All you have to know about

Industrial Predictive Maintenance

THE DEFINITIVE GUIDE

What you'll find in this guide

    • The evolution of Industrial maintenance
    • Industry 4.0 and Predictive maintenance
    • What is predictive maintenance?
    • What to look for in a predictive maintenance solution
    • From predictive to prescriptive, a step

"The world is changing. Our consumer habits are not the same as they were ten years ago, and probably none of us can define anymore whether these habits were a product of digitalization or the opposite. "

Patricio Saez

CEO

predictive-sigma

 The truth is that our way of consuming has impacted our society at all levels, and the industry has been the main affected by the challenge of responding to demand in the most agile, sustainable and profitable way possible.

At the entrance to the age of The Industry 4.0, companies are faced with a range of possibilities to face this new challenge. Surely, one of the areas with the most potential is that which has to do with the optimization of the performance of the machinery. That is to say, achieve full performance and minimize the possibility of failures. Thanks to the Advances in IoT, artificial intelligence and cloud computing, current maintenance managers now have more tools to design action plans that advance to machine failures and ensure optimal performance. This is Predictive Maintenance, Industry Development Branch 4.0 which promises to be one of the areas of greatest impact at the industrial level.

 

The guide below is intended to describe what predictive maintenance is and analyze it from the perspective of its applications, applied technology and its Differences from traditional maintenance. From the experience of a technology-based company, we also reflect on the challenges faced by predictive maintenance today and how it should evolve towards maintenance of the prescriptive type.

We hope that the information will be of value and will allow you to understand one of the most relevant areas of application in the 4.0 industry.

01.

Industry 4.0 and Predictive maintenance

02.

What is predictive maintenance?

03.

What to look for in a predictive maintenance system

04.

From predictive to prescriptive, a step

PREDICTIVE MAINTENANCE GUIDE

The evolution of industrial maintenance and the impact of Industry 4.0

QRobablemente one of the oldest tasks associated with industrial processes is that which has to do with the maintenance of the machinery used. Profitable companies have always invested in resources and mechanisms to ensure that their industrial assets produce in a continuous and predictable way, extending their useful life as much as possible.


Traditionally, maintenance has had as the only tool the reactive response to the failures of the machinery, that is to say, to repair that which is spoiled as fast as possible. This type of maintenance has as key tasks the identification of the critical parts of the machinery and securing a stock of them to be able to minimize the spare time. It is also crucial to have specialized technical personnel who are able to solve the breakdown as quickly as possible. In conclusion, the slogan is to plan to reduce reaction times.

YOU MAY BE INTERESTED IN: What types of predictive maintenance exist?


The evolution of traditional maintenance comes with the so-called preventative maintenance. The knowledge of the machinery over the years and the historical behavioral data that both companies and manufacturers have been able to collect has allowed us to make estimates of their useful life quite accurately. This has allowed to plan maintenance actions whose objective is to replace those parts and mechanisms that the statistic says will fail in the short term, before they fail. This has been an important qualitative leap in productivity, but with the limitation that it is only profitable in cases where the cost of replacing a part is more economical than the repair of the entire machine. On the other hand, companies sometimes invest in this type of maintenance without really exhausting the life of the machinery, either because of the particularity of their production process or the volume of output.

Want to know more? Find out what types of machinery you can apply the maintenance of predicitive-Sigma Predictive


The increase in consumption, the optimization strategies of the time to shelf and the emergence of disruptive businesses that increase competition has made companies need to optimize their maintenance strategies. In this way, the main challenges or difficulties of the current industrial maintenance are:

Cost reduction

To ensure the proper functioning of the machinery at the lowest possible cost.

Eliminate stops

of production or service. In case of having them, that they have a duration of the shortest possible time, so as not to lose productivity.

Extending the service life

of the asset: to have machines in optimal operation for as long as possible.

Find out how we do it in predictive-sigma

we'll tell you. The future is predictive

PREDICTIVE MAINTENANCE GUIDE

Industry 4.0 and Predictive maintenance

According to a study by the consultancy Happiest Minds, the main responsible of the industrial revolution is the immense reduction that technology has suffered in the last decade. The reduction of these costs allows that, today, any device can connect to the Internet and send data that provide real-time information to its users, what we know as Internet of Things. This is the real trigger of the Industrial Revolution 4.0, since information allows to automate and optimize processes in ways that had previously been very costly or slow to implement.

Industrial maintenance experts have been able to use these new technologies and develop practices called Predictive Maintenance, which is based on the use of actual machinery performance data to take corrective action before failures or breakdowns occur.

 

Want to know more? Discover how the industrial predictive maintenance platform works predictive-sigma 

HOW THE TRADITIONAL PREDICTIVE MAINTENANCE DIFFERS

 

Predictive maintenance responds to the challenges of industry transformation because, unlike traditional methods,

  • It uses sensors that allow real-time data capture of the machine or process in question.
  • Performs data processing by artificial intelligence systems that allow the detection of behavioral patterns.
  • Generates predictions of value in relation to the performance of the machinery.
You may be interested: what are the differences between predictive and traditional maintenance?

4.0 Technologies for Predictive maintenance

Internet Of Things (IoT)

Network of physical objects-from vehicles, appliances and "wearables" to industrial machines-that have integrated electronic components, software, sensors and network connectivity that allows to collect and exchange data over the Internet. One study estimates that the cost of the sensors has decreased by 60% and the cost of processing data has decreased 60 times. The IoT is intrinsically related to the Big data, which allows to transform data into information.

Elements to look for in the choice of predictive maintenance

Big data

It is the management and analysis of data whose quantities would be impossible to process with the tools and conventional human resources. The Big data manages to provide valuable information about the behavior of different processes and services, which can be used to prevent problems, among other purposes.

