Predictive pull-based control of an unmanned manufacturing cell, accounting for robot mobility

In order to remove the cell size limitation and to make cellular manufacturing systems more flexible, a manufacturing cell has been equipped with a mobile robot which moves from machine to machine, much like its human counterpart would do. To further acquire a portion of the attributes (intelligence and adaptability of the human worker) lost in robotizing the traditional manned manufacturing cell, a “limited” knowledge-based cell control algorithm is developed. This algorithm maintains standard pull control logic as its underlying basis while utilizing heuristically, knowledge of the ongoing processes and their current status. In this paper, an analysis of this ‘new’ predictive pull-based controller is presented. The simulation results from the predictive pull-based controller are evaluated and compared with the standard pull control logic, first come first serve and shortest distance. A test-bed of an unmanned manufacturing cell with a self-propelled mobile robot was used for experimental verification. Simulation and experimental results show that the degree of improvements gained in the cell performance over the standard pull control method is dependent on the robot’s mobility velocity and how much part inspection/part rejection is performed within the cell. In conclusion, the predictive pull-based algorithm is relatively simple variation of pull control which has the ability of enhancing the “near optimal” pull cell control performance.     

Network flow control under capacity constraints: A case study

In this paper, we demonstrate how tools from nonlinear system theory can play an important role in tackling “hard nonlinearities” and “unknown disturbances” in network flow control problems. Specifically, a nonlinear control law is presented for a communication network buffer management model under physical constraints. Explicit conditions are identified under which the problem of asymptotic regulation of a class of networks against unknown inter-node traffic is solvable, in the presence of control input and state saturation. The conditions include a Lipschitz-type condition and a “PE” condition. Under these conditions, we achieve either asymptotic or practical regulation for a single-node system. We also propose a decentralized, discontinuous control law to achieve (global) asymptotic regulation of large-scale networks. Our main result on controlling large-scale networks is based on an interesting extension of the well-known Young's inequality for the case with saturation nonlinearities. We present computer simulations to illustrate the effectiveness of the proposed flow control schemes.     

Stability and performance analysis of mixed product run-to-run control

Run-to-run control has been widely used in batch manufacturing processes to reduce variations. However, in batch processes, many different products are fabricated on the same set of process tool with different recipes. Two intuitive ways of defining a control scheme for such a mixed production mode are (i) each run of different products is used to estimate a common tool disturbance parameter, i.e., a “tool-based” approach, (ii) only a single disturbance parameter that describe the combined effect of both tool and product is estimated by results of runs of a particular product on a specific tool, i.e., a “product-based” approach. In this study, a model two-product plant was developed to investigate the “tool-based” and “product-based” approaches. The closed-loop responses are derived analytically and control performances are evaluated. We found that a “tool-based” approach is unstable when the plant is non-stationary and the plant-model mismatches are different for different products. A “product-based” control is stable but its performance will be inferior to single product control when the drift is significant. While the controller for frequent products can be tuned in a similar manner as in single product control, a more active controller should be used for the infrequent products which experience a larger drift between runs. The results were substantiated for a larger system with multiple products, multiple plants and random production schedule.     

Optimal operation of Petlyuk distillation: steady-state behavior

The “Petlyuk” or “dividing-wall” or “fully thermally coupled” distillation column is an interesting alternative to the conventional cascaded binary columns for separation of multi-component mixtures. However, the industrial use has been limited, and difficulties in operation have been reported as one reason. With three product compositions controlled, the system has two degrees of freedom left for on-line optimization. We show that the steady-state optimal solution surface is quite narrow, and depends strongly on disturbances and design parameters. Thus it seems difficult to achieve the potential energy savings compared to conventional approaches without a good control strategy. We discuss candidate variables which may be used as feedback variables in order to keep the column operation close to optimal in a “self-optimizing” control scheme.     

Multidimensional visualisation for process historical data analysis: a comparative study with multivariate statistical process control

This paper describes a comparative study of a multidimensional visualisation technique and multivariate statistical process control (MSPC) for process historical data analysis. The visualisation technique uses parallel coordinates which visualise multidimensional data using two dimensional presentations and allow identification of clusters and outliers, therefore, can be used to detect abnormal events. The study is based on a database covering 527 days of operation of an industrial wastewater treatment plant. It was found that both the visualisation technique and MSPC based on T² chart captured the same 17 days as “clearly abnormal” and another eight days as “likely abnormal”. Pattern recognition using K-means clustering was also applied to the same data in literature and was found to have identified 14 out of the 17 “clearly abnormal” days.     

