Multilevel Flow Modeling Research Articles

A Proposal to Define and to Treat Alarms in a Supervision Room

Automation offers the possibility of building production systems more and more complex. This evolution leads to a modification of human operator’s activities which sets out problems, particularly when a failure occurs in the process. The first part of the paper presents, after the human operator’s work and manmachine interfaces in a supervision room, a synthesis of different problems linked to alarms. The second part of the paper is a proposal to design and to treat alarms. The goal is to design and display “educational” alarms, well defined and well treated, giving to the human operator the means to perform correctly his work. On that subject, the Multilevel Flow modelling (MFM) for the functional analysis and methods from the dependability science like FMEA (Failure Mode and Effects Analysis) are interesting.

Status and Challenges of Intelligent Plant Control

Intelligent plant control is a rapidly expanding research area that is nourished by the industrial interest in improving the efficiency and safety of plant operations, by the general academic interest in developing theories for control of complex systems and by results from AI research. Currently a variety of techniques for building advanced controls has been developed but there is no general consensus on the meaning of basic terms and there seems to be a limited recognition of the importance of common sense engineering knowledge in realising intelligent controls. In the paper it is argued that common sense knowledge of the purpose of plant components and subsystems play a significant role in building dynamic models, in diagnosis and planning and in the design of supervisory displays for the human supervisor. The lack of understanding and formalisation of common sense engineering knowledge within the control domain limits the levels of plant automation that can be achieved and impede the development of methods for design of supervision and control systems. Recent research results of the author on the modelling of plant goals and functions (Multilevel Flow Modelling) show that modelling research that supplement and expands the traditional modelling paradigms of control theory may lead to significant advances in the design of intelligent plant controls.

マルチレベルフローモデリングの概念を用いた運転員メンタルモデルの提案

マルチレベルフローモデリングの概念を用いた運転員メンタルモデルの提案

Combining Multilevel Flow Modeling and Hybrid Phenomena Theory for Efficient Design of Engineering Systems

A combination of Multilevel Flow Modeling (MFM) and Hybrid Phenomena Theory (HPT) is proposed for efficient design of engineering systems. The MFM is a methodology to model an engineering system from the standpoint of means and goals. One useful application of the HPT is to derive mathematical equations describing a system's behavior. A conceptual design procedure composes of 1) goal decomposition. 2) listing up design constraints. 3) function assignment, 4) finding out suitable parts/components and their configuration. and 5) examination of the design's behavior. The paper presents extensions to MFM representing detailed information about goals and parts/components configuration so as to be applied to the steps 1) to 3). The HPT is applied for the step 5). Two design examples by the proposed combination are discussed.

A Simulation, Environment for Eyaluation of Kowledge Based Fault Diagnosis Systems

In the paper a simulation environment for evaluation of knowledge based fault diagnostic systems is presented. The contribution deals with the concept of the environment that provides an integration of qualitative and quantitative models. This concept is presented on a simulated water-column process controlled by feedback. The simulation environment is used to evaluate the potentials of Multilevel Flow Modelling (MFM) approach to system diagnosis. The environment is realized within the real-time expert system tool G2.

Diagnostic reasoning strategies for means-end models

This paper describes three diagnostic methods for use with industrial processes. They are measurement validation, alarm analysis and fault diagnosis. Measurement validation means consistency checking of sensor and measurement values using any redundancy of instrumentation. Alarm analysis is the analysis of multiple alarm situations to find which alarms are directly connected to primary faults and which alarms are consequential effects of the primary ones. Finally, fault diagnosis is a search for the causes of and remedies for faults. The three methods use multilevel flow models (MFM), to describe the target process. They have been implemented in the programming tool G2, and successfully tested on simulations of two processes.

MODELING GOALS AND FUNCTIONS OF COMPLEX INDUSTRIAL PLANTS

Abstract The purpose of this paper is to describe a modeling methodology called multilevel flow modeling (MFM), which the author has developed for the representation of goals and functions of complex industrial plants. The idea of the methodology is to apply functional concepts to represent a plant at multiple interrelated levels of abstraction. MFM is currently used in supervisory control applications for aiding the operator in diagnosis and planning. It is also used in the conceptual analysis and synthesis of control systems. The paper provides an introduction to the basic concepts of MFM, details two modeling examples, and describes the object-oriented tool Abstractions used for the implementation of MFM models and for diagnosis and planning applications. Finally, the paper presents a review of the use of MFM in previous and ongoing international projects.

