Predictive x Prescriptive - what is the future of maintenance?

Since the First Industrial Revolution, the development of operation and maintenance techniques in the industry has always been a constant process of evolution. The deployment of new technologies has accompanied this progress and joined these efforts with the aim of achieving more efficient industrial processes and increasingly longer useful life of the assets involved.

The maintenance routine in the industry started in a corrective way, meaning that only when equipment or structure has a problem that a correction or substitution occurs. This type of action has the lowest efficiency since the problem only starts to be solved when it has already happened, increasing downtime and involving greater operational risks. According to The Welding Institute, in Cambridge, corrective maintenance

?"is inefficient as a general working practice and can lead to greater expense if one failure leads to other parts failing and an eventual catastrophic failure of the whole asset. Such serious problems can lead to long periods of downtime, expensive repairs or even the need to replace an entire piece of equipment. All of these factors can quickly cause costs to increase dramatically."

With the evolution of industrial processes, a new type of maintenance has emerged: preventive. In this case, equipment inspection occurs periodically, even if it seems to be working perfectly. Despite the improvements in preventive maintenance, what happens in many cases is less optimized work, because there is not always the need to inspect an equipment at that particular moment and it involves very analogical techniques, generating diffuse data and spending more time than necessary.

There came a point when it was necessary to balance equipment and maintenance costs, in search of greater optimization of operation and profitability. This balance began to emerge from the investment in Industry 4.0, with the development of new technologies and the increased use of data within the industry.

This increased focus on cutting-edge technologies even led to the results of a survey conducted by McKinsey with the Global Economic Forum, indicating a potential of $3.7 trillion by 2025 in value creation from manufacturers and suppliers that implement Industry 4.0.

And one of the justifications for this growth is the efficiency of processes that the industry digitalization brings. In this case, we can cite two very important examples within this scenario: predictive and prescriptive maintenance. Both based on data collection and contextualization, they represent the latest levels of evolution in the industrial maintenance process and are essential to be understood by those who want to achieve the highest levels of efficiency in the industry.

What is Predictive Maintenance

Predictive maintenance focuses on presenting patterns and predictions about the efficiency of a particular machine or structure, based on data generated by a sensor or equipment. The objective of this type of inspection is to act on a problem before it happens, but in a more optimized way than the preventive one, since it uses data to understand the best moment to act on a certain asset.

In this case, a number of sensors are used that can capture everything from the operating temperature of a piece of equipment to the internal corrosion degree of a pipe. These sensors then generate time-based data on the operation and allow professionals in the field to analyze the data and make decisions based on it.

Predictive maintenance is often linked to sensor monitoring because it is a great way to remotely control the integrity of the equipment. However, it is data that really matters when considering this maintenance strategy. It is possible to correlate maintenance work orders and asset failures through historical data with sensors, for example, and use Artificial Intelligence outputs to indicate when equipment is likely to fail. Therefore, it is possible for operators to analyze the data and take action before accidents happen.

What is Prescriptive Maintenance

The last type of maintenance, the prescriptive one, evolves a lot in terms of the technologies involved. Developed in the midst of the 4th Industrial Revolution, this type of maintenance adds tools such as Digital Twin, Artificial Intelligence, Machine Learning, and the Internet of Things to contextualize the data generated by sensors and suggest actions from them.

In other words, unlike predictive maintenance, which requires a person responsible for cross-referencing, understanding the collected data, and building an action plan based on it, in this case, the technology itself indicates the best next steps.

For example, in prescriptive maintenance, it is possible to obtain when is the best day to perform the inspection of a piece of equipment, what is the best temperature for the operation of a machine, and even when it is necessary to repaint a structure. And all these insights are delivered by the technologies involved and the learning from them over time.

In this way, ?is only by combining experience with a technology strategy that prescriptive or predictive maintenance can enable the industrial environment to become much more data-driven, increasing the efficiency of the operation and reducing costs.