Variable structure exponential stabilization of chained systems based on the extended nonholonomic integrator

In this paper, a new variable structure control strategy for exponentially stabilizing chained systems is presented based on the extended nonholonomic integrator model, the discontinuous coordinate transformation and the “reaching law method” in variable structure control design. The proposed approach converts the stabilization problem of an n-dimensional chained system into the pole-assignment problem of an (n−3)-dimensional linear time-invariant system and consequently simplifies the stabilization controller design of nonholonomic chained systems.     

A human-assisted knowledge extraction method for machining operations

This paper deals with a human-assisted knowledge extraction method to extract “if…then…” rules from a small set of machining data. The presented method utilizes both probabilistic reasoning and fuzzy logical reasoning to benefit from the machining data and from the judgment and preference of a machinist. Using the extracted rules, one can determine the values of operational parameters (feed, cutting velocity, etc.) to ensure the desired machining performance (keep surface roughness within the stipulated range (e.g., moderate)). Applying the presented method in a real-life machining knowledge extraction situation and comparing it with the inductive learning based knowledge extraction method (i.e., ID3), the usefulness of the method is demonstrated. As the concept of manufacturing automation is shifting toward “how to support humans by computers”, the presented method provides some valuable hints to the developers of futuristic computer integrated manufacturing systems.     

Non-hierarchical control of a flexible manufacturing cell

As hierarchically controlled computer-integrated manufacturing systems growthey tend to become complex and their designability, maintainability, expandability and fault tolerance deteriorate. As an alternative, herterarchical control architectures offer prospects of reduced compexity, reduced software development costs, high modularity, high flexibility, and improved fault tolerance. By locating decision making where information originates, global information is reduced to a minimum, scheduling becomes dynamic, machines and parts become “intelligent” entities that cooperatively interact, and the overall system is decomposed into functionally simplified, modular parts.Three flexible machining cell control systems have been constructed at the University of Wisconsin-Madison and are described in this paper: a centralized controller, a hierarchical controller with dynamic scheduling, and a fully distributed heterarchical controller with “intelligent parts”. Comparative results are reported showing that the heterarchical approach possesses a number of advantages including increased fault-tolerance, inherent adaptability and reconfigurability, decreased complexity, and reduced software development cost.     

Results of the Eureka FAMOS research programme: A European contribution to manufacturing technology

Market competition is leading to direct confrontation between products coming from the main manufacturing areas: Europe, the U.S.A. and Asia, where Japan is leading. For Europe to maintain and improve its role, it is important to develop “winning” products. This requires an advanced, technology based industry (secondary sector) and a highly effective research and innovation network (quaternarium), developing new design tools and new configurations for products and production systems. To foster and strengthen an advanced industry and a highly effective quaternarium, Europe has launched various initiatives, from EEC Programmes to the Eureka initiative.In this paper the contribution of the Eureka FAMOS projects is assessed by using a model linking the research innovation cycle, “actors” (companies and research institutions) and activation mechanisms (programmes and initiatives). With the Eureka FAMOS Umbrella Project acting as an activation mechanism, the analysis of the specific FAMOS projects carried out through MONITECH—an assessment activity set up by IMU-CNR—shows how the “actors” (companies and research institutions) have reacted and launched specific projects (36 at an estimated cost of 600 MECU), in 17 industrial sectors, leading to new production paradigms (production systems and related technologies). Considering the results already achieved (pilot production systems and technologies) the Eureka FAMOS Research Programme may be seen as a relevant European initiative in manufacturing technology.The synergy which the Eureka FAMOS Programme is starting to provide with EEC Programmes will give the European research network and industry further benefits.     

Artificial neural networks in process estimation and control

In this contribution, the suitability of the artificial neural network methodology for solving some process engineering problems is discussed. First the concepts involved in the formulation of artificial neural networks are presented. Next the suitability of the technique to provide estimates of difficult to measure quality variables is demonstrated by application to industrial data. Measurements from established instruments are used as secondary variables for estimation of the “primary” quality variables. The advantage of using these estimates for feedback control is then demonstrated. The possibility of using neural network models directly within a model-based predictive control strategy is also considered, making use of an on-line optimization routine to determine the future inputs that will minimize the deviations between the desired and predicted outputs. Control is implemented in a receding horizon fashion. Application of the predictive controller to a nonlinear distillation system is used to indicate the potential of the neural network based control philosophy.     

