<p>This thesis considers the problems of modelling and control of batch processes, a class of finite duration chemical processes characterized by their absence of equilibrium conditions and nonlinear, time-varying dynamics over a wide range of operating conditions. In contrast to continuous processes, the control objective in batch processes is to achieve a non-equilibrium desired end-point or product quality by the batch termination time. However, the distinguishing features of batch processes complicate their control problem and call for dedicated modelling and control tools. In the initial phase of this research, a predictive controller based on the novel concept of reverse-time reachability regions (RTRRs) is developed. Defined as the set of states from where the process can be steered inside a desired end-point neighbourhood by batch termination subject to input constraints and model uncertainties, an algorithm is developed to characterize these sets at each sampling instance offline; these characterizations subsequently play an integral role in the control design. A key feature of the resultant controller is that it requires the online computation of only the immediate control action while guaranteeing reachability to the desired end-point neighbourhood, rendering the control problem efficiently solvable even when using the nonlinear process model. Moreover, the use of RTRRs and one-step ahead type control policy embeds important fault-tolerant characteristics into the controller. Next, we address the problem of the unavailability of reliable and computationally manageable first-principles-based process models by developing a new data-based modelling approach. In this approach, local linear models (identified via latent variable regression techniques) are combined with weights (arising from fuzzy c-means clustering) to describe global nonlinear process dynamics. Nonlinearities are captured through the appropriate combination of the different models while the linearity of the individual models prevents against a computationally expensive predictive controller. This modelling approach is also generalized to account for time-varying dynamics by incorporating online learning ability into the model, making it adaptive. This is accomplished by developing a probabilistic recursive least squares (PRLS) algorithm for updating a subset of the model parameters. The data-based modelling approach is first used to generate data-based reverse-time reachability regions (RTRRs), which are subsequently incorporated in a new predictive controller. Next, the modelling approach is applied on a complex nylon-6,6 batch polymerization process in order to design a trajectory tracking predictive controller for the key process outputs. Through simulations, the modelling approach is shown to capture the major process nonlinearities and closed-loop results demonstrate the advantages of the proposed controller over existing options. Through further simulation studies, model adaptation (via the PRLS algorithm) is shown to be crucial for achieving acceptable control performance when encountering large disturbances in the initial conditions. Finally, we consider the problem of direct quality control even when there are limited quality-related measurements available from the process; this situation typically calls for indirectly pursuing the control objective through trajectory tracking control. To address the problem of unavailability of online quality measurements, an inferential quality model, which relates the process conditions over the entire batch duration to the final quality, is required. The accuracy of this type of quality model, however, is sensitive to the prediction of the future batch behaviour until batch termination. This "missing data" problem is handled by integrating the previously developed data-based modelling approach with the inferential model in a predictive control framework. The key feature of this approach is that the causality and nonlinear relationships between the future inputs and outputs are accounted for in predicting the final quality and computing the manipulated input trajectory. The efficacy of the proposed predictive control design is illustrated via simulations of the nylon-6,6 batch polymerization process with a different control objective than considered previously.</p> / Doctor of Philosophy (PhD)

Master of Agribusiness / Department of Agricultural Economics / Ted C. Schroeder / The CNH Wichita Product Center has had a chronic leak problem with the Skid Steer Loaders. The objective of this project was to analyze the manufacturing plant leak data and make improvements to correct the issue. The objective is twofold: 1) to make process or design improvements on current products produced in the plant and 2) to make recommendations for future designs to prevent such leak issues from reoccurring. The manufacturing data had to be transformed into usable form and then it was analyzed mostly by utilizing Pareto Charts. The highest six problem leak points were chosen from the manufacturing data. Process changes were implemented on these particular leak joints and the results were analyzed using two proportions hypothesis tests. The process changes reduced the leak rate by an average percent reduction of 86 percent. The process changes implemented will also be applied to other similar joints, and results documented in the future. The future design recommendations made from the analyzed data included the increased use of o-ring face seal connections at certain locations and where possible, reducing the number of joints per machine.

Healthcare is facing great challenges and psychiatric care is no exception. Extensive attempts to improve quality are made. It is essential to use methods that enable learning from experience, to improve performance. The core feature of Statistical Process Control (SPC), the control charts, are in use in various settings to enable learning and to support quality improvement work, but its use in psychiatric settings are scarce. This master´s thesis explores the usefulness of control charts, in quality improvement work. This was done in a case study at a department of psychosis by addressing two questions related to: a) control chart’s contribution to knowledge on if, when, where and how changes occur, and 2) how usefulness of control charts is perceived at the department. Control charts were applied to important variables and development officer’s and manager’s thoughts on usefulness were analysed using pattern matching. The use of charts shows shifts and differences between wards related to ongoing improvement projects. There is a readiness to start using control charts. The perceived usefulness matches the benefits and challenges identified in literature. Control charts as a tool supporting continuous improvement work in a psychiatric context, has a great potential still awaiting its use.

