Since operation error is a major factor contributing to industrial disaster, it is necessary to develop safety interlock systems which prevent modes of operation that are known to be dangerous. It is not possible to foresee all the disruptions that might occur and, for this reason, an a priori analysis of safety problems is incomplete. At each step in operation the state of the process must be updated if hazardous conditions are to be avoided. To a large extent this is a problem in logic which can be handled rapidly by the computer using the methods developed here. Systems that handle chemically reactive, toxic, radioactive, and other dangerous materials must be designed and operated with special care to avoid loss of life and property. Provisions must be taken for the reduction of the frequency of equipment and human failure, for the containment or release to a safe area of materials during unintended excursions from normal operation, and for a multitude of other factors which determine the disaster tolerance of a system. Studies of loss prevention have been extensive and a large body of empirical information has evolved. However, major industrial disasters still occur, indicating a critical need for further study (Kletz, 1972; Browning, 1970). DALEAn enormously large number of ways exist for manipulating most industrial processes, many of which are extremely dangerous and a few of which achieve useful processing objectives. In the sample process considered in this report seventeen valves can be opened or closed leading to S7 = 131,072 final valve positions. Considering that the sequence in which the valves are manipulated during transient operation forms an even larger combinatorial problem, it is hopeless to examine the safety of all possible operating problems that might arise. In practice only normal operation, start-up, shut-down and major emergency situations can be examined a priori, leading to the unfortunate possibility that an extremely hazardous mode of operation may be entered into unknown to the operators as they attempt to contend to' other innocuous situations which might arise.We expand on the conjecture that the violent and destructive events which are so prominent in industrial disaster to a large extent are the effects of an inability to anticipate the long-range effects of current actions. Sequences of seemingly innocuous events occur before the first major disruptions, which force the system into a mode of operation from which it cannot be extricated without disaster. Further, could the effects of these events be foreseen, actions could be taken to intercept and quench the impending disaster.The process system is viewed as a network of connectors through which material flows, the routes taken being determined by the position of valves. Symbolic logic is used to model the system and to determine the long-range effects of valve operation policies. Proposed changes in operation can be assessed in matters of seconds by the computer providing an interface between the operator and the process to p...
Methods are developed for the computer-aided synthesis of sequences of valve operations to reach complex operation goals with safety. Given dangerous events which must not occur and operation goals to be reached, sequences of valve openings and closings are formed rapidly for industrially significant problems. and SCOPEThe engineer is guided by experience and intuition during the synthesis of failure-safe operating instructions, and these methods cannot be automated directly. Especially during emergency situations there is a need for computer assistance to speed up and increase the accuracy of the synthesis. This is the first step taken in that direction.The process is modeled as a network of connectors through which material is driven by pressure differences guided by the position of valves. The variety of hazardous conditions to be avoided are given, along with general statements of operating objectives. The synthesis involves the sequencing of valve opening and closings to reach the operation objectives while avoiding the hazards. This forms an enormous combinatorial problem, similar to that found in games such as chess.Synthesis is accomplished by the formation of a hierarchy of goals which identify critical operations and the order in which they are to be performed. Industrially significant problems involving a score or more operations can be handled rapidly and accurately. CONCLUSIONS AND SIGNIFICANCEThis work is significant in that it deals with an important field of engineering problem solving that had received little or no systematic study. It has now been demonstrated that the difficult problems in inductive reasoning associated with operation instruction synthesis can be formulated and solved. Optimism is warranted in the generality and usefulness of this approach.The fields of artificial intelligence and pattern recognition have reached the point where the computer can perform tasks which might be thought to require human intelligence (Newel1 and Simon, 1971;Uhr, 1973). The inductive reasoning used in the synthesis of complex plans of action, such as the development of failure safe operations, in the past has been completely dominated by the human mind. We present our latest developments in the use of the computer in the synthesis of intelligent appearing plans of action (Masso and Rudd, 1969;Lee, Masso, and Rudd, 1970; Siirola, Powers, and Rudd, 1971; Siirola and Rudd, 1971).The computer-aided analysis of proposed industrial operations yielded to sequential logic, leading to the practical solution of the enormous combinatorial problems arising in safety interlock system design (Rivas and Rudd, 1974). Given a proposed sequence of valve operations, the computer can perform the bookkeeping required to detect hazardous operations at speeds sufficient to monitor industrially large processes. However, an equally important and intellectually different problem is the sjmthesis or creation of the operating procedure rather than analysis. Analysis requires deductive reasoning and synthesis appears to require...
Disaster-resistant industrial operations are created to avoid certain classes of events known to be dangerous. While a rapid and accurate synthesis of safe operations is necessary during a disaster, it is hindered by the magnitude of the logic-analysis problems encountered and by the sophistication of the operating goals. This paper develops practical methods for computer-aided synthesis of disaster-resistant operations.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.