A novel framework for integrating data mining with control loop performance assessment

Data driven control loop performance assessment techniques assume that the data being analyzed correspond to single plant‐controller configuration. However, in an industrial setting where processes are affected due to the presence of feedstock variability and drifts, the plant‐controller configuration changes with time. Also, user‐defined benchmarking of control loops (common in industrial plants) requires that the data corresponding to optimal operation of the controller be known. However, such information might not be available beforehand in which case it is necessary to extract the same from routine plant operating data. A technique that addresses these fundamental requirements for ensuring reliable performance assessment is proposed. The proposed technique performs a recursive binary segmentation of the data and makes use of the fact that changes in controller settings translate to variations in plant output for identifying regions corresponding to single plant‐controller configurations. The statistical properties of the data in each such window are then compared with the theoretically expected behavior to extract the data corresponding to optimal configuration. This approach has been applied on: (1) raw plant output, (2) Hurst exponent, and (3) minimum variance index of the process data. Simulation examples demonstrate the applicability of proposed approach in industrial settings. A comparison of the three routes is provided with regard to the amount of data needed and the efficacy achieved. Key results are emphasized and a framework for applying this technique is described. This tool is of significance to industries interested in an automated analysis of large scale control loop data for multiple process variables that is otherwise left unutilized. © 2015 American Institute of Chemical Engineers AIChE J, 62: 146–165, 2016     

Estimation of Variance Reduction Opportunities Through Cascade Control

We describe an approach that is useful in deciding if significant benefits, in terms of control loop performance index (through variability reduction), will be achieved by a change in control loop configuration from simple feedback (SFB) to cascade control. The problem is considered in a stochastic setting and solved using the variance decomposition technique. The proposed methodology requires only routine operating data from an existing simple feedback control loop and knowledge of the process delays. Several simulation examples and one experimental case study exemplify the utility of this approach.     

Performance assessment of multivariate control loops on a paper-machine headbox

A large number of control loops in the pulp and paper industry are multivariate in nature. The main control objective of these loops is to reduce process variation. The satisfactory performance of these loops is therefore important for maintaining product quality. This paper introduces a multivariate performance measure of such control loops. The multivariate minimum variance control performance which provides an absolute lower bound on process variance can be estimated from routine operating data, and serves as a good benchmark to assess control loop performance. The proposed measure of performance is evaluated by application to the multiloop control of a headbox machine.     

Detection and Diagnosis of Plant‐Wide Oscillations

This paper presents some emerging techniques for detection and root‐cause diagnosis of plant‐wide oscillations, and demonstrates their efficacy through a successful industrial case study. The recently proposed autocorrelation function based method (Thornhill et al., J. Proc. Control 13, 91–100, 2003a) is used for detection of oscillations in the process measurements. Signals having common oscillations are analyzed for the presence of valve stiction using higher order statistical methods (Choudhury et al., Automatica 40, 1719–1728, 2004b) . A method employing changes in controller gain is proposed for distinguishing an internally generated oscillation from an external oscillatory disturbance. This method of changing controller gain is used to confirm the presence of control valve stiction. The proposed techniques have been used successfully to identify the root cause of plant‐wide oscillations in an industrial case study using routine operating data.     

A review of control loop monitoring and diagnosis: Prospects of controller maintenance in big data era

Owing to wide applications of automatic control systems in the process industries, the impacts of controller performance on industrial processes are becoming increasingly significant. Consequently, controller maintenance is critical to guarantee routine operations of industrial processes. The workflow of controller maintenance generally involves the following steps:monitor operating controller performance and detect performance degradation, diagnose probable root causes of control system malfunctions, and take specific actions to resolve associated problems. In this article, a comprehensive overview of the mainstream of control loop monitoring and diagnosis is provided, and some existing problems are also analyzed and discussed. From the viewpoint of synthesizing abundant information in the context of big data, some prospective ideas and promising methods are outlined to potentially solve problems in industrial applications.     