Thesis (MScIng)--University of Stellenbosch, 2004. / ENGLISH ABSTRACT: As result of the increasingly competitive performance in today’s industrial environment, it has become necessary for production facilities to increase their efficiency. An essential step towards increasing the efficiency of these production facilities is through tighter processes control. Process control is a monitoring and modelling problem, and improvements in these areas will also lead to better process control. Given the difficulties of obtaining theoretical process models, it has become important to identify models from process data. The irregular behaviour of many chemical processes, which do not seem to be inherently stochastic, can be explained by analysing time series data from these systems in terms of their nonlinear dynamics. Since the discovery of time delay embedding for state space analysis of time series, a lot of time has been devoted to the development of techniques to extract information through analysis of the geometrical structure of the attractor underlying the time series. Nearly all of these techniques assume that the dynamical process under question is stationary, i.e. the dynamics of the process did not change during the observation period. The ability to detect dynamic changes in processes, from process data, is crucial to the reliability of these state space techniques. Detecting dynamic changes in processes is also important when using advanced control systems. Process characteristics are always changing, so that model parameters have to be recalibrated, models have to be updated and control settings have to be maintained. More reliable detection of changes in processes will improve the performance and adaptability of process models used in these control systems. This will lead to better automation and enormous cost savings. This work investigates and assesses techniques for detecting dynamical changes in processes, from process data. These measures include the use of multilayer perceptron (MLP) neural networks, nonlinear cross predictions and the correlation dimension statistic.The change detection techniques are evaluated by applying them to three case studies that exhibit (possible) nonstationary behaviour. From the research, it is evident that the performance of process models suffers when there are nonstationarities in the data. This can serve as an indication of changes in the process parameters. The nonlinear cross prediction algorithm gives a better indication of possible nonstationarities in the process data; except for instances where the data series is very short. Exploiting the correlation dimension statistic proved to be the most accurate method of detecting dynamic changes. Apart from positively identifying nonstationary in each of the case studies, it was also able to detect the parameter changes sooner than any other method tested. The way in which this technique is applied, also makes it ideal for online detection of dynamic changes in chemical processes. / AFRIKAANSE OPSOMMING: Dit is belangrik om produksie aanlegte so effektief moontlik te bedryf. Indien nie, staar hulle die moontlikheid van finansiële ondergang in die gesig – veral as gevolg van toenemende mededinging die industrie. Die effektiwiteit van produksie aanlegte kan verhoog word deur verbeterde prosesbeheer. Prosesbeheer is ‘n moniterings en modellerings probleem, en vooruitgang in hierdie areas sal noodwendig ook lei tot beter prosesbeheer. Omdat dit moeilik is om teoretiese proses modelle af te lei, word dit al hoe belangriker om modelle vanuit proses data te identifiseer. Die ongewone optrede van baie chemiese prosesse, wat nie inherent stogasties blyk te wees nie, kan meestal verklaar word deur tydreeks data vanaf hierdie prosesse te analiseer in terme van hul nie-liniêre dinamika. Sedert die ontdekking van tydreeksontvouing vir toestandveranderlike stelsels, is baie tyd daaraan spandeer om tegnieke te ontwikkel wat inligting uit tydreekse kan onttrek deur die onderliggende geometriese struktuur van die attraktor te bestudeer. Byna al hierdie tegnieke aanvaar dat die dinamiese proses stationêr is, m.a.w dat die dinamika van die proses nie verander het tydens die observasie periode nie. Die vermoë om hierdie dinamiese proses veranderinge te kan identifiseer, is daarom baie belangrik. Ook in gevorderde beheerstelsels is vroegtydige identifisering van dinamiese veranderinge in prosesse belangrik. Proses karakteristieke is altyd besig om te verander, sodat model parameters herkalibreer moet word, modelle opgedateer moet word en beheer setpunte onderhou moet word. Meer betroubare tegnieke om veranderinge in prosesse te identifiseer sal die aanpasbaarheid van proses modelle in hierdie beheerstelsels verbeter. Dit sal lei tot beter outomatisering en sodoende lei tot enorme kostebesparings. Hierdie werk ondersoek tegnieke om dinamiese veranderinge in prosesse te identifiseer, deur die analise van proses data. Die tegnieke wat gebruik word sluit die volgende in:multilaag-perseptron neurale netwerke, nie-liniêre kruisvoorspelling statistieke en die korrelasie dimensie statistiek. Die tegnieke is op drie gevallestudies toegepas om te sien of hulle die dinamiese veranderinge in die data kan identifiseer. Vanuit die navorsing is dit duidelik dat proses modelle nadelig beinvloed word deur niestationêre data. Dit kan dien as ‘n indikasie van veranderinge in die proses parameters. Die nie-liniêre kruisvoorspellings algoritme gee ‘n beter indikasie van dinamiese veranderinge in die proses data, behalwe waar die tydreeks baie kort is. Toepassings van die korrelasie dimensie statistiek gee die beste resultate. Hierdie tegniek kon dinamiese veranderinge vinniger as enige ander tegniek identifiseer, en die manier waarop dit gebruik word maak dit ideaal vir die identifisering van dinamiese veranderinge in chemiese prosesse.

