COMPUTER INTEGRATED MANUFACTURING:CIMS IN PROCESS INDUSTRY

CIMS IN PROCESS INDUSTRY

Introduction

Process industry, by which we refer to continuous or semicontinuous production industry processes, principally includes the petroleum industry, the electric power industry, the metallurgical industry, the chemical industry, the paper industry, the ceramic industry, the glass industry, and the pharma- ceutical industry. Process industry is a kind of highly complicated industrial system that not only includes biochemical, physical, and chemical reactions but also transmission or transition of matter and energy. Most process industries are subject to the interlocked relations of enterprise decision making, business marketing, schedule planning, material supplying, repertory transportation, and product R&D, in addition to the characteristics of continuity in wide scope, uncertainty, high non- linearity, and strong coupling. All these factors are responsible for the unusual difficulty of compre- hensive management, scheduling, optimization, and control in process industry enterprises. Therefore, these problems cannot be solved relying on either control and optimization theory, which are based on accurate mathematical models and exact analytical mathematical methods, or automation tech- niques alone (Ashayberi and Selen 1996). The CIMS technique is one possible solution to complex, comprehensive automation of process industry.

Definitions

• Process industry: Those industries in which the values of raw materials are increased by means of mixing and separating, molding, or chemical reaction. Production can be continuous or batch process. The characteristics of process industry must be considered when CIMS is applied to those industries.

• Architecture structure: The models that reflect these characteristics of production and business in process industry. The models represent all aspects of CIMS in the multiview and multilayer approach.

• Models: The structural representations of object concepts. Models include rules, data, and formal logical methods that are used to depict states, behaviors, and the interactive and inferential relations of objects or events.

• Reference model: The model definition for the architecture structure.

• Modeling method: According to the architecture descriptions, designers obtain the descriptions of all the states in an enterprise by abstracting the business function, business data, and business period.

• Information integration: Information integration activities in the production process or enterprise or even group can be described as a process of obtaining, handling, and processing information so that accurate information can be sent punctually and in the right form to the right people to enable them to make correct decisions.

Key Technologies

Because CIMS in process industry is in the developmental stage, some key technologies still need to be developed further:

1. Total technology:

• Architecture structure and reference model of CIMS in process industry

• Business reengineering model and managerial modes of enterprise

• Control modes of Material and cost streams

• Modeling methods for CIMS in process industry

• Structural design methods for CIMS in process industry

• Design specifications for CIMS in process industry

2. Integration technologies:

• Information integration in enterprise and between enterprises

• Integration of relation database and real-time database systems

• Core data model, data booting, data compression, and data mining

• Integration and Utilization of development tools and applications

• Information integration-based Internet, data navigation, and browser technology

3. Network technologies:

• Architecture structure of computer network system

• Openess, reliability, safety, expandability, monitoring and management of networks

• Speed, collisions resolution, concurrency control of networks

4. Supervisor control technologies:

• Distributed intelligent decision-making support system-based intelligent agent

• Optimization model establishment of large-scale systems

• Description and analysis of hybrid system

• Multimode grouping modeling and production operation optimization

• Advanced process-control strategy and intelligent coordination control

• Production safety monitoring, fault diagnosis and isolation, failure forecast

• ‘‘Soft’’ measurement, intelligent data synthesis and coordination

Reference Architecture of CIMS in Process Industry

CIMS in process industry involves complex systematic engineering. Since its inception, advanced management theories, such as BPR (business process reengineering), CE (concurrent engineering), and TQM (total quality management), have been introduced. Using these theories, managers could reorganize departments that overlapped in function so as to facilitate the development of the enter- prise. The realization of CIMS in these enterprises must build a clear reference architecture that can depict all functions in various phases and levels. Under the guidance of the reference architecture, the designers can simulate all potential solutions in an appropriate workbench and determine the total integration solution. The reference architecture of CIMS in process industry can refer to the frame of CIMS-OSA and PERA. The CIMS-OSA frame has many definitions and modeling approaches, in which the concepts are very clear. The PERA frame is very suitable for the definition of every phase in the CIMS life cycle, which considers every human factor that will affect the enterprise integration.

