GOALI: Coordination of Multi-Stakeholder Process Networks in a Highly Electrified Chemical Industry

点击翻译按钮获取中文摘要

The chemical sector is currently the largest industrial consumer of oil and gas and a major emitter of greenhouse gases. As the demand for chemicals continues to grow, there is an urgent need to decarbonize global chemical production. Electrification using power-to-chemicals and power-to-heat technologies is considered one of the most promising emission reduction strategies for the chemical industry, provided that renewable sources of electricity are used. As electrification efforts intensify in the coming years, the chemical industries will see a dramatic increase in the number of chemical processes that consume large amounts of electricity. These processes will face the challenge of increasingly time-sensitive availability and pricing of electricity, such that significant operational flexibility will be required to ensure safe and cost-effective operation. A major source of flexibility lies in the coordination of multiple interconnected plants (and the respective multiple self-interested stakeholders). Importantly, with a large number of power-intensive processes, effective coordination must happen in real time, much akin to the way modern power grids operate. This would, however, be a major paradigm change for the chemical industry, which is used to highly steady operation and long response times, where different companies coordinate their operations mainly through long-term bilateral contracts. In this research, the research team aims to develop efficient computational methods for the coordination of multi-stakeholder process networks that improve the whole system's performance while ensuring that all stakeholders benefit from cooperation. To this end, new fair and privacy-preserving coordination mechanisms will be designed that can consider multiple products, plants, stakeholders, and spatiotemporal scales. Implemented at large scale, such coordination will not only enable the electrification of the chemical industry in the most cost-effective manner but also help improve grid reliability and facilitate further growth in renewable energy generation. This project is jointly funded by the Electrochemical Systems and the Process Systems, Reaction Engineering, and Molecular Thermodynamics programs of ENG/CBET.The proposed research plan is organized into three specific aims: (i) Distributed cooperative industrial demand response (DR), where DR refers to the adjustment of an electricity consumer's load profile in response to price changes. The researchers will develop capabilities for the distributed coordinated optimization of multiple connected plants (possibly across an entire supply chain), which maximizes the overall benefits from DR while accounting for local objectives, data privacy requirements, and fair profit allocation. (ii) Risk-aware cooperative provision of interruptible load, which aims to further incorporate uncertainty and risk measures to consider participating in the reserve market through the provision of interruptible load. Here, the major technical challenge is the computationally efficient modeling of endogenous uncertainty and multistage recourse. (iii) Market-based coordination of an electrified chemical industry, where an alternative approach to distribution optimization for real-time coordination will be explored. Specifically, a responsive and robust market mechanism will be developed based on the idea of trading operational flexibility. This project will be conducted in a close industry-university collaboration between the University of Minnesota and The Dow Chemical Company, which is crucial as it requires both the development of new computational methods at the cutting edge of academic research as well as their application to industrial use cases to gain practical insights and demonstrate the value of the proposed approaches. This research will advance the state of the art in the broad area of multi-agent decision making. Although this work will focus on a specific manufacturing context, the methods will be sufficiently general such that they can be adapted for many other applications. Generally, they can be applied to multi-stakeholder systems in which conditions and resource availability are highly time-sensitive and uncertain.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.