Site: TOKYO ELECTRIC POWER COMPANY (TEPCO)
Tokyo 100, Japan
Date Visited: March 27, 1992
TEPCO is the largest privately owned electric utility in the world, and is one of ten electric power companies in Japan. Total revenues for the fiscal year ending March 1991 were 4.4 trillion yen. TEPCO's generating capacity is over 46 billion watts, and it services over 22 million customers. There are approximately 40,000 employees. Our hosts were from within two departments of the Engineering Research & Development Administration (approximately 400 people, with an annual budget of about 68 billion yen): the R&D Planning Department and the AI Technology Department within the Computer & Communication Research Center (C&CRC). There are approximately 100 people in the C&CRC. The AI Technology Department employs 14 technical staff.
Our hosts prepared a very thorough response to the questionnaire. They distributed a report on AI applications within TEPCO, plus a table listing all their application systems -- a total of 30 -- including the purpose of the system, stage of development, AI tool or language used, and size of the KB (see Table TEPCO.1). We were also given copies of about 17 technical papers that have been published and presented at various conferences, symposia and workshops.
The R&D strategy of the AI Technology Dept. has three thrusts:
The department is actively pursuing fuzzy logic, neural networks, genetic algorithms and computer graphics in addition to expert systems. Our hosts made it clear that at TEPCO "AI" means not only Artificial Intelligence but also Advanced Information Technology.
Most of the technical discussion focussed on one particular application, a system for forecasting maximum daily load on the power system.
TEPCO has developed 30 systems, of which 11 are in routine use. The application domains for these 11 include design, consultation, control, prediction, planning and scheduling, fault location, hot-line service, and computer operations. Three systems are in field test, 14 in the prototyping, and two in the feasibility stage. The most successful system is the daily maximum load forecasting system. Measures of success have been user satisfaction, a reduction in the absolute forecasting error rate from 2.2 percent to 1.5 percent, and a four-fold speedup in forecast generation. The system is actually quite small, with only about 100 rules, and was built using Toshiba's TDES3 tool. It runs on a Toshiba minicomputer and also on Toshiba workstations (Sun workstation compatible). The forecasting system is one component of and integrated with a much larger load forecasting system called ELDAC. The system was developed at a cost of approximately $2 million over a period of about 20 months. Two researchers and two experts at TEPCO designed the system, and three system engineers from Toshiba built it. It is now used routinely by load dispatchers. Although the ROI is difficult to estimate, the use of the system precludes the need for a standby generator at a power station.
About 50 percent of TEPCO's ES projects get from the prototype stage to an operational system.
TEPCO's AI group is not formally trained. Some of the people were transferred in, e.g., the young woman engineer, Ms. Hiroko Miyamoto, who built the AI part of the load forecasting system. They have a 14 person group in the lab.
The systems they have built are typically small, in the range of 100 to 300 rules. Applications are usually built by the people who need them (at least by the organization that needs them), with help from the lab if needed.
Expert System Applications at Tokyo Electric Power Company