This study utilizes a Monte Carlo simulation to examine the performance of multivariate geometric moving average control chart schemes for controlling the mean of a multivariate normal process. The study compares the performance of the proposed method with a multivariate Shewhart chart, a multiple univariate cumulative sum (CUSUM) control chart, a multivariate CUSUM control chart and a multiple univariate geometric moving average control chart.

Textured surfaces are universally adopted in the steel sheet production industry, and manufacturers are continuously improving the quality of the finished products through intense research in the surface characterisation field. Deterministic Surfaces are textured with specifically designed rolls in order to present a certain degree of regularity, which allows better control over the functional behaviour of the metal sheets. The regularity of the texture impressed on the steel sheets also allows unconventional approaches to surface characterisation and to the assessment of the texture's structure. Statistical analysis is the most effective way to target the isolation of the deterministic part of the surface, which represents the desired product, from the stochastic part, called ‘noise’ and associated with the inaccuracies of production and measurement. This work addresses the problem of characterisation of deterministic textures through statistical analysis, proposing innovative filtering techniques aimed at the realisation of an On-line Process Control System. Firstly the techniques proposed are theoretically formulated and studied, addressing in particular the physical meaning of the geometrical parameters extracted through statistical analysis of highly correlated portions of the textures. A method for isolating the deterministic textures present on a surface, called the Statistical Surface Filter, is presented and discussed in detail, and tested on existing laboratory samples. Secondly the filter is applied to preliminary measurements acquired by an innovative on-line measurement system currently under development, and evidence is shown that the technique is effective in separating the information regarding the regular patterns from the stochastic noise. The possible applications to on-line Statistical Process Control are discussed. Thirdly, the Statistical Surface Filter is tested on a set of measurements representing texturing rolls and textured sheets with different characteristics; statistical analysis of the surface parameters extracted from the filtered surfaces show that the technique allows the assessment of the different contributions of the various stages of the texturing process to the final product. Finally, a software package is implemented for the practical application of the filtering techniques and the parameters extraction; the algorithms that perform the statistical filtering are described and discussed, concluding with the operations of optimisation and fine-tuning for production-line implementation.

Thesis submitted in compliance with the requirements for the Doctor's Degree in Technology: Electrical Engineering, Technikon Natal, 1999. / This study relates to applications of static artificial neural networks (ANNs) to two basic problems of process control: (a) process model identification, and (b) optimal controller tuning. The emphasis is on model identification, where several novel techniques are introduced. A review of the use of ANNs for determining optimal controller settings is included as a logical adjunct which would make the complete system suitable for realisation as a portable or networked system. Three methods for obtaining good approximations for the parameters of first-order processes with long dead time using artificial neural networks (ANNs) are proposed and described. These are termed in this study: time-domain, frequency-domain and model-based methods. In each case the aim was to develop a brief one-shot test that could be applied with minimal disturbance to a closed loop control system. These methods build on existing techniques, but introduce the following novel aspects: 2. The frequency-domain method makes use of the first 81 components of the FFT without further selection as input to a static ANN to yield process parameter estimates. 3. The model-based method uses a simple single-neuron implementation of an ARX model and uses a static ANN to relate process parameter values to the weights of this neuron. In making the analysis, the process input and output are applied repetitively to the neuron model with delays getting progressively larger. Useful effects arising from this are explored. A technique in which ANN training sets are slightly distorted in a random way during training of a radial basis function is developed as part of the time- and frequencydomain methods. The benefits arising from this technique are demonstrated. These experimental ANN-based control methods are evaluated by means of simulations in which accuracy in the presence of measurement noise and performance with higher order processes is measured and analysed. Although the main theme of this study is first-order-plus-dead-time (FOPDT) processes, the full autotuning scheme is tested with some representative higher order processes. Finally, the composition of a complete autotuning scheme is proposed which includes the automatic generation of controller parameters by means of ANN s. / M

