirements of NIMC are considerably less than NMPC, but the NIMC approach is currently restricted to nonlinear models with well-defined and stable inverses. The NIMC controller is shown to provide superior servo and regulatory performance to a linear IMC controller for a continuous stirred tank reactor.     

Stability radius optimization: A geometric approach

The subject of this paper is a geometric approach to the robustness optimization problem for uncertain finite dimensional linear systems. For the structurally perturbed stabilizable pair (A, B), we show that optimally robust feedback stabilization is possible via the class of feedbacks parameterizing (A, B) feedback invariant subspaces of codimension Rk B. A direct consequence of this geometric approach is a reduction in dimension of the optimization problem.     

复杂生产流程协同优化与智能控制

我国流程行业原料来源复杂,如何优化调控工艺指标使复杂生产流程适应原料波动,是保障产品质量、降低物耗能耗的关键.本文结合全流程、工序、反应器等不同生产层级的工艺特点,系统研究复杂生产流程协同优化和智能控制方法.针对全流程多工序关联的特点,提出了操作模式优化方法和操作模式动态匹配的全流程多工序协同优化方法;针对单元工序多反应器级联的特点,分析了工序内不同反应器的物质转化效率差异,提出了反应器指标梯度协同优化方法;针对反应器多反应共存、工况多变的特点,研究了基于完备状态空间的动态特性描述框架,建立了竞争-促进反应体系机理模型,提出了工况全覆盖的模型参数自主辨识方法和基于分工况智能综合调节的反应器操作参数精细化调控方法.通过锌冶炼智能工厂建设案例阐述了所提方法在提高工艺原料适应能力、生产效率、质量稳定性等方面的成效.最后,结合我国流程行业智能化发展现状和需求,分析了需进一步研究的问题.     

Reliability-based robust Pareto design of linear state feedback controllers using a multi-objective uniform-diversity genetic algorithm (MUGA)

In this paper, fuzzy threshold values, instead of crisp threshold values, have been used for optimal reliability-based multi-objective Pareto design of robust state feedback controllers for a single inverted pendulum having parameters with probabilistic uncertainties. The objective functions that have been considered are, namely, the normalized summation of rising time and overshoot of cart (S_RO–C) and the normalized summation of rising time and overshoot of pendulum (S_RO–P) in the deterministic approach. Accordingly, the probabilities of failure of those objective functions are also considered in the reliability-based design optimization (RBDO) approach. A new multi-objective uniform-diversity genetic algorithm (MUGA) is presented and used for Pareto optimum design of linear state feedback controllers for single inverted pendulum problem. In this way, Pareto front of optimum controllers is first obtained for the nominal deterministic single inverted pendulum using the conflicting objective functions in time domain. Such Pareto front is then obtained for single inverted pendulum having probabilistic uncertainties in its parameters using the statistical moments of those objective functions through a Monte Carlo simulation (MCS) approach. It is shown that multi-objective reliability-based Pareto optimization of the robust state feedback controllers using MUGA with fuzzy threshold values includes those that may be obtained by various crisp threshold values of probability of failures and, thus, remove the difficulty of selecting suitable crisp values. Besides, the multi-objective Pareto optimization of such robust feedback controllers using MUGA unveils some very important and informative trade-offs among those objective functions. Consequently, some optimum robust state feedback controllers can be compromisingly chosen from the Pareto frontiers.     

基于神经网络和专家系统的铸造质量控制方法

提出了一种基于神经网络与专家系统,具有一定自学习能力的铸造质量控制方法,具体描述了其工作原理.该系统由工艺参数优化模块和缺陷诊断模块组成.工艺参数优化模块以神经网络为推理机,以过去生产数据和有限元数值模拟数据作为训练样本,建立工艺参数与铸件性能之间的非线性关系.缺陷诊断模块以基于产生式规则的推理方式,诊断缺陷类型、产生原因及防治措施,且诊断结果反馈到工艺参数优化模块,用于神经网络再学习.试验结果表明,该系统提高了工艺参数准确率,增强了缺陷诊断能力,减少了铸件次品率.     

Systematic top-down economic plantwide control of the cumene process

Plantwide control system design for economic operation over a wide through range (design throughput to maximum throughput) encompassing multiple active constraint regions, is studied for the cumene process. A unique feature of the process is that it recycles the heavy side-product to extinction. A novel top-down control system synthesis approach, where the control objectives for maximum throughput operation are first obtained using steady state optimization followed by control loop pairings with highest priority to economic objectives, is applied. The control structure thus obtained is unconventional with tight active constraint control requiring ‘long’ level loops that maintain the reflux drum and bottom sump levels of a column using the two process fresh feeds. This structure for maximum throughput operation is adapted for economic operation at lower throughputs. Rigorous dynamic simulations show that the structure provides acceptable process regulation for large disturbances despite the long level loops over the entire throughput range. More importantly, no back-off from the active hard equipment capacity constraints also ensure that the loss in throughput from the maximum achievable is negligible. This work is amongst the first reports illustrating the application of the top-down plantwide control system design approach for superior economic performance with robust process stabilization.     

Cartesian反馈线性化技术的延时补偿设计

Cartesian反馈线性化技术是减小射频功率放大器非线性失真的一种有效方式,但信号传输延时却严重影响反馈系统的稳定性,使得Cartesian反馈线性化技术受到带宽的限制.采用Smith预估补偿方法对Cartesian环路进行延时补偿设计来降低时间延时对反馈系统稳定性的影响,并给出了理论分析.计算机仿真结果也表明采用延时补偿设计的Cartesian反馈系统可以使本来早已不稳定的系统保持稳定,而且对输出功率在不同回退级时的线性化性能都有不同程度的改善.     

基于头脑风暴优化算法的Wiener模型参数辨识

Wiener模型是一种典型的模块化非线性模型,广泛应用于工业过程控制领域.由于其结构的非线性,参数辨识无法直接得到解析解.为此,将Wiener模型的参数估计转化为带约束的非线性优化问题,以头脑风暴优化(BSO)算法并行搜索该问题的最优解,并以搜索过程中的反馈信息调整BSO算法的变异过程,以改进算法的收敛速度和辨识精度.数值仿真和工业数据验证了所提算法的有效性.     

Designing MIMO controller by neuro-traveling particle swarm optimizer approach

A nonlinear, multiple input–multiple output controller called the quality controller of neuro-traveling particle swarm optimizer (QC/NTPSO) approach has been proposed in this paper. A reliable controller must stabilize the quality during the manufacturing process and bring the quality characteristics of the manufacturing process close to the target. This controller must also have an adequate feedback system with estimation technology and optimization algorithm. In addition, the artificial intelligence has reasonably been matured and is often used in dealing with construction problems. Therefore, this work constructed a controller with artificial intelligence technology by first using an artificial neural network as the predictor and then using the traveling particle swarm optimizer that is ideal for continuous optimization problems as the algorithm for optimization. The proposed approach has been tested through chemical mechanical polishing (CMP), an important process in semiconductor manufacturing. The result of the test shows that the proposed approach can bring quality characteristics closer to the target than any other approaches.     

A novel derivation methodology for polynomial-LQ controller design

A simple derivation method for optimization of linear-quadratic (LQ) controllers is presented in the discrete-time polynomial systems framework. A control-law variation, regarded as a potential feedforward from the innovations, is used. Orthogonality is evaluated in the frequency domain by collectively cancelling unstable poles by zeros. The suggested method, summarized as a three-step scheme, is exemplified on a disturbance measurement feedforward and an output feedback problem. It is a simple, more direct alternative to the completing-the-squares approach