Model-based quality monitoring of batch and semi-batch processes (original) (raw)
Related papers
Fault detection properties of global, local and time evolving models for batch process monitoring
Journal of Process Control, 2005
This paper discusses alternative methods for batch process monitoring. Two alternative methods are investigated and compared to an existing one (the benchmark). A description of the models is given and the performance is discussed by means of fault detection performance indices. The performance indices used are the overall type I error and the action signal time. In order to evaluate the performance of the models in terms of the overall type I error and action signal time, six different batch process data sets have been used. The data sets comprise four industrial data sets, one simulated data set and a laboratory spectral data set.
State Estimation In Batch Processes Using a Nonlinear Observer
Mathematical and Computer Modelling, 2006
In order to control or monitor many technological processes, the problem of states estimation constitutes a strategic topic. With the goal of process control and optimization, the knowledge of some physical state variables provides useful information. This is the case of many ...
2001
Together with some on-line measurements, a reliable process model is the key ingredient of a successful state observer design. In common practice, the model parameters are inferred from experimental data so as to minimize a model prediction error, e.g. so as to minimize an output least-squares criterion. In this procedure, no care is actually exercised to ensure that the unmeasured model states are sensitive to the measured states. In turn, if sensitivity is too low, the resulting state observer will probably generate poor estimates of the unmeasured states. To alleviate these problems, a new parameter identiÿcation procedure is proposed in this study, which is based on a cost function combining a conventional prediction error criterion with a state estimation sensitivity measure. Minimization of this combined cost function produces a model dedicated to state estimation purposes. A thorough analysis of the procedure is presented in the context of bioreactor modeling, including parameter identiÿcation, model validation and design of extended Kalman ÿlters and full horizon observers. ?
Hybrid extended Kalman filtering and noise statistics optimization for produce wash state estimation
Journal of Food Engineering, 2017
Food-borne diseases associated with fresh produce consistently cause serious public health issues. Although sanitization measures are utilized to enhance the safety of fresh produce, strategies that neglect the dynamic nature of commercial wash processes are limited, creating the potential for pathogen cross-contamination and major disease outbreaks. In light of this risk, there is an urgent need for new control approaches during produce washing to reduce the probability of outbreaks. As an important step in this direction, a hybrid extended Kalman filter (HEKF) and particle swarm optimization (PSO)based noise statistics optimization are designed for a produce wash system. The HEKF uses discrete-time free chlorine (FC) measurements, and PSO is used to optimize the noise statistics of the process noise model. The process model and HEKF enable the estimation of chemical oxygen demand (COD) in the water wash, FC concentration, Escherichia coli concentration (PC) in the water wash, and E. coli level (P) on the lettuce. Although control is not explicitly addressed in this paper, the estimation technique proposed here will enable not only monitoring but also advanced control methods. The HEKF is applied to estimate E. coli O157:H7 contamination of shredded lettuce during an industrial wash. The HEKF estimates COD with a root mean square error (RSME) of 8.24 mg/L, FC concentration with an RMSE of 0.09 mg/L, PC in the wash water with an RMSE of 0.19 MPN/ml, and P on the lettuce with an RSME of 0.04 MPN/g. A sensitivity analysis demonstrates that the estimator has good robustness.
Stability Monitoring of Batch Processes with Iterative Learning Control
Advances in Mathematical Physics, 2017
In recent years, the iterative learning control (ILC) is widely used in batch processes to improve the quality of the products. Stability is a preoccupation of batch processes when the ILC is applied. Focusing on the stability monitoring of batch processes with ILC, a method based on innerwise matrix with considering the uncertainty of the model and disturbance was proposed. First, the batch process with ILC was derived as a two-dimensional autoregressive and moving average (2D-ARMA) model. Then two kinds of stability indices are constructed based on the innerwise matrix through the identification of the 2D-ARMA. Finally, the statistical process control (SPC) chart was adopted to monitor those stability indices. Numerical results are presented to demonstrate the effectiveness of the proposed method.
Recursive data-based prediction and control of batch product quality
AIChE Journal, 1998
In typical batch and semibatch processes, process/feedstock disturbances occur fiequently and on-line measurements of product quality uariables are not auailable. A s a result, most batch processes have not been able to achieue tight quality control. Empirical, data-driven approaches are ueiy attractive for dealing with this problem because of the difficulties associated with developing accurate process models from first principles. An approach for recursive on-line quality prediction was deueloped around data-based model structures. Techniques designed to incorporate the predictive models into on-line monitoring and control of batch product quality were also examined. The proposed control approach can be viewed as shrinking-horizon model-predictiue control based on empirical models. The effectiveness of the proposed prediction and control methods are illustrated by using an industrially releuant simulated polymerization example.
Monitoring the Quality of a Chemical Production Process Using the Joint Estimation Method
Journal of Chemical Information and Modeling, 1995
The standard approach to monitoring a chemical production process is the control chart. Such a chart assumes that the data values are independent and identically distributed. It has been shown that such is not so for many production processes especially those of chemical interest. However a more realistic approach to the problem can be developed using time series modeling. A recent development in time series methodology, the joint estimation procedure, allows for the detection and identification of four types of outliers. This study draws the correspondence between these outlier types and out-of-control situations (that is, adverse process changes) in chemical process control. The correspondence is illustrated in this study using a dye liquor data set. Results show that the joint estimation procedure is appropriate for use in process control and provides advantages for dealing with out-of-control problems in production processes.
International Journal of Advanced Manufacturing Technology, 2011
Due to aging and environmental factors, system components may either fail or not function as expected, which causes unprecedented changes in the quality of the system. A timely detection of the onset of a fault in a component is crucial to a quality monitoring of a process if costly failures are to be avoided. However, finding the source of the failure is not trivial in systems with a large number of components and complex component relationships. In this paper, an efficient scheme to detect adverse changes in system reliability and find the failed component is proposed in order to have an effective process quality monitoring. The monitoring scheme has been made effective by implementing first the techniques of fixed-parameter Shewhart, MEWMA and Hotelling’s T2 control chart, and then the adaptive versions of Shewhart Chart, MEWMA and T2 control chart for counter checking the precision of quality reports. Once detected, the fault isolation scheme uses a Bayesian decision strategy based on the maximum correlation between the residual and one of a number of hypothesized residual estimates to generate a fault report. By doing so, the critical information about the presence or absence of a fault, and its isolation, is gained in a timely manner, thus making the quality monitoring system an effective tool for a variety of maintenance programs, especially of the preventive type. The proposed scheme is evaluated extensively on simulated examples, and on a physical fluid system exemplified by a benchmarked laboratory scale two-tank system to detect and isolate faults including sensor, actuator, and leakage ones.