Innovating the solid waste processing based on AI (iWASTE)

The disposal of SW has become a significant challenge to achieve sustainable development, and waste incineration is recognized as a major measure due to its advantages of recovering energy and other resources from waste and avoiding landfilling to minimize other environmental risks. In Sweden, SW has been commonly used as fuel in combined heat and power (CHP) plants. Both the operation optimization and the control of pollutant emission rely on accurate information on SW properties, such as element compositions, heating values and mass flowrates. However, SW is characterized by its diversity, which covers various plastics, composite materials, organic materials such as wood, paper and green waste, mineral fractions etc. Such a diversity makes it extremely difficult to obtain the properties. It is of great importance to develop a method that can be used online for waste recognition.

The main tasks of this project includes: (I) adopting regenerative AI to generate data that can be used for training the image processing model, which is to tackle the barrier of data lacking; (II) developing novel hybrid models based on image processing and near infrared (NIR) spectroscopy that can improve the accuracy of the waste classification; and (III) developing advanced optimization algorithms based deep reinforcement learning to achieve multi-objective optimization and multi-level control by using MPC that can increase the profit of waste fired power plants.

This project will be implemented through a close collaboration between EST and IDT, between MDU and Hong Kong Polytechnic University, and between academia and industry. It is built on our previous research projects with Mälarenergi and Eskilstuna Energi och Miljö on using solid waste in CHP plants. The findings can also be further extended to waste sorting and recycling for other purposes.

Project objectives

The objective of this project is to develop new solutions for solid waste (SW) classification and utilization based on advanced AI technologies, such as regenerative AI, image recognition and reinforcement learning, that can improve the energy efficiency and reduce the operating cost and emission of pollutants when using SW in CHP plants.