New Sustainable Operation Method for a Power Grid without an Energy Storage System: A Case Study of a Hospital in Japan

This paper presents a new sustainable operation method for running the power system of a disaster base hospital without the use of an energy storage device. There is a diesel generator for islanded operation in the hospitals in the event of a disaster, but it keeps emptying due to the issue that the fuel stored in the tank deteriorates. In consequence, diesel generators fail to start up and medical services cannot be kept. To prevent fuel deterioration, it is deemed necessary to refuel the tank with occasional use of fuel. Even in hospitals, installing backup power systems like photovoltaics is a common way to reduce energy use. In such hospitals, there is a demand to combine diesel generators and photovoltaics to respond a demand side response. Since it is challenging to operate a complicated system of diesel generators and photovoltaics, it is necessary to install more energy storage system. But several hospitals do not want to install it because energy storage system is so pricey. This paper proposes a method that can correspond demand side response as a virtual power plant in a power grid without an energy storage system to improve the operational issues of a complex combination of diesel generators and photovoltaics. It requires a load prediction first. The prediction method is the load one step ahead prediction and providing the optimized output distribution and rate setting to the diesel generators, stable operation is possible without energy storage system. The proposed method is evaluated by employing a simulation model using the measured photovoltaics output and the actual load at a hospital. As a result, it shows that it can correspond a demand side response of +/- 10 % in the season when the load is low at the hospital with a contract demand 980 kW with 20 % of a photovoltaics. Furthermore, it is clarified that it can correspond a demand side response of +/- 25 % in the season during peak load seasons.