Diagnostic Methods for Means-end Models

This paper describes three diagnostic methods for use with industrial processes. They are measurement validation, i.e., consistency checking of sensor and measurement values using any redundancy of instrumentation; alarm analysis, i.e., analysis of a multiple alarm situation to find which alarms are directly connected to primary faults and which alarms are consequential effects of the primary ones; and fault diagnosis, i.e., a search for the causes of and remedies for faults. The three methods use multilevel flow models, (MFM), to describe the target process. They have been implemented in the programming tool G2, and successfully tested on two small processes

015 Model-based fault diagnosis using MFM

Abstract Expert systems have been used with great success in fault diagnosis. However, a large rule database is not easy neither to build, update, nor document. This paper presents a method for building knowledge databases based on an explicit means-end model of the target process. The modeling technique used is multilevel flow modeling (MFM), as described by Lind (1990b). This allows the knowledge database to be built, updated, and documented in a graphical language, with internal consistency guaranteed. First, the basics of MFM are described, and then an example of how such a model can be used in fault diagnosis is given.

Human-computer interaction and complex systems : Weir, G R S and Alty, J L (Eds). Academic Press, London, 1991, 292 pp, 256 refs

G.R.S. Weir, Living with Complex Interactive Systems. J.L. Alty and R.A. Ritchie, An Algebraic Approach to Interface Specification Analysis. F. Decortis, V. de Keyser, P.C. Cacciabue, and G. Volta, The Temporal Dimension of Man*b1Machine Interaction. E. Hollnagel, The Phenotype of Erroneous Actions: Implications for HCI Design. P.A. Booth, Modelling the User: User-System Errors and Predictive Grammars. S. Grant and T. Mayes, Cognitive Task Analysis? D.D. Woods, The Cognitive Engineering of Problem Representations. L. Bainbridge, Multiplexed VDT Display System: A Framework for Good Practice. B.-B. Borys, Ways of Supporting Ergonomically and Technically Correct Display Design. M. Lind, Representations and Abstractions for Interface Design Using Multilevel Flow Modelling. E. Edmonds and J. Ghazilkhanian, Cooperation between Distributed Knowledge-Bases and the User. P. J*aearvinen, Human*b1Computer Interaction Research in the Light of Aulin's Foundations of Mathematical Systems Dynamics.

Model-Based Fault Diagnosis using MFM

Abstract Expert systems have been used with great success in fault diagnosis. However, a large rule database is not easy neither to build, update, nor document. This paper presents a method for building knowledge databases based on an explicit means-end model of the target process. The modeling technique used is multilevel flow modeling (MFM), as described by Lind (1990b). This allows the knowledge database to be built, updated, and documented in a graphical language, with internal consistency guaranteed. First, the basics of MFM are described, and then an example of how such a model can be used in fault diagnosis is given.

Model-Based Measurement Validation using MFM

Abstract Most industrial processes are equipped with a large number of sensors, of which several directly or indirectly measure the same variables. Especially when material and energy balance equations are taken into account, the total set of measurements commonly gives rise to redundancy, which can be used to check the consistency of the signals, i.e., to validate them. This paper presents a method for measurement validation, which is fairly general and built on model-based reasoning. The modeling technique used is multilevel flow modeling (MFM), as described by Lind (1990b). First, the basics of MFM are described, and then an example of how such a model can be used in validation is given.

Use of Abstractions within Fault Diagnosis of Shipssystems

This paper describes a fault diagnosis system developed for ships. The system is designed to support the operator in dealing with malfunctions in a ships machinery installation. The scope of the diagnosis system is to diagnose malfunctions in the ships auxiliary and ancillary systems. Processes internal to the main engine are not within the scope. The main engine is thus handled as a single component. The purpose of the diagnosis system is not only to detect malfunctions in terms of possibly faulty components, but also to assist in determining consequences of the detected malfunctions and to suggest appropriate actions according to the situation. The diagnosis system is based on an intentional model of the ships machinery installation. The model is described in a language called Multilevel Flow Modelling (MFM) which support derivation of statements about failure to achieve system intentions. MFM describes causal and continuity relationships in the physical process in terms of flow of mass and energy. Reasoning within the model is based on qualitative values given in terms of normality. These qualitative values are derived by means of a filter which continuously calculates RMS values of measurements. The system is implemented in C++, executes under the Unix operating system, connects to a conventional alarm system and has an X-Windows based user interface.

Model-based alarm analysis using MFM

Most industrial processes are equipped with a large number of alarms. In a failure state it is quite usual that many of the alarms will trigger. Some of them will be directly connected to the primary sources of error, but others may be secondary, i.e., not connected to any failed equipment, but due only to consequential effects of the primary failures. In a failure state it is vital for the operator to separate the primary from the secondary alarms. This paper describes a new method for automatically recognizing the primary failures. It is fairly general and built upon model-based reasoning. The modeling technique used is multilevel flow models (MFM), as described by Lind (1990b). First, the basics of MFM are described and then an example of how such a model can be used in alarm analysis is given.