Search

Disturbance Management is a major issue in process industries like the pulp and paper (P&P) industry. In this paper, a case study in an integrated P&P mill is examined. Production plans for the whole mill need not only to be optimized concerning company's indicators, but also to be robust so that disturbances can be avoided. We present a simulation-optimization approach that generates plans, correctly weighting their quality (regarding various indicators) and robustness. A discrete-event simulation model replicates the dynamics of implementation and adaptation of production plans in practice. The simulation model gives then feedback to optimization, in order to enhance the analytical model, which is thus able to generate robust plans.

During paper manufacture, system actuators need to control the properties of the entire sheet based on a restricted set of data measured by a scanning sensor that traverses the moving sheet. Cross direction variations (CD) are those along an axis perpendicular to the motion of the sheet, while machine direction (MD) variations are those along the axis of motion, and are assumed uniform in CD. Current industrial practice is to separate the relatively slow variations of the CD profile from the higher bandwidth MD variations using low pass filtering, although the spacing and timing of the scanned data measurements makes it inevitable that some process variations will be distorted or lost to aliasing in the filtered data. In this paper, a novel approach to estimation of MD and CD variations is proposed – compressive sensing. In this approach, knowledge of the process is used to help characterize the expected process variations, allowing accurate reconstruction of the true process variations from far fewer measurements than would be indicated by simple bandwidth-based uniform sampling theory. Instead, a random sampling protocol is used to accurately reconstruct the sheet properties. The approach is found to be effective, using simulated and actual industrial process data.

This paper presents an enhanced dynamic causal digraph reasoning method for fault diagnosis and its application to the short circulation process of a paper machine. In order to improve the isolation capability of the original method, an inference mechanism between the arcs of the graph is proposed to locate process faults on the arcs. Application of the proposed method to the paper machine short circulation process is presented at the end of the paper. The results show that the proposed method is able to identify the responsible arcs when the system is affected by a process fault.

In this contribution, the dynamical behavior of a paper-making system is presented and an advanced controller is developed to minimize the amount of recycled paper used in the paper making process. The advanced controller is a model based control strategy, and therefore, a model is developed in order to evaluate the loss of paper during the transition between two points of operation. The wastage model is adapted to be used in the cost function of an in-house developed predictive control algorithm. Performance evaluation is described and the corresponding results are presented.

A general strategy to perform water and element balances was developed and implemented for a pulp and paper mill to identify the flow profiles of water and key elements. The materials covered in the balance were water, sodium (Na), sulfur (S), calcium (Ca), potassium (K), and chlorine (Cl). A sampling strategy for the whole plant was first developed and then a number of analytical techniques were chosen, taking into account the special/difficult characteristics of the samples found in a typical Kraft pulp and paper mill. To perform the balances, boundaries around each section of the plant were selected and additional information was created by incorporating extra knowledge/relationships within each section, thus providing redundant information. Error analysis and data reconciliation were performed to evaluate the reliability of the results and correct them, to fmally obtain an adjusted set of data consistent with the balance equations. The reconciled data obtained, as expected, showed better consistency in regards to the balance equations and reduced uncertainties. The final mass balance simulation toolkit developed enhanced the calculation and data processing activities, allowing material balances to be perfonned continuously to provide up-to-date results to serve for “on time” monitoring and “on time” decision making by mill staff.

Typically in pulp and paper processes raw material quality variation due to seasonal deviations as well as measurement drifts cause difficulties in setting the tight alarm thresholds for quality and control measurements. By using adaptive thresholds, more sensitive measurement range and thus reduced quality variation can be achieved. In this paper, new adaptive classification algorithm is proposed and validated using simulated and real mill data.

In this study an indicator for the prediction of paper machine runnability was developed as a Case-based reasoning (CBR) type application with linguistic equation (LE) approach and fuzzy logic. The foundation of this application is a case base containing models of various operating situations with different amount of breaks. The indicator compares online measurements to the examples in case base and uses the information of the best fitting case to identify the current situation. This paper presents the application with some test results.

Recently, La Papelera del Plata implemented the Paper Machine IT project in zarate, Buenos Aires. The integration of control system with supervisory sensor based systems, actuator systems, PLC systems and drives was considered the most important. True system integration requires system level interoperability. ABB has achieved this by offering a wide range of pulp & paper solutions built upon a common system platform. This paper describes how this project was executed and presents the results achieved by La Papelera del Plata paper mill through the application of integrated system solutions, on the Advant Open Control System (OCS) with Master software foundation.

