The manner in which the software components are integrated into the knowledge-based modelling environment KEMS is referred to as 'medium coupling'. This is unlike 'loose coupling' which uses simple fil...
详细信息
The manner in which the software components are integrated into the knowledge-based modelling environment KEMS is referred to as 'medium coupling'. This is unlike 'loose coupling' which uses simple file transfers between modules, and 'tight coupling' which uses a single language for implementation. KEMS runs on PC 286 and 386 systems under MS DOS. An important reason for targetting onto these machines is the desire to make simulation tools available to a range of users at widely differing sites. The concepts within the prototype software are being evaluated using process models for gas distribution and control supplied by British Gas. The knowledge-acquisition front-end is being used to increase the model repertoire, and is itself being evaluated for use by engineers. Further knowledge-based tools which are to be added include an expert system adviser for simulation (to aid novice users), and model validation aids. The overall aim is to make complex modelling and simulation methodologies accessible to a much wider range of engineers and other users than is currently achievable.< >
GRADIENT is a research project within the ESPRIT programme. The project aims to prove design principles, for the development of advanced interface systems for process control. The primary result is a graphical dialogu...
详细信息
GRADIENT is a research project within the ESPRIT programme. The project aims to prove design principles, for the development of advanced interface systems for process control. The primary result is a graphical dialogue environment in which knowledge-based systems and dialogue design combine to provide advanced information processing for complex process control domains. From its inception, the run-time GRADIENT system was conceived to embody a software architecture derived from the philosophy of user interface management systems (UIMS). Within GRADIENT two monitoring and response systems are assigned, one to the process, and another to the operator. The former, a specialised expert system module, called QRES, attacks the problem of intelligent alarm handling. In his actions on the process, the operator is monitored by a plan-recognition system, called RESQ. This knowledge-based system attempts to interpret the operators' likely control objectives, given the process state and his recent activity. Behind these two monitoring advisory systems, GRADIENT provides a support system with more extensive knowledge of the process plant.< >
A project is underway to construct an IKBS for fault diagnosis in a process control system able to cope with both single and multiple faults, and which generate explanation describing the reasoning underlying a partic...
详细信息
A project is underway to construct an IKBS for fault diagnosis in a process control system able to cope with both single and multiple faults, and which generate explanation describing the reasoning underlying a particular diagnosis. The diagnosis system here uses qualitative modelling and deep knowledge techniques in which a behavioural and structural model based on the physics of a system, is formulated instead of using the classical shallow knowledge paradigm (using rules for example). So given knowledge of the input to a device and an understanding of how it is supposed to work, then expectation about its actual behaviour can be generated. Then a comparison between these two, specially the difference between them, provides the foundation of the trouble-shooter.< >
In order to comply with NEI Parsons' CIM strategy, a system was required which would integrate with other systems, particularly the factory information control system (FICS) and the tender estimating system, both ...
详细信息
In order to comply with NEI Parsons' CIM strategy, a system was required which would integrate with other systems, particularly the factory information control system (FICS) and the tender estimating system, both being in-house developments. The measurement and routing system has already been successfully implemented in the turbine blade manufacturing area and is currently supplying FICS with accurate and timely information. The engineer can produce routes in a matter of minutes rather than hours for blade types stored in the system. Further by-products of the system are already becoming apparent. Information on tooling requirements can be produced automatically in readiness for component production to start, information for process planning can be supplied and tender estimates can now be produced to a higher degree of detail within a very short space of time. The system takes the form of an expert system shell and is now being adapted for use in other areas. It is envisaged that many of the manufacturing departments within the company will eventually benefit from its introduction.< >
Explores the underlying basis of the fault finding process for electronic systems and maps the development of an expert system for inertial navigation systems production fault diagnosis. A computer based expert system...
