OOP techniques for control systems software design

The process of software design for nuclear power plants has several stages that must be fulfilled for a successful application. The first deals with the suitability of the concepts being applied to control the process itself, i.e. the control theory approach used, rationale for imposed limits, etc., that could be termed the formal specification generation. The second encompasses the processes of insuring the correctness of the software according to its formal specification. There are several approaches to dealing with these issues being developed worldwide, from the purely analytical to exhaustive testing methods. The purpose of our research effort is to explore and develop new tools for designing and testing control algorithms based in object-oriented programming (OOP) techniques and iconic interfaces. These two technologies will empower the designer by providing powerful yet simple means to introduce faults, investigate their effects and introduce configuration changes. Software development approaches that both encourage and force modularity with explicit interfacing definition could be of great help towards automatic software verification techniques. It is in this context that OOP techniques could play an important role, since they could be thought of as independent units interacting through allowed communication channels. Object classes implementing most common algorithms have been developed and used in our research laboratory computing environment with good results. It is feasible, in principle at least, to design a topology advisor/verifier that parses the control algorithm to detect anomalous configurations. Future research topics include the development of a prototype topology parser to explore its potential applicability in software design. (Author) 6 refs., 2 figs