Predictive maintenance: science fiction? No, reality

Interconnessione tardiva
       Written by Michele Zubani

Thanks to the increasingly flourishing development of technologies in the field of industrial digitization, companies have started to speed up work in the production department.

This facilitation, however, has also led to an increase in the amount of maintenance that needs to be done to machinery within the various production lines.

Whare are the different types of maintenance?

A type of preventive maintenance that each of us habitually performs is the coupon of the car: we know that we must make the coupon or because we have reached a certain threshold of Km or because it has passed a year since the last coupon.

  1. Reactive or breakdown maintenance: when a component of a machine breaks down, it is only fixed after the damage has been done. The apparent advantage of carrying out this type of repair is the decrease in time/cost of the work done. I say apparent advantage because in reality this type of maintenance can cause long machine downtimes, with the consequence of delaying the production of the entire line and waiting for days for the spare part.

  2. Preventive maintenance: This is based on planned repair activities at elapsed time or working time. This type of maintenance is based on the assumption that all machinery components can degrade over time and that, in order to avoid stoppages/failures during production, repairs to the machinery are scheduled on a regular basis.

  3. Predictive maintenance: is based on objective data collected and processed directly from the machinery of a plant. This allows SMEs to maximize production efficiency by predicting possible production defects and ensuring the maximum interval between repairs, thus minimizing the number of interventions to be made on the machinery.

 

In this article, we'll focus primarily on predictive maintenance and the benefits it brings to SMBs.

Before we start, however, we must make a small premise: history has taught us that the tried and tested maintenance plans (i.e. TPM - Total Productive Maintenance and RCM - Reliability Centered Maintenance) have not brought the hoped-for benefits to companies due to the great lack of attitude in the workplace to properly use data analysis tools to improve production efficiency.

The imminent advent of digitization of the production process has made, therefore, essential the change of approach with which maintenance was carried out in companies until you get to the great demand from Italian SMEs to implement predictive maintenance within their production.

What are the advantages of predictive maintenance?

I vantaggi della manutenzione predittiva

What are the costs involved in predictive maintenance?

  • Measuring devices: The MEASUREMENT CHAIN consists of several elements:
  •  Sensors have different types of operating principles, depending on the physical phenomenon they are going to measure
  •  Electrical disturbances in the working environment are a cause of compromise in the quality of measurements.
  •  Systems for converting measurements into digital format must be appropriate to the phenomenon we want to investigate.

Fortunately, today there are several standards that ensure different levels of accuracy and performance, so it's fairly easy to find the right combination of tools to link together.

  1. Computing systems: today it is possible to process a huge amount of data in a short time and therefore it is possible to use analysis techniques that were reserved for niche markets until a few years ago.
  2. Agile Data Management: BIG DATA defines the problems associated with managing the large amounts of data that are available from machines today.

The main issues are related to three key factors:

    1. Save the data (write the data and be able to access it in reasonable time)
      b. Process the data in reasonable time
      c. Display the data in a way that is understandable to humans.
  1. Technologies and competences

You need to rely on experts who know the ins and outs of each type of technology to avoid making the following mistakes:

  • Using the wrong data to support your analysis
  • Generate the wrong data (measurement chain designed inappropriately for the context)
  • Using analysis tools in a way that is inconsistent with one's purposes.

What needs to be improved in companies in order to begin an effective predictive maintenance project?

The main steps are:

  1. Cultural change: All employees need to know how to use technology correctly through general training. This type of information transfer allows the individual "actor" in the change to be aware of all aspects of the work being done, making them actively involved and willing to improve their work procedures.
  2. Start with the most burdensome and frequent problems: These types of problems are the most dangerous and should be addressed with higher priority. For example, if in 3 months of production, the gasket on my machinery keeps breaking every month, it may mean that there are underlying problems and that they need to be addressed promptly to avoid machinery breakdown.
  3. Step-by-step data collection and analysis design: once you've done the previous steps, you need to start setting up a process data collection and analysis model that is optimal for different SMB needs.

Circolo del data mining

Tecnichnician Analyst
Experience data history
Context and informations quantitative analysis

What are the different predictive maintenance techniques?

  1. Visual inspection: it allows to ascertain the state of the various components of the plant during scheduled maintenance operations. This examination consists in evaluating the hygienic state of some critical points of the plant and their functionality.
  2. Vibrational analysis: through the analysis of spectra, it allows to recognize the most common damages or malfunctions of rotating machines.
    Monitoring of process parameters: it allows to collect data instantaneously thanks to specific devices that are able to detect the smallest anomalies in the signal received and send error signals before there is any significant damage to the machinery.
  3. Tribology: is the science that studies all the problems that can occur in the relative motion between interacting surfaces subjected to load, such as friction, lubrication and wear.
    Thermographic analysis: consists in the detection of thermal radiation emitted by a body through a thermal imaging camera. This activity allows to know the different surface temperatures and to detect possible failures or anomalies in SMEs.

 

To recap, Predictive Maintenance allows companies to:

a. extend the life of machinery and improve its performance over time
b. Reduce downtime, costly raw material waste and replacement costs for machinery or parts of machinery
c. increase earnings
d. strengthen the quality of products and last but not least the corporate image.

 

→ If you are looking for a solution that allows you to implement a proper predictive maintenance strategy, AI.Tech Digital Factory and iDaq Analytics are for you.

→ Request a demo, we will be happy to show you the solutions that best meet your needs.

 

 

Subscrive to the newsletter to stay updated on other articles and events.


SHARE:      

Add comment

Subscribe to our newsletter

Subscribe to get notified about the next update.

Read the privacy policy

Lingua