详细信息
Explores the underlying basis of the fault finding process for electronic systems and maps the development of an expert system for inertial navigation systems production fault diagnosis. A computer based expert system has been implemented based on both surface level rules and on a knowledge of structure. General fault finding rules are defined and interviews with a test engineer conducted to provide unit specific fault finding rules. Prolog was chosen for system implementation.< >
A rule basedcontrol system, known as 'RUBIC' (rule-based intelligent control), has been implemented in SmallTalk to run on a personal computer. RUBIC uses fuzzy logic to decipher the vaguely defined terms suc...
详细信息
A rule basedcontrol system, known as 'RUBIC' (rule-based intelligent control), has been implemented in SmallTalk to run on a personal computer. RUBIC uses fuzzy logic to decipher the vaguely defined terms such as 'high, low, OK', etc which are used in the control rules, e.g. IF (Temp is LOW) THEN (make a MEDIUM increase to fuel rate). A graphical interface tool, which is linked to every instance of a rule-based object, has been developed to make the system easy to use. The graphics make extensive use of the SmallTalk graphics primitives. The graphics tool is used for entering the control-rules and the fuzzy sets associated with ill-defined quantities such as LOW, HIGH, etc. RUBIC has been used to control temperature in a model of an exothermic reactor and level in a model of a spherical vessel. It is planned to use RUBIC for temperature control on an exothermic pilot plant reactor.< >
The Conceptual And Relational Database System is a knowledge-based server for very large knowledge-basedapplications supported by a high performance relational database. control of the CARDS engine is maintained by a...
详细信息
The Conceptual And Relational Database System is a knowledge-based server for very large knowledge-basedapplications supported by a high performance relational database. control of the CARDS engine is maintained by a conceptual model interpreter which provides knowledge-based programs with a high level modelling language that supports inference and consistency checking through constraint propagation within a network of inter-related concepts. The project includes the development of a large scale knowledge-based system which provides intelligent support to molecular biologists and cancer research scientists in the analysis and prediction of protein structures from gene and amino acid sequences.< >
To improve the performance of expert systems a deep knowledge approach is recommended, whereby the process is understood in terms of the underlying physical principles and theories. This is important in process contro...
详细信息
To improve the performance of expert systems a deep knowledge approach is recommended, whereby the process is understood in terms of the underlying physical principles and theories. This is important in process control since industrial processes are composed of physical processes. An intelligent knowledgebased system is being developed in Queen Mary College for the tasks of monitoring, fault diagnosis and safety action in process control. The problem considered is to monitor the process behaviour and diagnose any of the following faults in the system: component faults; sensor faults and controller faults.< >
knowledge-based system (KBS) technology is considered as a general framework within which systems and technologies currently used in traffic monitoring and control tasks could be potentially implemented, to improve th...
详细信息
knowledge-based system (KBS) technology is considered as a general framework within which systems and technologies currently used in traffic monitoring and control tasks could be potentially implemented, to improve the functionality of single systems and to achieve broader integration among different (sub-)systems. Many technical problems surrounding KBS technology such as the computational efficiency of available tools, the capabilities of integration with database techniques and conventional numeric processing software, etc., are still open questions. Hence, for many tasks, especially real-time applications, feasible solutions are not readily available. Qualitative considerations regarding potential long-run applications of such techniques.< >
The author outlines the motivation behind ESPRIT project p820: Design and Experimentation of a KBS Development Toolkit for Real-Time Process controlapplications. He is developing a novel architecture capable of repre...
详细信息
The author outlines the motivation behind ESPRIT project p820: Design and Experimentation of a KBS Development Toolkit for Real-Time Process controlapplications. He is developing a novel architecture capable of representing different sources of knowledge, an empirical model that uses observed knowledge of the system's behaviour and an ontological model based on the application of general physical laws to obtain an explicit qualitative model of the process. This partitioning of process knowledge provides a clear separation between the informal experienced based approaches and established formal methods based on the application of appropriate theories. In general, problem solving in complex domains requires a sophisticated and subtle blending of these two sources. The toolkit architecture allows for the blending of these approaches at the task level thereby making explicit the distinction and therefore allowing more general task level problem solvers to be constructed, rather than ad hoc combinations.< >
暂无评论