Architecture Structure Model

The architecture structure model of CIMS in process industry has four phases (Aguiar and Weston 1995): the strategic planning, requirement analysis, and definition phase; the conceptual designs phase; the detailed design and implementation phase; and the operation and maintenance phase, as shown in Figure 31. They reflected all the aspects of building process of CIMS. The strategic planning and requirement definition phase relates to senior management. The models in this phase manipulate information with reference to enterprise models and external factors to assess the enterprise’s behav- ior, objective, and strategy in multiview and multidomain so as to support decision making. The conceptual design phase is the domain of system analysis. According to the scope defined in the previous phase, the models in this phase give a detailed description of a system in formalized system modeling technology. A solution will be found that satisfies the demands for performance and in- cludes what and how to integrate. In general, the solution is expressed in the form of functions. Detailed design and implementation will be carried out in the system design phase. In this phase, the physical solutions should be specified, which include all subsystems and components. The models given in this phase are the most detailed. The models in the operation and maintenance phase embody the characteristics of the system in operation. These models, which define all active entities and their interaction, encapsulate many activities in enterprise operation.

The reference models depicted in Figure 31 consist of run-time models, resource models, inte- gration models, system models, and business models, which correspond to the descriptions of the AS-IS system and the TO-BE system in the process of designing CIMS in process industry. Their relationships are abstracted step by step from down to up, opposite to the process of building CIMS.

• Run-time models encapsulate the information related to system operation, such as dynamic math models of production, input–output models, and order management models.

• Resource models contain the information related to relationships between the resource and the satisfaction of demands. In these models, the resource information that designers will add for some special functions is still included.

• Integration models present the way in which various component elements of the AS-IS system and the TO-BE system could be integrated to complete an integrated system.

• System models capture the structure and knowledge of the department in related domains that are currently organized or should be organized to improve the performance of the system. They encapsulate the experiences of system analysts and the descriptions of the prototype.

• Business models collectively contain the business knowledge required to accomplish strategic analysis and requirement definition, including business rules and the accumulated experience from analyzing enterprise performance.

With these reference models, CIMS in process industry could be built from up to down. Each phase is dynamic in nature and can be related to each other phase. That is, the implementation in every phase can be modified to adapt to changes in environment and demands.

Hierarchical Structure Model

The hierarchical structure model is a structured description of CIMS engineering. It is an aggregation of models and their relationships in the whole CIMS of an enterprise. It is the foundation of the design and realization of CIMS. A hierarchical structure model used in CIMS in process industry is shown in Figure 32. It has five levels and two supporting systems. The five levels are the direct control system, the supervisory control system, the production scheduling system, the management information system, and the enterprise decision making system. The two supporting systems are the database system and the computer network system.

The main function of the hierarchical structure model is:

• Direct control system level: This is the lowest level of automated system in the production process, including the distributed control system used in production devices and the fundamental automated equipment used offsite. The parameters of the production process are measured and controlled by the automated system. They also receive instruction from the supervisory control system level and accomplish the process operation and control.

• Supervisory control system level: The system in this level fulfills supervisory control of main production links in the whole production process. According to the instructions from the sched-

uling system level, it formulates process tactics and conducts the actions at the direct control system level, including operation optimization, advanced control, fault diagnosis, and process simulation.

• Production scheduling system level: At this level, the production load is determined and the production planning is decomposed into five days rolling work planning for every month, ac- cording to the information from the decision-making system and the material-stream and energy- stream data. By optimizing scheduling, allocating energy, and coordinating operations in every workshop, the production becomes balanced, stable and highly efficient.

• Management information system level: The system at this level accomplishes the MIS function for the whole enterprise and carries out integrated management of production and business information. According to the instructions from the decision-making system, it makes logical decisions. It is in charge of day-to-day management, including business management and pro- duction management.

• Enterprise decision-making system level: The system at this level comes up with decisions supporting enterprise business, product strategy, long-term objectives, and developing planning and determines the strategy of production and business. Within the company group, it aims at integration optimization in the whole enterprise so as to yield the maximum benefit.

Approach to Information Integration for CIMS in Process Industry

The core of CIMS in process industry is the integration and utilization of information. Information integration can be described as follows: The production process of an enterprise is a process of obtaining, processing, and handling information. CIMS should ensure that accurate information is sent punctually on in the right form to the right people to enable them to make correct decisions.

Production Process Information Integration

Production is the main factor to be considered in CIMS design for a process industry. Driven by the hierarchical structure model discussed in Section 7.2.2, these information models of every subsystem at all levels are built. In these models, the modeling of production process information integration is the crux of the matter. This model embodies the design guidance, centering on production in three aspects:

1. Decision → comprehensive planning → planning decomposition → scheduling → process optimization → advanced control

2. Purchase → material management → maintenance

3. Decision → comprehensive planning → planning decomposition → scheduling → product storage and shipment control

The computation of material equilibrium and heat equilibrium and the analysis / evaluation of equipment, material and energy can be done using this model so as to realize the optimized manip- ulations in the overall technological process.

