Artificial intelligence (AI) is growing in use across a variety of sectors, and manufacturing is no different. The value-added AI provide in safety, efficiency and cost savings is incredibly valuable for any business.
The manufacturing industry is constantly under pressure to increase productivity to remain profitable and at the same time prioritise safety whilst also maximising productivity. This requires a delicate balance and keen oversight to ensure equipment is operating optimally and effectively, and that staff are working both safely and efficiently. Manufacturing presents a challenging predicament, requiring firms to emphasize improving efficiency whilst also ensuring the health and safety of workers remains a priority.
In particular, the manufacturing industry is susceptible to workplace accidents, machinery problems, and other issues that impact productivity. The total impact of health and safety failings on economic cost in 2016/17 was estimated at £1.1 billion in the manufacturing sector, with £580m of this due to injury.
Manufacturing spans an array of products and businesses, meaning across the sector a variety of distinctive operational practices are in place. Each business will have safety risks unique to their specific set-ups, meaning mitigation and prevention approaches have to be adaptable and flexible to each businesses’ requirements.
AI has the power to transform factories and is positively impacting manufacturing by increasing efficiency and improving health and safety. Through the use of AI real-time visualization, manufacturers can more easily monitor equipment and staff simultaneously, and more reliably than human monitors. The use of AI can provide a smart solution to provide constant and consistent digitisation of their shop floor – allowing for persistent monitoring and alert mechanisms. ThingTrax has developed the Factory Smart Vision System (FSVS) to do just that.
AI and Machinery
The Factory Smart Vision System is an industrial digital technology (IDTs) project focused on the use of Artificial Intelligence (AI) to create a smart, connected factory through the use of machine vision-based solutions that are better able to monitor and alert to production and safety issues in real-time. Using AI to provide firms with real-time visualization of workers and processes using edge computers and cameras, FSVS alerts when needed to enable quick action, saving time and money, preventing costly and dangerous errors.
AI Digital Manufacturing Platform models helps power ThingTrax’s vision devices. Using AI on Edge, ThingTrax can analyse what is occurring in the physical environment around a machine, as well as use PPE detection, hazardous condition detection, sudden fall detection, and fire and smoke detection to help manufacturers monitor their operations in real-time. Deep learning-based machine vision software helps to minimize human intervention, providing real-time solutions by distinguishing the acceptable variations in products and the defects in manufacturing industries.
Health and Safety
Through the use of ThingTrax solutions, manufacturers will be able to receive warnings, alerting them before an accident is about to happen. Such advanced notifications allow the teams involved to intervene in time to avert issues. The data gathered through this technique can also simulate the skills matrix of employees, enabling management to deploy the right resource to the right job, avoiding accidents and injuries.
This is especially useful, given that manufacturing has a higher rate of fatal injury (around 1.5x the average rate across all industries, according to HSE). Traditional approaches to manually monitoring factory shop floors and manual reporting and tracking are costly and impractical, so innovations in this area are greatly needed across the sector. ThingTrax’s piloted AI solution indicates a reduced risk of HSE incidents and compliance failings by up to 90% through the use of effective live monitoring, reporting and trackable action recording. Thus, AI provides a key social impact through improved monitoring and prevention of health and safety incidents.
In the simplest terms: AI works better than humans to reduce health and safety compliance monitoring costs, lone worker requirements, and avoid the tedious and slow reporting and tracking of compliance actions.
AI allows for improved data and insight into operations in real-time, providing ready access to information that simply isn’t feasible from staff. Typically, manufacturers are able to capture Overall Equipment Effectiveness (OEE) in a mostly automated way but Overall Labour Effectiveness (OLE) is more difficult to monitor. Specifically capturing accurate labour cost on each job is critical data that is often missing, as it can only be captured by manual field monitoring and recording. A costly and unreliable method.
Using various pre-trained object detection models to detect people and other commonly used objects (not industrial), ThingTrax created a pipeline of the AI model to achieve calculation of dwell time, PPE detection, and other manufacturing-related needs. The results highlighted our human detection and tracking model ran with almost 70% accuracy.
Manufacturing companies regularly incur high costs for employing monitoring operators, and quality assurance and support staff. ThingTrax’s use of AI, machine vision and deep learning allows for analysis of real-time production data and historical data to make production planning more accurate and real. This valuable insight helps manufacturers produce more orders per machine, deliver goods on time, improve quality and ensure the safety of workers – without the need for specialised staff.
AI provides an array of cost-savings to manufacturing, ensuring that investment is well rewarded. For example, fines for manufacturers have drastically risen in recent years. The total sum collected in fines from 1st February 2016 to 31st January 2017 was £22,781,021.23, with an average fine standing at £157,110.49. Since AI is more effective than human counterparts at reducing compliance costs, the use of FSVS can help companies avert such fines.
AI algorithms can make complex decisions, analyse and present information from multiple sources in ways human operators are not capable. This results in more effective monitoring and utilization of the QC team and reduction in downtime, more effective allocation of operators during the product change, and being better able to identify the right operator for the right job (using the colour of their jacket) – resulting in productivity gains every day. Using this technology, manufacturing companies can target up to an 8% decrease indirect labour cost, a 25% decrease in labour management costs and up to a 16% increase in efficiency due to effective monitoring and insights provided by ThingTrax’s system.
FSVS technology will help manufacturers to monitor the unmonitored, leading to manufacturing success. Currently, a majority of firms rely on expensive and inefficient field monitoring (humans) or hire expensive consultancy firms to develop bespoke solutions, with high up-front costs and little guarantee of future cost savings. AI works more effectively than humans to improve safety, reduce compliance costs and ensure greater efficiency. AI algorithms can make complex decisions, analyse and present information from multiple sources in ways human operators are not capable. Deep learning-based machine vision software helps to minimize human intervention, providing real-time solutions by distinguishing the acceptable variations in products and the defects in manufacturing industries.