基于动态多属性决策方法的工业过程控制性能评价

工业控制系统由于设备老化而导致控制性能发生变化,而传统的控制性能评价方法对于这种慢时变特性具有局限性。提出了一种基于变权动态多属性决策的控制性能评价方法。首先在对系统性能进行判断时,利用系统故障率与运行时间相结合的计算方法对系统的运行状态进行划分,获得不同运行阶段的决策信息;采用四个评价指标分别为超调量、非线性指标、输出方差、阀黏滞指标构造多属性决策判断矩阵,对老化慢时变系统进行评价,并运用色谱分析法计算决策过程中属性权重的变化,最终确定当前性能运行状态;将该方法应用于某工业DMF回收装置中,对比实验结果验证了该方法的有效性。     

Dynamic real‐time optimization with closed‐loop prediction

Process plants are operating in an increasingly global and dynamic environment, motivating the development of dynamic real‐time optimization (DRTO) systems to account for transient behavior in the determination of economically optimal operating policies. This article considers optimization of closed‐loop response dynamics at the DRTO level in a two‐layer architecture, with constrained model predictive control (MPC) applied at the regulatory control level. A simultaneous solution approach is applied to the multilevel DRTO optimization problem, in which the convex MPC optimization subproblems are replaced by their necessary and sufficient Karush–Kuhn–Tucker optimality conditions, resulting in a single‐level mathematical program with complementarity constraints. The performance of the closed‐loop DRTO strategy is compared to that of the open‐loop prediction counterpart through a multi‐part case study that considers linear dynamic systems with different characteristics. The performance of the proposed strategy is further demonstrated through application to a nonlinear polymerization reactor grade transition problem. © 2017 American Institute of Chemical Engineers AIChE J, 63: 3896–3911, 2017     

催化裂化反应再生系统的先进控制策略

催化裂化装置具有复杂的过程动态特性,操作约束与经济目标之间存在矛盾,从而为自动控制提出了深入的研究课题.从过程控制的角度讨论了催化裂化反应再生系统的过程特点,并在此基础上提出预测控制与协调决策一体化的先进控制策略.实际工业对象的应用结果表明,这样的控制方案是切实可行的.     

Dynamic analysis and open‐loop start‐up of an integrated radiant syngas cooler and steam methane reformer

The transient performance of an integrated radiant syngas cooler (RSC) of an entrained‐bed gasifier and steam methane reformer (SMR) is investigated. Base‐case designs using either co‐current or counter‐current configurations are subjected to operating transients to evaluate the feasibility to transition to new steady states. Each system, under open loop, is subjected to changes in key variables of the SMR feed on the tube side and disturbances to variables of the coal‐derived syngas on the RSC side to determine the dynamics and stability of the integrated system. The results indicate that the co‐current configuration is flexible to move to new operating steady states and more safe than the counter‐current configuration, although it provides less cooling and has poorer methane conversion. The variables likely to violate the design limit in the event of a disturbance are identified. A start‐up procedure is also established based on industrial practices employed for entrained‐bed gasifiers and methane reformers. © 2016 American Institute of Chemical Engineers AIChE J, 63: 1602–1619, 2017     

The identification of activity profiles in an industrial tube cooled ammonia converter

This paper addresses the problem of identifying space dependent catalyst bed activity profiles in tube-cooled fixed bed reactors from plant data. The method is applied to an industrial tube cooled ammonia converter. Errors in plant data and in the estimated temperature gradients are used to assign confidence limits to the results. The estiamted activity profiles can be used on a routine basis to provide the feedback for a supervisory control, maintaining the reactor at optimum operating conditions. It may also guide the plant operator to timely corrective actions that could forestall catalyst damage, extending its useful life.     

Error‐triggered on‐line model identification for model‐based feedback control

In industry, it may be difficult in many applications to obtain a first‐principles model of the process, in which case a linear empirical model constructed using process data may be used in the design of a feedback controller. However, linear empirical models may not capture the nonlinear dynamics over a wide region of state‐space and may also perform poorly when significant plant variations and disturbances occur. In the present work, an error‐triggered on‐line model identification approach is introduced for closed‐loop systems under model‐based feedback control strategies. The linear models are re‐identified on‐line when significant prediction errors occur. A moving horizon error detector is used to quantify the model accuracy and to trigger the model re‐identification on‐line when necessary. The proposed approach is demonstrated through two chemical process examples using a model‐based feedback control strategy termed Lyapunov‐based economic model predictive control (LEMPC). The chemical process examples illustrate that the proposed error‐triggered on‐line model identification strategy can be used to obtain more accurate state predictions to improve process economics while maintaining closed‐loop stability of the process under LEMPC. © 2016 American Institute of Chemical Engineers AIChE J, 63: 949–966, 2017     

Energy‐efficient complex distillation sequences: Control properties

Four thermally coupled distillation systems were designed for the separation of five‐component mixtures (the light‐ends separation section of a crude distillation plant); their steady‐state design was obtained by starting from a conventional distillation sequence and then optimizing for minimum energy consumption. The thermally coupled distillation systems were compared to sequence based on conventional columns design. Comparison was based on controllability properties under open and closed loop operation, following the dynamic behaviour after common industrial operating disturbances. Simulation results were analyzed by the singular value decomposition technique and with the performance examination of elimination of feed disturbances using PI controllers. It was found that thermally coupled distillation systems are controllable and, sometimes, they exhibit dynamic responses that are easier to manage than in the case of conventional distillation sequences; this result is innovative in the study of this kind of systems.     

