And tomorrow, what maintenance?

Prof. Pierre Dehombreux, University of Mons, Mechanical Engineering Department - 10 years of 4.0

Industrial 4.0 environments bring together, within a platform for logistical coordination and control of production resources, autonomous electronic and computer systems - embedded systems - and sensors communicating their data via Internet services - Industrial Internet of Things or IIoT - which ensure a secure interconnection of databases, digital services and multimodal man-machine interfaces. Production system control decisions benefit from fault detection, diagnostic and prognostic information generated by artificial intelligence applied to the processing of massive data (Big Data).

The Walloon Maintenance Forum of June 2021 allowed to appreciate the reality that this 4.0 revolution is progressively taking in several maintenance departments. Priority is given to the revision of the management of data from production and maintenance, in its various phases: collection, storage, sharing, analysis.

Of course, there is still a lot of room for improvement for a generalized and successful 4.0 deployment.
For example, predictive maintenance (for the purists) or predictive maintenance (for the others), which aims to estimate the time before a failure occurs, is only applied in a limited number of industrial facilities. Conditional preventive maintenance enables maintenance action to be triggered before failure occurs, based on the periodic or continuous measurement of one or more degradation indicators (vibration level, temperature, etc.). Predictive maintenance goes further than the latter because it aims to estimate, for a component or a system, the time span before its failure occurs. Conditional preventive maintenance is based on a comparison of indicators with warning threshold values; predictive maintenance extrapolates their evolution to provide a prognosis, i.e. an estimate of the residual useful life (RUL). Convincing results have been obtained on simple machine elements, such as roller bearings or cutting tools used in machining. 

Figure 1: Comparison of corrective, preventive and predictive maintenance.

Predictive maintenance will certainly develop through the integration of sensors and monitoring systems into machine elements. A good example - which is a widespread automotive application of 4.0 - is the integration of the Tire Pressure Monitoring System (TPMS), which allows the driver to be informed of a leak that could lead to an accident due to the loss of stability it causes. The 4.0 implementation of this TPMS will be complete when this anomaly detection is accompanied by a safe mileage (which corresponds to the prognosis linked to predictive maintenance), or better still, by a recommendation to carry out a check at a garage located not far from your journey and indicated on the vehicle's GPS. This last stage is the example of tomorrow's maintenance, which is known as prescriptive maintenance: maintenance that is accompanied by recommendations on the decisions to be taken, not only according to the health of the system, but also according to the constraints and possibilities of the moment.

A notable advance in Industry 4.0 is that it leads to the decompartmentalisation of production and maintenance, since these activities now feed into databases that are used for a common objective: ensuring the performance of assets.

4 fundamental axes of 4.0

Maintenance is fully recognised as a lever for the performance of a production system for goods or services. It cannot be reduced to a set of technical interventions. It must be organised around 4 axes:

• a strategic axis leading to clearly targeted objectives: classically: reliability, availability, cost, quality, lead time, safety, but now including those of sustainable development;
• a methodological axis where a global maintenance policy is thought out according to the objectives: Reliability-centered maintenance, Total Productive Maintenance, 6 sigma, operational excellence, Value Driven Maintenance: whatever the framework, continuous improvement will always find its place in a participative approach;
• a technological axis: IIoT, Big Data, artificial intelligence: the current tools are prodigious, provided they are used for reporting, detection, diagnosis or prognosis needs clearly established in the maintenance policy in place;
• a human aspect: more than ever, respect for the men and women who carry out maintenance, whatever the nature of the action, is essential to the success of the industrial performance project, and all the more so as they are expected to have an increasingly global understanding of the issues at stake and increasingly specialised skills. Corporate culture and leadership are essential. The announced hybridisation of post-covid work will lead companies to rethink the organisation of work in depth.

It is often said that it is difficult to attract people to maintenance because of its long-standing image of hands-on work. Maintenance respects human work, contributes to the economic health of companies and, above all, makes it possible to achieve the expected reductions in energy and material consumption by chasing losses. 

The world of maintenance must communicate more on the axis of sustainable development and on its societal usefulness. Shouldn't we adopt KPIs that better reflect these objectives? ... This is the basis for further debate within BEMAS and beyond!

Do you have a question? Contact me at pierre.dehombreux@umons.ac.be