Since most of the chemical processes are represented typically by nonlinear models, the use of nonlinear geometric control in chemical processes is straightforward. This paper presents a decoupling application in a paper machine headbox. The nonlinear geometric control is implemented with blocks for reference trajectory, tracking error, integrators of error feedback and with an anti-windup scheme. The results show a clear implementation scheme and good dynamical responses.

The topics of recovering of materials, quality control, and environmental impact have gained a raising attention in industrial activities, in order to enforce the quality parameters imposed by the new laws. An innovative approach to control and optimisation of the manufacturing process becomes necessary to cut the industrial costs. For this reason, the interest for the intelligent methods for monitoring and control several process phases has been growing since 80's. This paper deals with the development a Neuro-Fuzzy (NF) system for the corrosion control in Pulp & Paper production. Furthermore, the performance of the NF expert system has been compared with that of the classical Neural Network MLPs (NN).

Because of tightened quality requirements of the control of CD (Cross Direction) profiles dilution control headboxes are getting more common. Dilution control utilizing white water makes the short circulation more complicated compared to the traditional short circulation. This paper discusses control problems related to the short circulation of the paper machine having a dilution headbox. A list of the most common disturbance sources is given based on the experience gained from production machines. Consistency control applications based on Kajaani RMi consistency analyzer are presented with results. Control applications are dilution water consistency control, break time headbox ash consistency control and thick stock ash content feedforward control. These applications are used to improve the stability of the machine during normal operation and to recover faster from web breaks.

The pressurized headbox is one of the most important pieces of equipment on a modern paper machine. It is a special two-inputs-twooutputs process, and proper matching of input and output of control loops is very important. Relative gain analysis and computer simulation based on a mathematics model of the pressurized headbox shows that the common used configuration in paper industry is not the best. For this kind of configuration, a multi variable predictive control algorithm is designed. Computer simulation shows that this kind of control algorithm can deal with the headbox control very well. A control system with this kind of controller for full scale pressurized headbox has a good performance with excellent control precision, very short response time and very small offsets.

In order to prepare a research project a study was conducted in the pulp and paper industry. The identified problems were around the risks associated with the design of control systems. The investigation suggested that a training program was the best mean to improve the design in order to reduce the risks associated with the considered machinery. This training program would be oriented toward personnel who design or prepare specifications for control systems. This paper will present the result of the study, which will be materialized in the project currently carried out throughout the pulp and paper industry in Quebec.

In papermaking, as in other industrial activities, increasingly stringent customer requirements for product quality and consistency necessitate tighter control of the manufactwing processes. This paper describes a neural network vibrationbased condition monitoring system for providing advance warnings of malfunctions in the paper machine. Spectral analysis of the raw vibration signals involves the technique of enveloping for isolating the spectral components that reflect the process defects. Feature extraction is then carried out by a suitably designed peak detection algorithm. This is followed by a neural network classifier for discriminating between the fault and the no-fault conditions. The system's performance is assessed on industrial data from a major paper manufacturer.

This paper describes a money paper quality control system based on image processing techniques. The system inspects paper sheet dimensions and verifies whether the money paper has any surface quality or dimensional defect (holes, stains, correct mark positioning,…) while it is moving at a very high speed 20-100 m/min. Due to this high advancing speed of the paper, the high number of inspections to accomplish and the minimum size of the defects to detect (1 mm 2 ), there is no other solution to automate this process but the use of an image processing system based on a high resolution (2048 pixels) lineal camera working at 20 Mhz.

Two methods for instrumental measurement of the dry line have been developed. They are based on innovative arrangements of illumination and related observation under which the pulp on wire is seen by video camera as a composition of dark and light fields. The complete dry line is extracted from the image and used industrially for feedback control of the dry line in machine direction (MD). Recorded correlations between changes of dry line and dry basis weight profile justify conclusions on their control in cross direction (CD) and on incorporation of dry line measurement into conventional MD and CD controls which are handicapped by a long process delay.

This paper is concerned with the task of sensing in web-forming process control loops. Conventional and newly proposed measurement techniques are discussed and a new type of measurement is suggested. The new measurement method adresses the problem of paper machine 'wet-end' sensing and is based on a CCD camera system which can detect the dry-line. Both the theoretical and practical aspects of the invention are investigated and presented using non-parametric techniques based on analysis of variance. Compared to traditional scanning gauge systems the dry-line detector allows for a large reduction in the transport delay, higher data density, and better machine directional (MD) and cross directional (CD) signal decomposition.