基于加权偏离度统计方法的预测控制性能评估算法

针对带区域约束条件的预测控制系统性能评估问题,在考虑过程输出变量约束类型的基础上,提出了基于加权偏离度统计方法的控制性能评估算法。该方法依据控制要求的不同,将输出变量分为质量变量和约束变量,并结合工程经验合理选择变量的权重。基于系统闭环运行数据和约束设置,通过计算变量的加权偏离度得到控制系统的性能评估指标,从而为预测控制器的参数调整和性能提升提供了决策依据。系统仿真实例和工程应用证明了该评估算法对区域预测控制系统性能评估的有效性。     

An extended logical analysis of data approach to fault detections of industrial hybrid systems

It is difficult to deal with industrial hybrid systems involving both continuous and discrete variables using conventional data-driven fault detection methods. While logical analysis of data (LAD) methods are able to effectively explore hidden rules in discrete and continuous data by means of logical analysis for variable associations. However, conventional LAD has the problem of losing trend change information when extracting features of continuous variables. And when processing industrial data with high-dimensional, multivariate features, it will cause a lot of redundancy in the extracted rules. Motivated by these observations, this paper presents an extended logical analysis of data (ELAD) approach to fault detections of industrial hybrid systems. Therein, correlated variables are selected according to the association degree with key variables and additive variable trends are employed to characterize process status changes, creating an explicable fault detection model. The proposed method is applied to the steam drum process of an industrial coal gasification plant in detecting the influence of key hybrid variables on the fault of steam drum level. The results verify the feasibility and effectiveness of the contribution.     

面向工业混杂系统故障检测的扩展数据逻辑分析方法

常规的数据驱动故障检测方法难以处理同时包含连续和离散变量的工业混杂系统,数据逻辑分析（logical analysis of data, LAD）方法通过对历史数据中变量组合的逻辑分析,能够有效地挖掘离散和连续变量数据中存在的隐含规则。然而,常规的LAD在提取连续变量特征时存在对趋势变化信息丢失的问题,并且在处理具有高维度、多变量特征的工业数据时会导致提取的规则存在大量冗余。为此,本文提出一种基于扩展数据逻辑分析（extended logical analysis of data, ELAD）的工业混杂系统故障检测方法,根据与关键变量的关联度选取相关变量,增加变量的趋势信息以进行过程状态变化的表征,生成可解释的故障检测模型。应用于工业煤气化汽包过程,有效地检测了关键混杂变量对汽包液位故障的影响,实验结果验证了所提方法的可行性和有效性。     

Closed-loop dynamic real-time optimization with stabilizing model predictive control

Dynamic real-time optimization (DRTO) is a supervisory strategy at the upper level of the industrial process automation architecture that computes economically optimal set-point trajectories that are in turn passed on to the lower-level model predictive control (MPC) for tracking. The economically optimal solution, in several process industries, could lead to operating the plant at or around an unstable steady state. The present article accounts for this by developing a closed-loop DRTO (CL-DRTO) formulation that enables handling unstable operating points via an underlying MPC with stability constraints. To this end, a stabilizing MPC that handles trajectory tracking for unstable systems is embedded within the upper-level DRTO. The resulting CL-DRTO problem is reformulated by applying a simultaneous solution approach. The economic benefits realized by the proposed strategy are illustrated through applications to both linearized and nonlinear dynamic models for single-input single-output and multi-input multi-output continuous stirred tank reactor case studies.     

