AI Makes Waste Generation More Efficient
As a solution to the energy problem in the sustainable development goals (SDGs), waste power generation, which generates electricity from high-temperature and high-pressure steam generated when garbage is burned in incineration facilities, is increasingly attracting the attention of stakeholders. Waste power generation, also known as refuse power generation, has many advantages, including the capacity to reduce the use of fossil fuels, procure fuel waste domestically, and recycle waste collected nationwide as power for local use.
Waste power generation is considered a potential power generation base for local energy production and consumption that could supply the energy used in a region. However, the amount of steam generated varies depending on the nature and shape of the waste used, and it is challenging to control the amount of steam because there are diverse parameters associated with controlling the volume of steam, which in turn makes stable power generation challenging.
NTT Communications Corporation (NTT Com) and Kubota Corporation are jointly conducting a study with the aim of stabilizing steam generation, which is a key condition required before the widespread adoption of waste power generation, using artificial intelligence (artificial intelligence). NTT Com is a wholly-owned subsidiary that handles the long-distance and international communications business of Nippon Telegraph and Telephone Corporation (NTT Corp). In recent years, it has developed various services using AI for corporate clients. For example, in 2016, it concluded the “Partnership Agreement for ICT Innovation in Agriculture, Water, and Environmental Infrastructure” with Kubota, and has conducted various joint experiments using AI.
In the current demonstration experiment, an AI analysis tool, “Node-AI”, developed by NTT Com was used to generate a prediction model for waste incineration, and the process during waste incineration was visualized using a unique analysis technology. By comparing the results of the prediction model with the results of Kubota’s waste incineration expertise, key data could be narrowed down from approximately 300 parameters used to control steam production. After that, an analysis would be carried out to discern the steam volume trends, and a prediction model for the waste incineration status one minute in advance could be generated. A system based on the prediction model will be introduced to waste incineration facilities in operation, enabling operators to monitor the steam volume one minute ahead in real-time.
NTT Com and Kubota aim to achieve more accurate steam volume prediction by generating prediction models for the status 5 minutes and 10 minutes in advance. In addition, by accelerating technological development aimed at stabilizing steam volume, such as the “digital twin”, which digitally models waste incineration facilities similar to the actual operating environments, we will improve the efficiency of renewable energy production for the future. It may not be long before waste power generation becomes adopted as a stable power generation option.