SiemensEPSRC Digital Twin with Data-Driven Predictive Control: Unlocking Flexibility of Industrial Plants for Supporting a Net Zero Electricity System

Lead Research Organisation: Cardiff University


In the net-zero transition of the UK by 2050, electricity demand will increase and more renewable power generation will be installed in industrial plants. The bulk electricity system also faces the challenges of increased total and peak demand, increased difficulty in balancing supply and demand, and increased network issues. The flexibility of industrial plants, i.e., the ability to change the normal electricity generation/consumption patterns, can be utilised to address these challenges, through improving the utilisation of renewable power generation onsite and providing balancing and network services to the bulk electricity system. However, the scheduling and control for tapping this flexibility are subject to great difficulty due to significant uncertainties and computational complexity.

Digital twins are systems of advanced sensing, communication, simulation, optimisation and control technologies, and can provide updating system states and prediction, based on which data-driven approaches can be developed to tackling the uncertainties and computational complexity in scheduling and control. Specifically, a kernel-learning based method is proposed to characterise the uncertainty sets, and an artificial neutral network based method is proposed for predictive control of industrial plants in real-time operation.

A test digital twin platform is established in the lab to demonstrate and assess the proposed data-driven solutions. The platform adopts a two-level structure, with the upper-level global digital twin for whole-plant level predictive control and lower-level local digital twins representing industrial processes, renewable power generation and energy storage systems. The measurements are taken from sensors or a data generator which produces mimic data flow. Two industrial case studies with real data are tested on the platform. One case is an industrial site with a number of bitumen tanks and PV panels, and the other is a paper mill with onsite wind turbines and battery storage.


10 25 50
Description Electrification is one of the key ways to achieve the net zero transition, but brings significant challenges as well, including soaring peak and total electricity demand and increasing impact of the volatility from renewable power generation. Industrial plants, with high energy consumption, are an important part of electrification and can contribute to addressing the challenges in the process of electrification.

The key way in which industrial plants can contribute is to adjust their electricity generation/consumption patterns in response to the needs of themselves and also those of bulk power systems (i.e., unlocking and utilising their flexibility). As a result, the electricity costs as well as the emissions of the industrial plants can be reduced, and the pressure on the bulk power networks can be relieved. Nevertheless, utilising flexibility, via controlling the flexible power generation, energy storage and demand in industrial plants, is faced with two major challenges - uncertainties (in renewable power generation and industrial processes) and computational complexity.

In this project, we found that the challenges can be well addressed via data-driven methods. A support vector clustering method with the generalised intersection kernel was found to be able to characterise the uncertainty set more accurately and effectively, thus better addressing the challenges of uncertainty. An artificial neutral network method was used to generate real-time operational decisions, instead of solving time-consuming optimisation problems, thus better addressing the challenges of computational complexity.

We tested the two methods, tackling uncertainty and computational complexity, in two case studies - an industrial site with multiple separate bitumen tanks and a paper mill involving continuous production process. The two methods performed well in both case studies.
Exploitation Route The findings will be published on academic journals so others can access the outcomes and further develop them in their own ways. In our view, our current artificial neural network method may not very well deal with the situations where the final schedules obtained violate some production constraints. This may be a significant issue for large complex industrial plants with continuous processes (like steelmaking plants). This may be a future research direction.

Furthermore, the research still stays at a low Technology Readiness Level (TRL). Further work can be done to see how the methods can be integrated in real industrial plants considering their existing digital and control infrastructure.
Sectors Digital/Communication/Information Technologies (including Software),Energy,Manufacturing, including Industrial Biotechology

Description A presentation at EPSRC SUSTAIN Future Steel Manufacturing Research Hub 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Industry/Business
Results and Impact The conference was organised by EPSRC SUSTAIN Future Steel Manufacturing Research Hub, and was attended by experts from both the academia and industry in the steelmaking sector. I gave the presentation on how the digital twin and data driven approaches can help steelmaking plants to utilised their flexibility to reduce electricity costs, reduce carbon emissions and also support the net zero transition of the bulk power system. My presentation generated a lot of interests from the audience - we received many questions and discussions during the conference. Liberty Steel also showed interests in further collaborating with us in using our approaches for their plants.
Year(s) Of Engagement Activity 2022