Model-Driven Approach to Information Integration

Figure 33 depicts the mapping relationship of the models in the building process of CIMS in process industry. It demonstrates that the designed function related to every phase of building CIMS can be depicted from the design view using the structural and model-driven approach (Aguiar and Weston 1995).

The realization of model mapping relies on the building of software tools supporting every phase in a hierarchical way. Model mapping refers to the evolutionary relationships of models between the phases of building CIMS. As the enterprise hierarchy is developed downwards, the description in the models becomes more detailed. In contrast, with the increasing widening of modeling scope, the granularity of descriptions in models will be reduced so as to form more abstract models. For ex- ample, at the detailed design and implementation level, various dynamic math models should be used, and detailed IDEF0 and static math models should be used in the conceptual design phase. We can conclude that the models of various phases in the building CIMS can be evolved step by step in the model-driven approach from up to down.

In the previous analysis, the realization of the model-driven information integration method re- quires a workbench. This consists of a series of tools, such as modeling tools from entity to model, simulating tools supporting the simulations in various levels from higher-level strategic planning to lower-level detailed design, and assessing tools appraising the performance of solution simulation at various levels.

An Application Example

We will discuss an application example of CIMS in a giant refinery enterprise. The technological process of the refinery is continuous, the material stream cannot be interrupted, and strict real-time demands for production manipulation are made. The enterprise aims at the following objectives: material equilibrium, energy equilibrium, safety and high efficiency, low cost and good quality, and optimized operation of the technological process. The realization of CIMS in this type of enterprise requires the consideration not only of problems such as production management, production sched- uling, operation optimization, and process control, but also of business, marketing, material supply, oil product transport and storage, development of new products, capital investment, and so on (Fujii et al. 1992). The computer integrated production system of the enterprise is constructed according to changes in crude oil supply, market requirements for products, flexibility of the production process, and different management modes. The integration of business decision making, production schedul- ing, workshop management, and process optimization is realized in the giant refinery.

Refinery Planning Process

The refinery enterprise consists of many production activities (Kemper 1997). If the blend operation day is called the original day, then the production activities on the day 90 days before that day include crude oil evaluation, making of production strategy, and crude oil purchasing. In the same way, the production activities on the day 10–30 days after the original day include stock transportation and performance adjustment of oil products. Every activity in the production process is relevant to each other activity. For example, in crude oil evaluation, the factors in the activities following the making of production strategy must be analyzed. In another example, people in the activity of crude oil evaluation need to analyze those production activities following the refinery balance in detail. Deep analysis of those activities in the refinery enterprise is the basis of design of CIMS in that enterprise. Figure 34 depicts the refinery planning process.

Integrated Information Architecture

By analyzing of the refinery planning process, we can construct the integration frame depicted in Figure 35.

Using the model-driven approach to the modeling of all subsystems, the information integration model in this refinery enterprise could be built as shown in Figure 36 (Mo and Xiao 1999). The model includes the business decision-making level, the planning and scheduling level and the process supervisory control level. Their integration is supported by two database systems.

The relevant information, such as market, costing, financial affairs, and production situation, is synthesized to facilitate business decisions of the enterprise, and crude oil supply and oil product sale planning are both determined at the business decision-making level.

The planning and scheduling level synthesizes management information, decomposes production planning to short-term planning and executes the daily scheduling, and gives instructions directly to process supervisory control level. In the meantime, it accomplishes the management and control of oil product storage and transport, including the management and optimized scheduling control of the harbor area and oil tank area.

The process supervisory control accomplishes process optimization, advanced control, fault di- agnosis, and oil product optimized blending.

Advanced Computing Environment

The information integration model of the giant refinery depicted in Figure 36 is built using the model- driven method. The model is the design guidance of the realization of CIMS in the enterprise. Figure

37 depicts the computing environment for the realization of the information integration model, using the client–server computing mode (Kemper 1997).

7.5. Conclusions

The reference architecture of CIMS in process industry can instruct designers to optimize solutions by repeatedly optimizing and simulating so as to obtain the final physical system realization in an enterprise. Practical experience indicates that CIMS in process industry is not like selling a car. With the progressive development of technology and the changes in the external environment, CIMS in process industry needs to continue adjusting to yield the maximum economic and social benefit