Blog

Enhancing energy forecasting and optimisation: Delta Lab update

This article explores the research and developments from Powerstar’s innovation team within the Delta Lab, specifically, an energy forecasting and optimisation project underway by Knowledge Transfer Partnership (KTP) Associate, Dr. Rui Zheng.

A knowledge transfer partnership (KTP) is a three-way relationship between a company, a university, and an individual that enables Dr Rui Zheng to have guidance from Cardiff University colleagues, access to cutting edge research, and support from his employer, Powerstar.

The requirement for change

Powerstar’s ongoing development of an enhanced Energy Management System (EOS) capable of predicting, forecasting, and optimising energy flows for businesses is driven by the growth of renewable energy generation, enhanced requirement for power resilience, and greater uptake of smart microgrids.

In 2019 alone, production of renewable energy rose by 4.9%, and by December 2019, low-carbon generation (renewables & nuclear) reached 51.6% of the total[1]. Whilst this assists with decarbonising electricity generation, the intermittency of renewables presents new challenges in balancing energy supply and demand for the National Grid.

The imbalances caused by renewable sources connected to the grid, alongside overall growing demand for energy, and reliance on pan-European imported energy has led to a number of near-misses and power failures in the UK, including the blackout in August 2019, which resulted in millions of people, residents and businesses alike, being impacted.

This has led numerous businesses across all industries to install multiple flexible energy technologies such as on-site renewable energy generation, and electric vehicle charging assets combined with battery energy storage systems to form microgrids. This enables businesses to reduce reliance on the grid, enhance power resilience, and meet net-zero targets.

To balance the ever-changing priorities of sites, including supporting electric vehicle charging, maximising on-site generation, negating power failures, and lowering carbon emissions, greater innovation is required that could enable all connected technologies to operate as effectively and efficiently as possible in unison, with a site’s energy needs being continuously identified and met.

Gaining flexibility through systems

Traditionally, energy management systems (EMS) operate based on pre-defined algorithms to determine how asset operation responds to pre-determined needs. However, with the changing nature and increasing complexity of the grid and business needs, demand has grown for smarter alternatives less reliant on pre-set orders, especially when dealing with flexible assets with the capabilities to respond in real-time to changing requirements.

One such flexible asset, battery energy storage systems, require accurate, up-to-date and relevant data sets to be able to perform optimally, an area which has improved with the rise of big data and expansion of the internet of things (IoT)[2], enabling the collection of a greater volume of better quality data, and by processing this through an intelligent algorithm, the actions of the assets can be continually optimised.

In recent years, the adoption of electric vehicles (EVs) has increased with figures suggesting that by the end of 2019 there were over 275,000 vehicles on the road in the UK[3]. Whilst the charging infrastructure for these vehicles expands, to avoid significant costs associated with upgrading supply capacity many businesses are turning to electric vehicle charging supported by battery energy storage, known as battery buffered charging.

Driving the future of flexibility

Noticing this trend and realising the opportunity to enhance the flexibility of battery buffered charging, Dr. Rui Zheng has managed to further develop Powerstar’s Energy Optimisation Systems software to enable numerous data sources from a multitude of connected assets to be gathered but, most importantly, utilised quickly and intelligently. With the assistance of artificial intelligence, Dr Zheng’s software is able to gather insights from the connected hardware’s use to further optimise the output, a particularly important factor when user interaction is a factor, as with battery buffered charging.

The crucial aspects of such systems are their ability to identify and prioritise a variety of needs, such as optimising power flow to achieve the best revenues or savings opportunities and the ability to enter island mode to negate a power failure, therefore enhancing a sites power resilience. Through the aforementioned signalling and AI learning, the EOS is able to do exactly that.

The project is currently still in its development stages but has been created using leading modelling software, Typhoon HIL, to examine how the design performs in virtual scenarios as well as enabling a prototype hardware version to be tested at Powerstar’s head office and production facility in Sheffield. The results are currently being analysed to assess the accuracy and reliability of the system. Once this modelling and evaluation is complete, the system will be refined further and a project report delivered alongside a final prototype of the system for commercial use.

Forward thinking

As Dr. Zheng’s research project is nearing its latter stages, he is already considering how it can be developed further in the future. As battery energy storage systems grow increasingly important, with a broader range of applications, forecasting and optimisation of energy flows across an entire site will continue to evolve, and he will endeavour to develop new elements to ensure the system is prepared for these developments.

His work has already drawn significant interest and, in 2019, Dr. Zheng was a key note speaker at the Intelligent Transport System Young Researchers Conference (ITSYRC). Hosted by the Transport Futures Research Network (TFRN), a cross-disciplinary network for future transport and mobility researchers at Cardiff University. The conference aims to boost understanding of new and emerging transport technologies from both academic and industrial aspects, including electric vehicles, autonomous vehicles and mobility as a service.

Dr Zheng’s presentation covered his project, explaining Powerstar’s battery buffered charging system, the principle of his project and the challenges ahead before demonstrating how the latest AI technology, like machine learning, can be applied in the EV technologies.

[1] https://www.edie.net/news/10/Low-carbon-generation-edges-past-50–share-of-UK-s-power-demand/

[2] https://www.ibm.com/blogs/internet-of-things/what-is-the-iot/

[3] https://www.nextgreencar.com/electric-cars/statistics/