A dissertation submitted to the Faculty of Engineering, University of the Witwatersrand, in fulfilment of the requirements of the degree of Master of Science Engineering / A framework designed to contain and manage the use of knowledge in a real-time knowledge based system for high level control of an industrial process is presented. A prototype of the framework is designed and implemented on a static objectorientated shell. Knowledge is stored in objects and in forward chaining rules. The knowledge has a well defined structure, making it easy to create and manage. Rules are used to recognize conditions and propose control objectives. The framework uses the knowledge to determine variables that if altered will meet the objectives. Control actions are then found to implement changes to these variables The use of explicit control objectives makes it possible to determine if an action worked as intended and if its use is suitable for the present conditions. This enables a learning mechanism to be applied in the expert system. The prototype operated adequately, but the knowledge required to drive the. system was found to be very detailed and awkward to create. / Andrew Chakane 2018

An Investigational Project Report Submitted to the Faculty of Engineering, University of the W'itwatersrand for the Degree of Master of Science in Engineering. / This investigational project is concerned with quality management in the process plant manufacturing industry. Process plant manufacture can be defined as the design, manufacture and installation of pressure vessels, pipework, heat exchangers, storage vessels, etc., for the power generation, chemical and oil processing industries, food and beverage, metals producing and mining practices for similar quality assurance (Q.A.) and quality control (O.C.) practices to be employed. (Abbreviation abstract) / Andrew Chakane 2018

Este trabalho apresenta os resultados obtidos no desenvolvimento de um sistema capaz de estabelecer um controle adequado do fluxo da informação e a supervisão de uma rede de sensores. O sistema se caracteriza pelo alto grau de flexibilidade possuindo uma camada de comunicação de dispositivos, uma camada de gerenciamento de regras de negócio para controlar o fluxo de processos, uma camada de interface homem - máquina e de interface com outros sistemas. A camada de comunicação com os dispositivos é responsável pela interface entre os mesmos e o núcleo do sistema, integrando, de forma transparente, diferentes equipamentos, tecnologias e fabricantes. A interface do usuário foi projetada em um único módulo para facilitar modificações sem comprometer o funcionamento geral do sistema. A interface com outros sistemas é feita por meio de drivers de comunicação, permitindo qualquer tipo de integração. O núcleo do sistema faz todo o controle de eventos, do fluxo do processo e geração alarmes, assim como recebimento e envio das informações da camada de dispositivos e da camada de interface. São descritas as diversas camadas da interface e sua implementação analisando as possíveis aplicações do sistema, juntamente com um estudo de caso do controle do processo de descarga de carga a granel no porto de Santos, usando tecnologia RFID. Os resultados obtidos nessa aplicação são descritos e comprovam a utilidade prática do sistema. Apresentam-se inicialmente alguns conceitos básicos necessários ao entendimento deste projeto como: sistemas distribuídos, estrutura webservice, linguagem XML e de tecnologias passíveis de integração com o sistema. / This paperwork presents the results obtained with the development of a system capable of establishing a proper control of information flow and supervision of a sensors network. The system characterizes itself due to its great flexibility degree by having a devices communication layer, a business management layer to do the process flow control, a man-machine interface and interface with other systems layer. The devices communication layer is responsible for the interface between other devices and the system core, integrating in a transparent way, different equipment, technologies and manufactures. The user\'s interface was designed in a single module to facilitate modifications without compromising the system\'s general functioning. The interface with other systems is made throughout communication drivers, allowing any type of integration. The system core makes a control of all events, process flow and alarm generation, as well as receiving and sending information from the devices layer and interface layer. The several interface layers and its implementation are described analyzing the system\'s possible applications along with a case study of loading discharge in a bulk process control at Santos Port using RFID technology. The results obtained with this application are described and prove the practical utility of the system. Some necessary basic concepts to understand this project are presented initially as: distributed systems, web service structure, XML language and technologies subjected to integration with the system.