MODELING AND PREDICTIVE CONTROL OF MIMO NONLINEAR SYSTEMS USING WIENER-LAGUERRE MODELS

In this work, a Weiner-type nonlinear black box model was developed for capturing dynamics of open loop stable MIMO nonlinear systems with deterministic inputs. The linear dynamic component of the model was parameterized using orthogonal Laguerre filters while the nonlinear state output map was constructed either using quadratic polynomial functions or artificial neural networks. The properties of the resulting model, such as open loop stability and steady-state behavior, are discussed in detail. The identified Weiner-Laguerre model was further used to formulate a nonlinear model predictive control (NMPC) scheme. The efficacy of the proposed modeling and control scheme was demonstrated using two benchmark control problems: (a) a simulation study involving control of a continuously operated fermenter at its optimum (singular) operating point and (b) experimental verification involving control of pH at the critical point of a neutralization process. It was observed that the proposed Weiner-Laguerre model is able to capture both the dynamic and steady-state characteristics of the continuous fermenter as well as the neutralization process reasonably accurately over wide operating ranges. The proposed NMPC scheme achieved a smooth transition from a suboptimal operating point to the optimum (singular) operating point of the fermenter without causing large variation in manipulated inputs. The proposed NMPC scheme was also found to be robust in the face of moderate perturbation in the unmeasured disturbances. In the case of experimental verification using the neutralization process, the proposed control scheme was found to achieve much faster transition to a set point close to the critical point when compared to a conventional gain-scheduled PID controller.     

MODELING AND PREDICTIVE CONTROL OF MIMO NONLINEAR SYSTEMS USING WIENER-LAGUERRE MODELS

In this work, a Weiner-type nonlinear black box model was developed for capturing dynamics of open loop stable MIMO nonlinear systems with deterministic inputs. The linear dynamic component of the model was parameterized using orthogonal Laguerre filters while the nonlinear state output map was constructed either using quadratic polynomial functions or artificial neural networks. The properties of the resulting model, such as open loop stability and steady-state behavior, are discussed in detail. The identified Weiner-Laguerre model was further used to formulate a nonlinear model predictive control (NMPC) scheme. The efficacy of the proposed modeling and control scheme was demonstrated using two benchmark control problems: (a) a simulation study involving control of a continuously operated fermenter at its optimum (singular) operating point and (b) experimental verification involving control of pH at the critical point of a neutralization process. It was observed that the proposed Weiner-Laguerre model is able to capture both the dynamic and steady-state characteristics of the continuous fermenter as well as the neutralization process reasonably accurately over wide operating ranges. The proposed NMPC scheme achieved a smooth transition from a suboptimal operating point to the optimum (singular) operating point of the fermenter without causing large variation in manipulated inputs. The proposed NMPC scheme was also found to be robust in the face of moderate perturbation in the unmeasured disturbances. In the case of experimental verification using the neutralization process, the proposed control scheme was found to achieve much faster transition to a set point close to the critical point when compared to a conventional gain-scheduled PID controller.     

A convolutional neural network (CNN)-based direct method to detect stiction in control valves

Control valves are considered important capital assets in any process industry. A properly maintained control valve can have a significant impact on how well the process is controlled as well as the overall cost of the plant. However, control valves can suffer from poor control performance due to valve non-linearities. One of the main reasons for non-linearity is control valve stiction. Stiction not only causes oscillations in the process variables but also shortens the life of the control valve, resulting in an economic loss for the process. In a process plant, a control engineer generally analyzes the time series plot of process value (PV), set point (SP), and controller output (OP) data and identifies stiction based on the typical shape pattern of PV/SP/OP plot. In this study, the same shape pattern methodology is adapted to identify stiction using convolutional neural network (CNN) technique. A one-dimensional convolution neural network (Conv1D) algorithm is developed, which works directly on PV/SP/OP time series data for stiction detection. The proposed CNN algorithm is tested on both simulated and industrial control loop data. The suggested method provides promising results with a combined stiction prediction accuracy of 92% (92.2% in predicting non-sticky and 91.53% in predicting sticky loops) for the industrial loops data studied.     

MULTIVARIABLE CONTROL OF CATALYTIC CRACKING PROCESSES

Simple, explicit and physically intuitive Feedforward and Feedback control policies are designed for Fluidized Catalytic Cracking Processes. The Feedforward (FF) control algorithm compensates for changes in the feed rate and feed coking tendency by the use of the air flow and catalyst circulation rates as control variables to maintain the conversion and the reactor temperature at fixed levels. Through steady state and dynamic simulations the FF controller is shown to be very effective. To improve the dynamic response of the process and to account for the process/model mismatch a feedback (FB) controller is also designed to complement the FF action. The FB action is designed by use of the transformation related to the physical modes which correspond to the extensive variables of the process. It is shown that the required control structure consists of two loops. One uses the air flow rate to control the total sensible heat content of the reactor and regenerator solid phases. The other loop controls the regenerator enthalpy by changes in the catalyst circulation rate. The air flow rate controller includes an integral action to avoid reactor temperature offsets, while the catalyst circulation rate controller requires a nonlinear static observer to predict the coke concentration on the regenerated catalyst from dense bed and flue gas regenerator temperatures. The performance of the controller for changes on the oil feed rate, caking tendency of the feed, as well as for reactor temperature set point changes is faster and smoother than Kurihara's scheme.