NOX selective catalytic reduction control based on simple models

In the paper, a new approach to the control of the selective catalytic reduction process of NO_X is presented. The reduction process can be efficiently controlled using a combination of feed-forward and feed-back control structures. However, this approach requires expensive measurement equipment. It is shown that measurement of nitric oxide concentration at the input of the catalytic converter and measurement of the flue gas flow rate can be successfully substituted by corresponding “software sensors”. The software sensors are based on simple mathematical models and on measurements of excess oxygen concentration in flue gas and the fuel flow rate which are usually parts of already available standard combustion control system equipment. In the paper the complete derivation of the required mathematical models is given, and the proposed new control structure is described. The structure is tested using simulation and implemented on an industrial pilot plant.     

Plantwide control: the search for the self-optimizing control structure

Plantwide control is concerned with the structural decisions involved in the control system design of a chemical plant (C.S. Foss, Critique of chemical process control theory, AIChE Journal 19(2), 1973) 209–214; “Which variables should be controlled, which variables should be measured, which inputs should be manipulated, and which links should be made between them?” In particular, the first issue about which variables to control has received little attention. It is argued that the answer is related to finding a simple and robust way of implementing the economically optimal operating policy. The goal is to find a set of controlled variables which, when kept at constant setpoints, indirectly lead to near-optimal operation with acceptable loss. This is denoted “self-optimizing” control. Since the economics are determined by the overall plant behavior, it is necessary to take a plantwide perspective. A systematic procedure for finding suitable controlled variables based on only steady-state information is presented. Important steps are degree of freedom analysis, definition of optimal operation (cost and constraints), and evaluation of the loss when the controlled variables are kept constant rather than optimally adjusted. A case study yields very interesting insights into the control and maximum throughput of distillation columns.     

Cross-sectional comparison of production functions for US manufacturing industries: Some evidence on the urban productivity

This study attempts to show that the production function for a given industry associated with urban areas yields a greater output than that associated with rural areas for a given vector of inputs. For purposes of estimation and hypothesis testing, the production function is of the Cobb-Douglas form with three inputs.Empirical estimates are based on cross-sectional data from the census of manufacturing for each manufacturing industry at the two-digit level of the standard industrial classification. From the data, two subsamples—“urban” and “rural”—are created to obtain separate estimates of the parameters of the production function.The empirical results indicate that for 12 of the 19 industries studied, there is no significant difference between the urban and rural production functions, and, for the remaining 7 industries, the results are not conclusive but do suggest that output elasticities of some factors are greater for urban areas than for rural areas.     

Realization of STEP-NC enabled machining

A STEP-compliant CNC machine tool that demonstrated a G-code free machining scenario is presented. The aim of this research is to showcase the advantages of, and evaluate, STEP-NC—a new NC data model—by implementing it in a legacy CNC system. The work consists of two parts: retrofitting an existing CNC machine and the development of a STEP-compliant NC Converter called STEPcNC. The CompuCam's motion control system is used for retrofitting the machine, which is programmable using its own motion control language—6K Motion Control language and capable of interfacing with other CAPP/CAM programs through languages such as Visual Basic, Visual C++ and Delphi. STEPcNC can understand and process STEP-NC codes, and interface with the CNC controller through a Human Machine Interface. It makes use of STEP-NC information such as “Workplan”, “Workingstep”, machining strategy, machining features and cutting tools that is present in a STEP-NC file. Hence, the system is truly feature-based. The Application Interpreted Model of STEP-NC has been used.     

Steering rules for AGV based on primitive function and elementary movement control

The complete structure of an AGV control system is described in the first part of this paper. The AGV control system is hierarchical and consists of five levels. The structure of one level does not depend on the structures of the other levels. This means that the control system depends on the design of the AGV at the lowest level only, at the actuator servo-control level and its coordination in realizing AGV primitive functions.The second part of the paper describes rules applicable to AGV steering. The structure of these rules depends on two groups of factors. The first group is dependent on information groups fed to the AGV processor by the position sensor. The second group of factors represents aims and conditions and AGV steering such as positioning accuracy, positioning time, allowed room for maneuver, the shape of the given trajectory, etc. The AGV steering rules contain sequences of primitive functions. These primitive functions are of such types as “turn left”, “straighten” (correct), “go straight on”, etc. Trajectory, as one of the basic factors, is defined at the level of controlling an elementary movement. The term “to control an elementary movement” means to select a transport road throughout the transport network and to code it using “elementary movement” such as “go straight” (relating to road section), “turn left” (relating to turning at a crossroad) etc.The results of the AGV steering simulation are presented in the third part of the paper. An exact kinematic AGV model used for stimulating control models is also presented.     