Production rates in modern paper machines are increasing all the time. This poses challenging requirements for control of different subprocesses in the paper machine, e.g. for the headbox. Currently, scalar controllers are used for headbox control in industrial practice almost everywhere. Headbox control has been a favorite example of multivariable control in many simulation and laboratory studies but not in industry. This paper presents an application of multivariable PI controller to the paper machine headbox. The procedure takes into account the limited tests that can be performed in practice, simplicity in structure and tuning, and realization using modern distributed automation systems.

This work proposes an expert management system of winding as linked with paper machine production planning and coordination. The integrated management system operates at three levels which receives the information of deckle buckets sent by paper machIne production planning. The main goal of deckle optimization at the highest level is to minimize the paper waste in arranging the reels in a bucket and provide the deckle information to a lower level. The winder optimization at the lower level receives the deckle information. The general goal of winding optimization is to maximize the roll quality and to provide the winding rate equal to the paper machine production. Corresponding to the generated optimum results, the winder optimization block generates the supervisory set point values such as: for web tension, and the target diameter of finished roll, to provide a high quality roll with the desired final diameter.

Retention control has largely remained unsolved, although it is one of the most important control loops still to be closed in the paper machine. In this paper retention modelling is done using multivariable time-series approach. Based on careful selection of vanables and model structure, several new control strategies are developed. Different feedback controls are successfully tested on a pilot paper machine. It is believed to be the first time when successful retention control has been reported in the literature.

This paper presents an industrial application of multivariable combined self–tuning control techniques to a paper machine. The papermaking process is taken as a MIMO system with unequal time–lag, some time–varying parameters and unstable inverses. In order to design a suitable control scheme for the process, a new control strategy, called multivariable combined self–tuner(MCST), with the U–D factorization recursive algorithm is proposed. The capability of the proposed scheme is demonstrated by its application to the basis weight loop and the moisture loop of the Long Wire Paper Machine(LWPM) at Juxian Paper Mill, Shandong, P.R. China.

This paper presents an efficient real-time method of sensor fault detection, isolation and estimation for a class of MIMO nonlinear time-varying stochastic systems. The sensor faults may be a jump or slow drift type. The presented approach is based on the combination of the separate-bias estimation algorithm to randomly time-varying bias of nonlinear systems proposed by the authors and the modified Bayes' decision algorithm. The proposed technique is used to design a sensor fault diagnostics system for the headbox of an actual paper machine. Actual headbox data have demonstrated the effectiveness of the proposed method

Whereas condition monitoring as applied to the mechanical processes of papermaking are well established, the complex situation in the wet end is only beginning to be considered. Examples of the effect of retention on mean time, pH measurement and sizing are commented on. A simple model which can be used for thinking about and developing sophisticated control procedures is presented

Modem paper machines are very large and complex. The good runnability of a paper machine is very desirable but hard to achieve. Breaks are one of the most important problems considering the runnability and the productivity of a paper machine. In this paper we discuss the methods to detect situations, where the probability of a break is exceptionally high. Operators have usually some kind of feeling about situations, which can increase the probability of a break, but it is impossible to find a unique set of features which cause breaks. Our approach is to collect data from a production paper machine and then analyse the data. The measurement points have been selected in co-operation with the paper machine staff. Principal component analysis has been used to visualize the occurrence of breaks and the movements of the state of a paper machine. Because very little unique knowledge about breaks is available, we use unsupervised neural networks to classify the measurement data and to find situations which are sensitive to breaks. Kohonen self-organizing feature maps are suitable because they create classifications in which classes of similar features become near each other. This paper has two major contributions. First the paper describes the mathematical results for the detection of paper machine breaks and second the paper presents a concrete application, not a simulation study, of the use of artificial neural networks.

Inspired by the acrobatics of the lizard, we present a novel robot platform capable of performing a barrel roll from a ramp. The system is modeled using Euler-Lagrange mechanics, followed by controller design and numerical simulation. A robotic platform is then designed to perform the experiments. Finally, we show that purely by the use of the actuated tail, the robot can rapidly performing a 360 degree roll rotation in under a second.

By focusing on core competencies and the utilization of international location advantages such as access to low-wage labor and local sales markets, global production networks are becoming increasingly complex. Additionally, turbulent times with great uncertainty forces globally acting enterprises to adapt their production networks. Therefore, it is crucial to determine the right point in time for change and also to identify the beneficial production network alternatives. In other words, globally acting enterprises have to answer the two questions of what to change and when to change. The presented method describes an approach for the strategic planning of global changeable production networks, based on future scenarios and a multi-objective optimization to identify the most favorable network configuration during the planning horizon. The multi-objective optimization is modeled as a mixed-integer MODM problem (multiobjective decision making) including seven criteria i.e. costs, delivery time, quality, flexibility, customer proximity, coordination and plant qualification.

