Abstract :
The current General Electric Boiling Water Reactor process computer is a conventionally architectured, medium sized, monolithic system based upon the GE-PAC* 4010. The system performs scan, log, and alarm functions, as well as some rather complicated nuclear steam supply and turbine generator performance calculations. However, over the years, the G.E. system designers in San Jose have become aware of both the limitations of the monolithic architecture as applied to the rapidly growing nuclear applications as well as the recent tremendous advances in computer hardware technology. What has emerged is the concept of a distributed computer control system primarily designed for the G.E. BWR, but also ideally suited for other applications. Built around a high speed common data base memory, the distributed system yields a giant improvement in terms of overall computer system modularity, software simplicity, versatility, price/performance ratio, availability, and functional capability. The G.E. design is built around two major changes in philosophy in terms of applying process computers to an application: (1) Force the computer architecture to fit the man and the application, and (2) Increase the hardware content and complexity whenever it is consistent with reducing the software complexity and cost.
