81
4.2.4 Increased Sophistication of Demand Forecasting
by Utilizing Big Data
Approximately 300 million tons of fossil resources are imported to Japan. Roughly,
60% go to oil refineries, 25% to power plants, 5% to gas companies, and the remainder represents coal for iron making. Of these, oil refineries, power plants, and gas
companies do not aim to consume energy themselves. It is the daily life (transportation, households, and businesses) and monozukuri sectors mentioned earlier that
actually consume energy. Oil refineries, power plants, and gas companies convert
energy into different forms so that consumers can use it easily, so they are called the
energy conversion sector.
It is preferable to convert fossil resources without loss in the energy conversion
sector, but some energy is consumed here. In the case of thermal power generation,
the turbine is rotated by the power of steam generated by burning fuel, but since
steam does not flow unless one direction is of lower pressure, there is a step in
which steam is liquefied in a condenser. About 60% of the total energy from fuel
burning is released to the sea and the atmosphere here. Japan’s thermal power boasts
the world’s top efficiency, but the loss is still not small. Loss occurs other than in
thermal power generation. Power generation loss, daily life, and monozukuri each
account for about one third of the total energy consumption.
Besides improving power generation efficiency, power consumption prediction
is also important to reduce power generation losses. Electric power companies
adjust the amount of electricity to be shipped by predicting the power consumption
based on past power consumption conditions and weather forecasts (temperature
predictions). If the amount of electricity generated is larger than the amount consumed, we will discard unused electricity wastefully, but if the electricity generation
is less than the amount consumed we will face supply shortage, and society will fall
into turmoil. Therefore, electric power companies must constantly generate electricity to a level higher than the predicted power consumption. It is troublesome not to
be able to store the electricity once generated.
A research team of Center for Low Carbon Society Strategy (LCS) investigated
the power consumption prediction in areas served by TEPCO and the weather forecast data of the meteorological observatory and found that there is an error of several percent. In order to eliminate this error, the accuracy of both the weather
forecast model and the power consumption prediction model must be increased. If
the weather forecast model improves, it will be easier to set up the power generation
plan, and if the accuracy of the demand forecast increases, it will be easier to control
power transmission. If prediction accuracy improves, it is expected to reduce energy
loss and reduce cost in hundreds of billions of yen, so we should develop highly
accurate prediction models using artificial intelligence (AI) and big data (Fig. 4.5).
4.2 Innovations Emerging from Theory and IT
4.2.4 Increased Sophistication of Demand Forecasting
by Utilizing Big Data
Approximately 300 million tons of fossil resources are imported to Japan. Roughly,
60% go to oil refineries, 25% to power plants, 5% to gas companies, and the remainder represents coal for iron making. Of these, oil refineries, power plants, and gas
companies do not aim to consume energy themselves. It is the daily life (transportation, households, and businesses) and monozukuri sectors mentioned earlier that
actually consume energy. Oil refineries, power plants, and gas companies convert
energy into different forms so that consumers can use it easily, so they are called the
energy conversion sector.
It is preferable to convert fossil resources without loss in the energy conversion
sector, but some energy is consumed here. In the case of thermal power generation,
the turbine is rotated by the power of steam generated by burning fuel, but since
steam does not flow unless one direction is of lower pressure, there is a step in
which steam is liquefied in a condenser. About 60% of the total energy from fuel
burning is released to the sea and the atmosphere here. Japan’s thermal power boasts
the world’s top efficiency, but the loss is still not small. Loss occurs other than in
thermal power generation. Power generation loss, daily life, and monozukuri each
account for about one third of the total energy consumption.
Besides improving power generation efficiency, power consumption prediction
is also important to reduce power generation losses. Electric power companies
adjust the amount of electricity to be shipped by predicting the power consumption
based on past power consumption conditions and weather forecasts (temperature
predictions). If the amount of electricity generated is larger than the amount consumed, we will discard unused electricity wastefully, but if the electricity generation
is less than the amount consumed we will face supply shortage, and society will fall
into turmoil. Therefore, electric power companies must constantly generate electricity to a level higher than the predicted power consumption. It is troublesome not to
be able to store the electricity once generated.
A research team of Center for Low Carbon Society Strategy (LCS) investigated
the power consumption prediction in areas served by TEPCO and the weather forecast data of the meteorological observatory and found that there is an error of several percent. In order to eliminate this error, the accuracy of both the weather
forecast model and the power consumption prediction model must be increased. If
the weather forecast model improves, it will be easier to set up the power generation
plan, and if the accuracy of the demand forecast increases, it will be easier to control
power transmission. If prediction accuracy improves, it is expected to reduce energy
loss and reduce cost in hundreds of billions of yen, so we should develop highly
accurate prediction models using artificial intelligence (AI) and big data (Fig. 4.5).
4.2 Innovations Emerging from Theory and IT