This paper presents the results of a project aimed at minimising fuel usage while maximising steam availability in the power and steam plant of a large newsprint mill. The approach taken was to utilise the better regulation and plant wide optimisation capabilities of Advanced Process Control, especially Model Predictive Control (MPC) techniques. These have recently made their appearance in the pulp and paper industry but are better known in the oil and petrochemical industry where they have been used for nearly 30 years. The issue in the power and steam plant is to ensure that sufficient steam is available when the paper machines require it and yet not to have to waste too much steam when one or more of the machines suffers an outage. This is a problem for which MPC is well suited. It allows variables to be kept within declared constraint ranges, a feature which has been used, effectively, to increase the steam storage capacity of the existing plant. This has resulted in less steam being condensed when it is not required and in significant reductions in the need for supplementary firing. The incidence of steam being dump-condensed while also supplementary firing the Combined Heat & Power (CHP) plant has been reduced by 95% and the overall use of supplementary firing is less than 30% of what it was. In addition the plant runs more smoothly and requires less operator time. The yearly benefit provided by the control system is greater than £200,000, measured in terms of 2005 gas prices

As the speed of paper feeding processes in an ink jet printer is increased, the paper feeding mechanisms are usually prone to induce some paper transport problems such as paper jamming and slipping, and some mechanical problems such as gear cracking and chattering. Therefore, for ensuring printing qualities and for preventing destructive damages, the paper path detection in ink jet printers is required. In this paper, we explore the feasibility of using speed perturbation observer to detect paper path situations in an ink jet printer. Moreover, in comparing with the existing approaches, this paper presents an approach without additional sensors. The speed perturbation observer composed of a nominal plant, a low-pass filter, and an identification unit is developed to detect paper path situations during paper feeding processes. The design concept is based on the fact that external disturbances usually significantly affect the speed performances of an axial motion system. Since the paper feeding processes in ink jet printers are equivalent to that the interaction among paper and paper feeding mechanisms exerts external disturbances on the driven motor, it is interesting to detect paper path situations by estimating speed perturbations induced by different paper feeding actions. The proposed approach is implemented on an ink jet printer and two usual paper path situations are tested. The experimental results demonstrate that the proposed approach can accurately respond to the different paper path situations caused by different paper feeding actions.

The increase of educational activities in automatic control, based on the use of virtual and remote laboratories, is already a fact. For that reason, it is necessary to include and update educational contents in agreement with the technological reality. This implies the development of methods to integrate new educational tools in the aforementioned laboratories in an easy, secure and flexible way. This paper shows how to integrate an educational tool developed in LabVIEW, widely used software, into the structure of our remote laboratory (LRA-ULE). The educational tool provides students with a control problem on a physical system where they need to design a PID regulator. They can also simulate its closed-loop behavior and identify the system in order to compare it with the theoretical model.

Traditional H ∞ design methods available in commercial softwares lead to controllers which present the same order as the model of the plant. Considering the high order of teleoperation systems, implementing such controllers may require the use of order-reduction methods. To avoid this extra step in the design, we propose to produce fixed-order controllers directly based on a full-order model of the system. In this paper, a zero-order (static output feedback) controller is designed for a teleoperation system. The design procedure, based on a cone complementary formulation is described and applied to a 22 nd -order model of the system. The resulting controller is implemented on a one degree of freedom teleoperation system. The experimental results obtained with this simple control structure validate the use of very low order control structures in the case of teleoperation systems with the extra advantage, with respect to conventional H ∞ design techniques, that the involved and time-consuming order reduction step is avoided.

This paper proposes the design of an observer-based control law for a continuous stirred tank reactor with recycle time-delay. Recent results on differential geometry for nonlinear time-delay systems are used. The many performed simulations show the high performance of the proposed control law.

