Many technology trends go through hype cycles. At a certain stage of the cycle everyone is so enamored with the perceived benefits that investments are made without a sound rationale.
Riding on manufacturers’ need to curb rising maintenance costs, ensure seamless operations, and reduce downtime, Predictive Maintenance is being hailed as the gateway to Industry 4.0. In fact, according to a report by Allied Market Research, the global predictive maintenance industry is estimated to reach $23.01 billion by 2026, growing at a CAGR of 30.2%[i]. But will these investments yield results? Or is there a better way to accelerate towards smarter manufacturing? Will blindly following an industry trend cost you? Will you miss out if you do not take steps towards predictive maintenance?
Predictive maintenance is the future – but it’s not the Now!
As things stand today, predictive maintenance has been built into existing systems through trial and error. It’s not predictive as much as based on baselining historical data with a heavy influence of people and paper. Though manufacturing processes are being digitalized and automated, a majority of machines on factory floors are not connected devices with IP addresses. These legacy assets may have been around for decades and it is difficult to get away from paper-based processes and tribal knowledge.
Currently, an engineer or maintenance resource on the shop floor knows more about the legacy asset than a digital system. They know from their long years of experience of running these factories that a bearing in a machine or an axle or spindle will likely fail after 10,000 revolutions or a set number of hours of use. And because these assets, their components, their lifetime and likelihood of failure, the need for maintenance is very well understood, adding a digital dimension doesn’t add a whole lot of value. So, for most of the processes that are currently in action in the industry, predictive maintenance isn’t a disrupter.
This is not to say that predictive maintenance isn’t important. It will, in fact, play a critical role in the next 20-30 years as new processes around new assets are rolled out. Take 3D printing for example, a 3D printer is not well understood today. There hasn’t been a lot of 3D printing and we don’t have the tribal knowledge or experience from creating millions and millions of 3D printed units. So, we don’t know how well these 3D printers are going to hold up over 20 years. But we do have the capability to digitally collect telemetry data from 3D printers at a granular level and feed it into predictive models. This data collected without manual intervention, over a course of time, parsed through intelligent algorithms will help us predict and prevent failures. So, we can expect alerts like “this particular printing arm that is moving the unit creating the 3D model is getting too hot and is likely to fail soon” and do something about it.
So, the crux of the matter is that predictive maintenance is going to take a while to deliver value and will be better applied on the newer machines.
But what can manufacturers do in the interim?
Reap benefits of smarter manufacturing with predictive quality management
Currently, quality assessments are done by people. This judgement of quality varies from person to person creating a lot of subjectivity in the system. People can interpret test data in several ways impacting the overall output that could impact the business. For instance, a small variation in dimensions could make car windows rattle or a slight change in the antenna could distort audio quality. Digitization of quality assessments takes out the subjectivity from the analysis. Based on a series of tests (e.g. x-ray, sonic resonance etc.) on a sample, the system can determine if the end product meets the required quality standards.
With millions of test data instances, Artificial Intelligence and Machine Learning can quickly be applied to the system allowing for intelligent, real-time insights and action. Connected, intelligent testing systems enable manufacturers to test products in real time as they move along the assembly line. Digital systems also make quality testing faster and more comprehensive. For example, it can take up to three weeks to complete the end-of-line testing for Lithium-Ion battery cells and if it doesn’t meet the quality standard it must be scrapped – costing valuable time and resources. With predictive quality analytics, a research center in Germany identified the affected battery cells early on, improving the manufacturing yield by 16%[ii].
Benefits of a predictive quality system
- Reduce the cost of warranty claims and recalls
- Proactively identify product quality and safety issues
- Identify root cause of issues more effectively
- Reduce manual effort
- Stay compliant with regulatory requirements
- Improve brand reputation and customer satisfaction
Making smart investment decisions
Competitive advantage is a matter of making smarter investment decisions. While investments in predictive maintenance cannot be ignored, companies need to pace them for future gain. Predictive quality promises more immediate ROI and is easier to implement for larger volumes. But answering these questions and taking decisive action can be difficult. That’s why at Wipro, we help our clients rationalize their predictive analytics journey. With our experience and domain knowledge, we help you find the sweet spot to allocate scarce funding for maximum gains.
To understand why you need to prioritize investments in predictive quality, connect with me at karl.veidis@wipro.com
[i] https://www.prnewswire.com/news-releases/predictive-maintenance-market-to-reach-23-01-bn-globally-by-2026-at-30-2-cagr-allied-market-research-300918549.html
[ii] https://www.elisasmartfactory.com/how-predictive-quality-analytics-increased-battery-cell-manufacturing-yield-by-16-percent/
