Government Europa Quarterly explores the impact of machine malfunction on the safety of workers and the future of predictive maintenance.
In line with the vision for Industry 4.0, the automation of machines is becoming especially prevalent. Artificial intelligence (AI), the industry Internet of Things (IoT) and machine learning are just some examples of the revolutionary outcomes of this transition towards smarter industry practices. However, with these revolutionary technology comes an increased risk of machine malfunction, and a growing need to predict and prevent it.
Traditionally, field service – the field of fixing industrial machines – is used to encompass the many specialised engineers which are responsible for ensuring that industrial machines are active and have little downtime. However, with these advances, machines are increasingly capable of thinking for themselves, and are thereby able to give operators an insight into potential downtime through predictive maintenance.
Machine malfunction: a common issue
Machine malfunction is universal, whether in the technology of the past, or of the future. As the world takes a more focused view of future technologies, the associated risks are exemplified in a growing number of reports of machines malfunctioning in ways that are not only inconvenient, but dangerous.
Traditionally, if a machine were to malfunction, a technician would be dispatched, the problem identified, the relevant replacement ordered and installed a few days later. Now, AI technologies are offering a variety of benefits in diagnosing and combating the ever-present risk of machine malfunction, including:
- Monitoring equipment, tools, materials and people in real time;
- Close surveillance of heavy machinery to prevent malfunction; and
- Improving worker safety and job site productivity.
Moreover, specialised software – based on the foundations of facial recognition – could be developed into security systems which could grant authorisation and enforce policy and regulation to employees in industry before they can gain access to equipment which poses a threat to safety. Through analysis of a database of certification and facial recognition, AI could eliminate the threat which unauthorised access to machinery poses.
A case study – exemplifying the need for regulation
The complications presented by machine malfunction are two-fold. The ability for machines to “think” for themselves has led to several cases which recirculate the debate around the relationship between robots and workers. In 2015, a maintenance technician of Vetro Ionia in Michigan, US, was killed after a robot worker malfunctioned and gained access to an unauthorised area. Although the manufacturer had safety doors in place to prevent robots from entering unauthorised zones, the operating system for the zone enabled the robot to bypass these precautions.
Owing to such cases, the demand for stronger regulations and legislation on robot workers is growing ever more urgently. Without the appropriate safety protocol and regulations, future incidents such as these cannot be mitigated.
In 2017, MEP Mady Delvaux presented a report to the European Parliament, urging for European authorities to draft new regulations on the use and creation of robots. These new rules would include providing “electronic personhood” – and with this rights and liability – to sufficiently advanced technological beings. As reported in The Guardian, Delvaux said: “A growing number of areas of our daily lives are increasingly affected by robotics.
“In order to address this reality and to ensure that robots are and will remain in the service of humans, we urgently need to create a robust European legal framework. What we need now is to create a legal framework for the robots that are currently on the market or will become available over the next 10 to 15 years.”
Predictive maintenance and deep learning
As machines are now being installed with their own monitoring technology, cloud connection and information on diagnostics, the machines of the future will be able to diagnose their own problems. As a result, machines will have the ability to alert their manufacturer as and when they are performing below their optimal levels, and through analysing patterns in data, outages can be predicted in order to avoid downtime.
One of the key developments which has stemmed from the pool of technological advancements is deep learning. Typically, the term is used to encompass software algorithms, or neural networks. Through filtering relevant data through several layers of artificial neurons, the machines can develop their capabilities to perform tasks. With a broad range of applications, deep learning is a technique set to revolutionise progress towards the vision of Industry 4.0.
Deep learning has already been used to attempt to understand the noise patterns of troubled machines, and with this, predict necessary maintenance in advance. The technique could be applied to cars and industrial machinery, and eliminates the possibility of missing problems during routine maintenance checks. As a result, through analysis of noise patterns, the machine will be able to identify that there is a problem, prior to malfunction.
Anticipating the benefits
In April 2017, US-based consultancy organisation McKinsey & Company released their report, ‘Smartening up with Artificial Intelligence (AI) – What’s in it for Germany and its Industrial Sector?’. The report anticipates that through the facilitation of predictive maintenance by AI technologies that asset productivity in Germany could increase by up to 20%, whilst maintenance costs may be reduced by around 10%. Further, it could reduce downtime by 20% and see a 25% reduction in costs relating to inspections.
Moreover, the company highlights that the greatest impact for predictive maintenance is likely to be evident in original equipment manufacturers in the automotive sector, as well as commercial vehicles, with industrial equipment also set to benefit significantly.
