Predictive Maintenance: Singapore’s Bridge and Foundation for Industry 4.0
In November 2017, the Singapore Economic Development Board (EDB) launched the Smart Industry Readiness Index (SIRI) and its accompanying Assessment Matrix. SIRI was created by EDB in partnership with a network of leading technology companies, consultancy firms, and industry and academic experts.
Comprising a comprehensive suite of framework and tools to help manufacturers start, scale and sustain their manufacturing transformation journeys, the three building blocks of process, technology, and organisation which SIRI lays out, have been key in assessing companies’ industry 4.0 readiness.
The advancements of Internet of Things (IoT) have been instrumental in paving the way forward for industry 4.0. While such a prospect is exciting, businesses need to understand the considerations needed to pivot towards industry 4.0. Businesses need to identify what is redundant and needs to be removed, learn how to adopt existing components into the new setup, and understand what is native for the new setup.
Improving the SIRI blocks of process and technology has to be balanced with practicality. Depending on the current stage of a company’s technological journey, embarking on a partial or complete system overhaul would be costly, and the disruption with downtime would be impractical. Although change is necessary for our manufacturing industries to stay relevant and competitive, it does not need to be painful or drastic and should involve a proper and appropriate planning process.
Predictive maintenance becomes the crucial factor that enables companies to fully optimise their current processes, regardless the stage of their business. It also functions as the bridge that smoothens the transition as they build up industry 4.0 capabilities and serves as a litmus test to determine the level of digitalisation that should be adopted by any process.
Prior to industry 4.0, maintenance in the manufacturing industry had moved through a few different phases, before landing on predictive maintenance as its latest iteration.
- Reactive maintenance – The most rudimentary approach where things are fixed only when they fail. For cheaper machines with redundant parts, this was often acceptable. However, this can only address the symptom and not the problem.
- Preventive maintenance – Replacement of parts before they fail. This timed approach often leads to higher costs as parts are binned while having considerable mileage left, requiring more planned downtime for machines that are working perfectly well.
- Proactive maintenance – Normally reserved to reduce the cost in money and time for complex and expensive machinery, companies address the peripheral symptoms that can lead to problems, and leads to a data-driven and analytical approach.
Predictive maintenance as a bridge
Though cheaper at the onset, the above-mentioned maintenance models resulted in productivity losses of between 5% to 20%, which will adversely impact any operation in the long run. By balancing costs with results, predictive maintenance can be progressively implemented to serve as a bridge. For companies that are embarking on this implementation, examining their processes and establishing a baseline for the condition of parts and machinery involved through any historical or recorded data from their current maintenance workflow would be a good start. As the effectiveness of predictive maintenance is predicated on the quality of data, the baseline helps to identify which parts of the process the team knows most about, and which parts have the greatest area for improvement.
Adopting a systematic approach and knowing the available key data points will then allow companies to map their way to other parts of their processes that have little to no data. Gradually, giving companies a better picture to navigate their workflow to the unknown will help companies develop a complete picture of their workflow and to ultimately decide what industry 4.0 means for them.
Predictive maintenance as a litmus test
For parts of the process that lack sufficient data to determine points of failure, businesses are often faced with a feast or famine scenario, where data is not collected until a failure occurs. Other times, the data collected is good for diagnostic purposes, but not comprehensive enough to aid in predictive maintenance. In such cases, creating a visual model or simulation would have to be done first before predictive maintenance could come into play.
Predictive maintenance gives companies the flexibility to be progressive and consistent with minimal disruption, and it has to be tailored to the industry and the workflow. As detailed in a case study by McKinsey & Company, predictive maintenance has to be weighed carefully as an option for these following reasons.
- Lack of data points for predictive maintenance to be truly effective
- The timelines and resources involved in the maintenance process
- Lack of significant impact, for example, in the case of multiple redundancies in place
- Lack of cost savings for predictive maintenance to be financially viable
Predictive maintenance as the future
There is still plenty of room for predictive maintenance to become an industry standard across the board. In a report by McKinsey & Company, only 30% to 40% of production lines in ASEAN are automated. By 2025, predictive maintenance will account for 10% to 40% reduction in spending, 3% to 5% improvement in equipment lifetime, and 50% reduction in equipment downtime. Predictive maintenance in ASEAN is projected to have an economic impact in the manufacturing industries of between US$38 billion to US$91 billion by 2025.
Adopting predictive maintenance will not be an overnight process, and the Singapore Government is dedicating resources and grants to elevate the manufacturing landscapes. Companies will have to go back to the SIRI board and make sure that their people and organisations are heading in the same direction on their industry 4.0 journeys.
The author is Kai Chua, CEO of Linkstuffs Pte Ltd. Linkstuffs is a tech solution provider for the manufacturing industry.