Solar Radiation Consortium's analysis of the amount of solar irradiation on the ground surface (source: Weathernews)
[Click to enlarge image]

Weathernews Inc announced Jan 31, 2018, that it has started to offer the "new solar power generation forecast" service to energy-related companies, providers of their services and IoT platform-related firms.

Using the latest physical meteorology, the service provides the forecast amounts of solar power generation and solar irradiation based on the light scattering/absorption caused by clouds with a resolution of 1km (mesh) in real time.

The previous solar power generation forecast service predicts the amount of solar irradiation by using a regression analysis model that performs statistical calculation based on data observed on the ground. And, based on the forecast amount of solar irradiation, the company provides the forecast amount of solar power generation up to seven days ahead.

However, with those conventional forecasting methods, the scattering/absorption of sunlight caused by cloud particles, which are the most uncertain element, is not taken into consideration. Also, the statistic values used for a regression analysis model depend on trends at observation points and the period of the statistic, making it impossible to calculate optimum values by considering real-time climate phenomena into consideration.

This time, Solar Radiation Consortium, a specified nonprofit corporation participated in by Weathernews, developed a cloud analysis algorithm that estimates the thickness of cloud, the amount of cloud water and cloud-top temperature by using satellite images and succeeded in increasing the speed of a radiative transfer model, enabling to analyze the amount of solar irradiation with high accuracy in real time.

As a result, it became possible to take in the analyzed actual status values of the amount of solar irradiation as "planes" in consideration of the light absorption/scattering caused by the cloud particles, air contaminants and yellow sand contained in clouds without depending on statistic values at specific observation points.

In a test conducted at a solar power plant, Weathernews confirmed that the amount of solar power generation estimated by the model has a good correlation with the actual amount of electricity generated. The company aims to further improve forecast accuracy by widely introducing and operating the service both in and outside Japan.