Dairies could slash cleaning costs with AI cleaning system
The group estimates that the food industry in the UK alone could save £100m ($132m) a year.
The group working on the technology say that in addition to the savings, the AI-driven monitoring system could lead to greater production capacity and cheaper food prices for consumers.
Food and drink production is the largest manufacturing sector in the UK, and the highest industrial user of water, at approximately 430m liters a day.
Cleaning machinery is imprecise
Cleaning accounts for 30% of energy and water use, and leads to down time and over-use of chemicals, at huge cost to manufacturers and the environment.
As current technologies cannot accurately determine exactly how dirty food and drink processing equipment is inside, cleaning to minimize food safety risks can last up to five hours a day.
The research project, led by Martec of Whitwell Ltd, in collaboration with the University of Nottingham and Loughborough University, has secured Innovate UK funding for the project.
Martec specializes in the design, installation and use of Clean-in-Place (CIP) and hygienic technologies in food and pharmaceutical manufacturing.
CIP operates 'blind'
Dr Nik Watson, assistant professor and chemical engineer specializing in food measurement systems is leading the University of Nottingham team.
Watson said that to prevent product contamination, many food and drink manufacturers use a non-invasive, CIP system to wash inside food processing equipment without disassembling it.
“As CIP has to operate ‘blind’, it is designed for the worst case scenario. In daily use this often results in the over-cleaning of production lines,” Watson said.
The research team will design and build a lab-scale experimental facility to reproduce common industrial cleaning problems in a typical food-processing plant, and run tests using various foods.
They will also assess the potential for an AI inspection system to measure precisely how much food residue and microbial debris is left inside the rig. Researchers will test ultrasonic sensing and optical fluorescence imaging technologies in comparison with existing detection methods to determine the best results.
Developing SOCIP software
Watson is working alongside Dr Elliot Woolley from Loughborough University. The two university partners have scientific expertise and industrial application experience in ultrasonic and optical sensing technologies respectively.
The year-long study will go on to develop bespoke software to process the sensor data results and generate algorithms for an AI-based monitoring system. This system will be able to autonomously optimize the cleaning process in plant equipment in real-time.
The resulting Self-Optimising-Clean-In-Place (SOCIP) system will be a world-first.
Watson said the project aims to reduce cleaning time and resource use by between 20% and 40%.
The technology could potentially be retrofitted on to existing CIP systems or incorporated into new installations, increasing its market potential.
The researchers point to the fact that of almost 9,000 UK manufacturers identified by the Food and Drink Federation, 1,000 plants currently use some type of CIP. Retrofitting those sites alone gives rise to a £50m market opportunity.
Potential for savings
Ian Sterritt, co-owner and director of Martec said the technology will likely become widely adopted because it can be applied to new installations and existing facilities.
The group says SOCIP requires no expertise and uses off-the-shelf electronic components, making it attractive to smaller users, and it would be a major cost-cutting technology for the food and drink industry with spin-out applications in other sectors such as pharmaceutical, FMCG and cosmetics.
For a medium-sized dairy, cleaning typically costs £1m a year with loss of production time responsible for at least half of that cost.
Using SOCIP on a dairy of this size is estimated to reduce annual water usage by 270,000 liters and energy consumption by 2,400 megawatt-hour, leading to net savings of £300,000 ($396,000) a year.