A modern paper machine is a very complicated piece of machinery and full of automation. Paper machine specialists can be in different locations around the world and that is why efficient communication networks are needed to transfer technical information between them. This paper specifies the terminology to cope with information transfer in a paper machine project. The information transfer techniques and system architecture during different generations of the paper mill automation systems are described briefly. Limitations of different generations of remote services are also discussed. The need for different communication networks is specified as well as the need for changes in management of cooperation between different partners in paper industry. The use of Internet technology is described in the communication systems. Copyright © 2002 IFAC

This paper presents a control design approach for robust sheet control in a printer paper path. The overall design question is formulated in terms of a hierarchical control set-up with a low level motor control part and a high level sheet control part. The sheet dynamics, subject to disturbances and parameter uncertainties, are captured in the piecewise linear modeling formalism. Based on this model, the control design yields robust sheet feedback controllers and a closed-loop system capable of rejecting disturbances up to a prescribed level. The effectiveness of the control design is shown on an experimental paper path setup.

Sheet-break is a long standing problem in pulp and paper industry. If advance warnings can be issued ahead of time, corrective actions may be taken to prevent some of the breaks. In this paper, we describe the development and implementation of an online monitoring scheme. A data based approach was taken in this analysis. PCA was used to model the process, and a combined index based on the Hotelling's T 2 and Q-statistics was developed as a sheet-break predictor. As the process is subject to external disturbances, grade changes, and frequent interruptions, preprocessing of the data played an important role in getting consistent results. We used several novel techniques for data selection, scaling and modelling. The models were validated using a large validation data set with known fault condition.

This study focuses on the applications of web based performance and knowledge support systems in the paper industry. Two different systems – a generic KnowPap learning environment and a customized electronic performance support system (EPSS) were evaluated. The results indicate a very limited usability of generic systems for operational and knowledge support. The customized KnowFine system was found very useful to support the operations and maintenance. In future this kind of customized systems will be integrated to process control and information management systems. Optimally the lice-cycle of these systems follows the life-cycle of the plant.

Enterprise-control system integration between business systems, manufacturing execution systems and shop-floor process-control systems remains a key issue for facilitating the deployment of plant-wide information-control systems for practical e-Business-to-Manufacturing industry-led issues. This achievement of the Integration-in-Manufacturing paradigm based on centralized/distributed hardware/software automation architectures is shifting by the Intelligence-in-Manufacturing paradigm addressed by the IMS industry-led R&D initiative in order to define and to experiment the next generation of manufacturing systems capable to cope with the high degree of complexity of meeting agility over flexibility and reactivity in customized manufacturing. This survey paper of the TC 5.1 summarizes these key problems, trends and accomplishments for manufacturing plant control before to emphasize for practical purposes some rationales and forecasts in deploying automation over networks, HMS and its related agent-based technology, as well as in applying formal methods to ensure dependable manufacturing.

This paper is aimed at exploring the understanding of the changes that face the lower levels of the supply chains as business moves from a global market into service with multiple highly changeable products and offerings, demanding interoperability across skills, customers and suppliers.

Paper manufacturing involves the sequential removal of water from pulp by means of gravity, vacuum dewatering, mechanical pressing and thermal drying. This research proposes a new control topology for moisture content using surrogate measurements to infer in-process moisture content, and the pressure settings in the vacuum dewatering section as actuators. The new topology has the potential to overcome the performance, robustness and energy efficiency limitations that the current process control topology has, due to a long dead-time and excessive use of the dryers. A preemptive control law was previously proposed for the new topology to decouple the upstream disturbances from propagating downstream in the process. In this paper, we develop an adaptive version of this pre-emptive controller to alleviate the need to know the dewatering coefficients precisely. The adaptive control problem has the interesting aspects that the unknown parameters are nonlinearly embedded in the error dynamics, and there is a time delay before the error in the estimated parameters can be observed.

An analytical model of moving paper in a printing station is proposed. Performance of basic PI control is discussed, and noninteracting control of machine speed and paper stress is analyzed. Theoretical results are illustrated with simulation data.

This paper presents a novel approach for the change detection of the output probability density functions for dynamic stochastic systems. Using the B-splines neural network, the measured probability density functions of the system output are represented by a set of weights which are functions of the control inputs to the system. This leads to a unique expression of the dynamic characteristics of the output probability density functions for the system. An observer based fault detection is developed which detects any unexpected changes of the parameters in the dynamic part of the system. An application to the detection of unexpected changes of particle size in paper-making is included to demonstrate the use of the proposed algorithm and desired results have been obtained.

In today's increasingly competitive markets, leading engineering organizations need to capitalize on every available opportunity to increase the quality oftheir products while reducing the length of the development cycle and manufacturing costs. This paper presents an integral approach for the reliability validation process of a product in the automotive industry. The stress/strength concepts are applied to estimate and validate the product reliability in which case both stress and strength are random occurring parameters within the reliability validation process. The simulation approach using vinual prototypes is presented as well.