Role of multiplicity in reactive distillation control system design

The impact of steady-state multiplicities on the control of a simulated industrial scale methyl acetate reactive distillation (RD) column is studied. At a fixed reflux rate, output multiplicity, with multiple output values for the same reboiler duty, causes the column to drift to an undesirable steady-state under open loop operation. The same is avoided for a fixed reflux ratio policy. Input multiplicity, where multiple input values give the same output, leads to “wrong” control action under feedback control severely compromising control system robustness. A new metric, rangeability, is defined to quantify the severity of input multiplicity in a steady-state input–output (IO) relation. Rangeability is used in conjunction with conventional sensitivity analysis for the design of robust control structures for the RD column. Results for the two synthesized control structures show that controlling the most sensitive reactive tray temperature results in poor robustness due to low rangeability causing “wrong” control action for large disturbances. Controlling a reactive tray temperature with acceptable sensitivity but larger rangeability gives better robustness. It is also shown that controlling the difference in the temperature of two suitably chosen reactive trays further improves robustness of both the structures as input multiplicity is avoided. The article brings out the importance of IO relations for control system design and understanding the complex dynamic behavior of RD systems.     

Price-driven coordination method for solving plant-wide MPC problems

In large-scale model predictive control (MPC) applications, such as plant-wide control, two possible approaches to MPC implementation are centralized and decentralized MPC schemes. These represent the two extremes in the “trade-off” among the desired characteristics of an industrial MPC system, namely accuracy, reliability and maintainability. To achieve optimal plant operations, coordination of decentralized MPC controllers has been identified as both an opportunity and a challenge. Typically, plant-wide MPC problem can be formulated as a large-scale quadratic program (QP) with linking equality constraints. Such problems can be decomposed and solved with the price-driven coordination method and on-line solutions of these structured large-scale optimization problems require an efficient price-adjustment strategy to find an “equilibrium price”. This work develops an efficient price-adjustment algorithm based on Newton’s method, in which sensitivity analysis and active set change identification techniques are employed. With the off-diagonal element abstraction technique and the enhanced priced driven coordination algorithm, a coordinated, decentralized MPC framework is proposed. Several case studies show that the proposed coordination-based decentralized MPC scheme is an effective approach to plant-wide MPC applications, which provides a high degree of reliability and accuracy at a reasonable computational load.     

Extraction of geometric characteristics for manufacturability assessment

One of the advantages of feature-based design is that it provides data which are defined as parameters of features in readily available forms for tasks from design through manufacturing. It can thus facilitate the integration of CAD and CAM. However, not all design features are features required in down stream applications and not all parameters or data can be predefined in the features. One of the significant examples is property that is formed by feature interactions. For example, the interaction of a positive feature and a negative feature induces a wall thickness change that might cause defects in a part. Therefore, the identification of the wall thickness change by detecting the feature interaction is required in the moldability assessment.The work presented in this paper deals with the extraction of geometric characteristics in feature-based design for manufacturability assessment. We focus on the manufacturability assessment of discrete parts with emphasis on a net shape process—injection molding. The definition, derivation and representation of the spatial relationships between features are described. The geometric characteristics formed by feature interactions are generalized as significant items, such as “depth”, “thickness”, “height” etc. based on the generalization of feature shapes. Reasoning on feature interactions and extraction of geometric characteristics is treated as a refinement procedure. High-level spatial relationships—“is_in”, “adjacent_to” and “coplanar” as well as their geometric details are first derived. The significant items formed from feature interactions are then computed based on the detailed spatial relationships. This work was implemented in a computer-aided concurrent product and process development environment to support molding product design assessment.     

A methodology for expressing CIM reference architectures

TNO is doing research in many areas of industrial automation and is heavily involved in European projects financed by R&D programmes such as Esprit, Eureka and Brite, and in many ISO and CEN standardization activities. From this experience it becomes clear that the I of Integration in CIM has not only to do with the integration of the so-called “islands of automation” but also with the integration of ”islands of manufacturing”: how we can improve the transfer of manufacturing knowledge. We have to increase the semantic content of our integration approaches, so that not only can computer scientist be involved, but also people from the companies we are trying to help, and people who are responsible for the development of new CIM components. The real problem is not a problem of technical integration of computers, but much more a “conceptual modelling” problem. Fundamental questions are, for instance, how we can, on the semantic level really required, model information transfer upstream and downstream in the product life cycle. Based on the analysis of existing CIM projects such as CAD*I, CIM- OSA, Open CAM Systems (Esprit I) IMPACT (Esprit II), CAM-I's CIM Architecture, the Danish Principal model for CIM, and more, we developed a generic and reusable CIM reference architecture. This architecture shows manufacturing activities, real and information flow objects, CIM components and industrial automation standards like STEP, MAP, TOP, EDIFACT, MMS etc. in an integrated way. In this paper we describe the CIM base model used to express the CIM reference architecture and give some details of the CIM reference architecture itself.