Present evolution of manufacturing systems is characterized by a progressive increase of the importance of knowledge-driven work in industrial plants of very large dimension. As the system complexity and the intelligent agents number grow, the system response uncertainty seems to grow also: this suggests to have greater attention to the agents intelligence. This paper first aims to verify if the increasing complexity of actual work systems is going to allow a greater and greater autonomy to individual agents. Then, taking into account that final scope of an enterprise is profit, a positive relation between global system efficiency and local agents autonomy is proved.

This paper addresses the problem of estimating the effective actuator centre locations in a paper machine whilst the process is operating in closed loop. A reduced order estimator employing a Radial Basis Function Network is described, where the basis functions in the hidden layer are defined using Discrete Chebyshev Polynomials. The problem is considered as a nonlinear optimisation problem and solved online using a recursive prediction error method.

This paper presents some results of field studies of six different paper production lines. The aim of the research was to analyze the need of flexibility in paper production and its impacts on both organizational and technological issues. This paper deals mainly with the latter ones. The results of the research show an increased need of flexible operation of the paper process. To effectively support these features some sophisticated properties are required of the automation and information systems. One main principle is integration of knowledge in both horizontal and vertical directions along the production line.

The paper describes advances made in a computer method for the detection of paper shrinkage during the drying stages. The technique is based on two-dimensional Fourier transformations used to detect those periodic marks left on the paper by the wire during the formation stage. These marks are obtained from images of paper taken by low-cost CCD cameras for a later matching with those of the wire. Differences in separation of lines for the two cases are due to paper shrinkage. Finally some comments on hardware considerations of the sensor design and potentials of the method for the paper and graphic industry are given

The aim of this paper is to present a methodology to design a generic Petri net based model for the class of paper mills finishing processes. This model is based on timed coloured Petri nets with decision solving procedures. This modeling tool is called by the generic term, high level Petri net. A special interest is given to represent the decision points of the system as conflicts in the net. The model is meant for simulation, to carry out qualitative analysis about the system. The purpose is to observe resource sharing and the consequences of different strategy choices on the production. Machine breakdowns, operation-machine allocation and times to change from one sort to another are taken into account. The basic concepts of the modeling tool and the methodology are first presented. Then, the characteristics of the class of processes are given. Finally, the generic model is explained.

This paper presents a model-based fault detection and fault-tolerant control technique for the pressurized headbox of a paper machine. A bank of Kalman filters is constructed corresponding to all the possible sensor failure modes. The possibility that each failure mode hypothesis is true is calculated using measurement innovations processes. The sensor failures are detected and located based on the calculated possibilities of the hypotheses. The controller and state estimator are automatically reorganized subsequent to the occurrence of failures to ensure the stability and good performance of the closed-loop system. The issues of system hardware redundancy and computational burden as well as implementai complexity are taken into account in the system design. Simulation results have shown satisfactory performance of the headbox control system after applying the presented technique.

We outline the needs for diagnostics in the paper industry and describe the state of the art. We have developed two next-generation diagnostic systems: WEDGE for the diagnostics of temporal variations, and FLEXPRO for spatial (cross-direction) variations. Both these systems combine powerful signal analysis tools with an intelligent user interface. Thus our systems can be used on a daily basis at paper mills, without an assisting process analysis expert. The flexibility of these systems allows easy adaptation to the continuously changing needs of a mill. Finally, we briefly discuss technological changes in papermaking which affect the diagnostics needs in the 1990's.

This paper presents a feedback control design and a statistical quality/process control (SQC/SPC) system implementable in real time via distributed control systems to provide a quality product. A case study is examined using a simulated paper production process, specifically, control of cross-directional basis weight of paper. As a model of the process, the paper variance is partitioned into its machine directional, cross-directional, and random components. Pole placement and Smith Predictor techniques are used to control the paper cross-directional profile. The proposed control scheme generates and implements the supervisory setpoint corrections for the cross-directional basis weight control. A distributed control hardware and distributed functional algorithms are utilized . Such a proposed system provides supervisory corrections during the production process prior to final measurements for feedback control for a quality product.

One main problem in dynamic system monitoring is to know over time in which physical situation the system is operating, in order to early anticipate possible failures. This paper addresses the problem of tracking a physical system and presents a model-based diagnostic system architecture, QDIAG, whose goal is to help operators early detect anomalies in dynamic systems in operation. QDIAG integrates both quantitative and qualitative techniques, using QSIM, Kuipers' qualitative simulation algorithm.