Artificial Intelligence

It is the general term to define a set of computer systems that can feel, think about learning and take action in response to what they are feeling and their goals. This is achieved through the use of tools such as bio-inspired algorithms, probabilistic reasoning and artificial neural networks. According to a study by the consultancy PwC, the value of artificial intelligence is found in the ability of this technology to increase productivity, workforce and custom demand. (6)

Machine Learning

Branch of artificial intelligence which refers to the ability of a machine equipped with artificial intelligence to administer and, above all, to automatically learn (2). Based on the identification and extraction of complex patterns from millions of data, an algorithm is capable of extracting information and obtaining high-value predictions about future behaviors for better decision-making. This implies that, automatically, systems are improved independently over time, without human intervention.

Find out how we do it in predictive-sigma

we'll tell you. The future is predictive

PREDICTIVE MAINTENANCE GUIDE

What is predictive maintenance?

Predictive maintenance is a modality of industrial maintenance that allows to know the operation of machines and equipment by means of non-destructive measurements in order to advance to possible failures and optimize their performance. Predictive maintenance makes use of the new technologies of the industry 4.0 to provide relevant information about the performance of the machinery and proposes maintenance actions that guarantee the operation and its useful life. A predictive maintenance system is ultimately intended to prescribe specific maintenance actions.

 

YOU MAY BE INTERESTED IN: What is predictive maintenance?

Key Predictive Maintenance Variables

Depending on the sensor used, predictive maintenance can exercise its function by monitoring different parameters. The most frequent are:

  • Temperature Analysis: Measures the infrared radiation emitted by the elements of the electrical system and allows to elaborate a heat map of a certain installation. Heat points can detect different events such as mechanical deterioration, overloading at specific points, etc.
  • Vibration Analysis: It measures, by mechanical parameters, the vibration of the machinery. The hypothesis behind its use is that an alteration in the standard vibration of a team is a consequence of events such as loss of energy, wear and damage due to material fatigue, among others. This technique is useful for rotary machines, in which there is a rotating element that produces vibration. For optimum operation it is necessary to have a sensor in each of the axes of motion.
  • Analysis of electrical parameters: measures the changes in the electrical behaviour of a machine or installation by means of high-precision sensors. Due to the high sensitivity of the electrical parameters, this method is capable of detecting deviations of the standard behavior in very embryonic phases, facilitating the maintenance actions. Its use is suitable for any type of machinery or installation connected to the electric current, whether large or small.

YOU MAY BE INTERESTED IN: What types of predictive maintenance exist?

  • Visual inspections: It ranges from simple direct visual inspection of the machine to the use of complex observation systems such as microscopes, endoscopes and strobe lamps. Its objective is to alert the appearance of cracks, leakages and signs of corrosion, among others. This method is useful in large installations with difficult access.


Areas of industrial application

Predictive maintenance has application areas along the industrial value chain. From productive systems to the tertiary sector, any organization that has an expense in maintenance of machinery is susceptible to the advantages of this new methodology.

Productive Sector-Industry

Unforeseen stops are the main problem in the Companies engaged in production. The fact that a part of a production process stops, leads to the delay of the expected production, loss of raw materials and semi-processed, stock breaks and unforeseen maintenance costs. That is why in production, the slogan of industrial maintenance has always been "to prevent production from stopping".

Predictive maintenance works continuously analyzing the working conditions of industrial assets, identifying deviations from optimal functioning in very embryonic states, informing the responsible team and allowing a performance on the An anomaly in a planned and organized way that ensures the availability of the asset in optimal conditions. "Listening to teams reduces unexpected stops."

Service Sector

The services sector, similar to the productive industry, needs to have its machinery in operation to guarantee the service to its customers. In this way, if in the productive industry the slogan is "To avoid stops", in services the slogan is "to guarantee the availability of the service infrastructure"

Predictive maintenance allows you to automate the analysis of all the assets that are part of a service infrastructure. From the control of pumping stations to ventilation systems, everything can be controlled, reducing the chances of a failure in the service.

Want to know more? Find out how to ensure the availability of your infrastructure with predictive-sigma

Advantages of predictive maintenance

01.

Increased service life

By being able to detect the failures in the machines before they happen, you can take the necessary actions and program them for the time that less affects the productive cycles. This way, it ensures that the life and availability of the machinery are extended. In other words, the stops are reduced and production increases.

02.

Cost reduction

It avoids having to completely replace a machine, since it is known exactly which part of the machine should be taken care of so that it does not fail as expected. That is to say, any anomaly in its earliest state can be prevented to drastically reduce the costs in spare parts or replacement of equipment, carrying out only the precise maintenance.

03.

Optimum production output

Identifies the deviations of the process that can cause an alteration in the performance of the machinery, as small as they are, assuring the optimum yield of the production.

PREDICTIVE MAINTENANCE GUIDE

What to look for in a predictive maintenance system

Not all predictive maintenance systems are the same. The accuracy of the technology and the ability to predict possible failures will depend on a number of factors. Although the adoption of a system as such can make a significant leap forward for the availability of machinery, we then give you the aspects that you have to keep in mind in order to opt for one solution or another.

01.

Data: Quantity and Quality

As mentioned in the article by Predictive Analytics in practice, by hbr.org, the source and quality of the data to be monitored is the basic pillar for the good deployment of a predictive maintenance strategy. The more variables the system is able to collect, the more information it can value for anomaly identification. The concept of big data applies in this aspect. A reliable predictive maintenance platform is one that can integrate data from different sources through data mining, consolidate it into a single point, cross it, and obtain relevant information from its analysis.

02.

Frequency is important

The more intense the monitoring, the faster the system can detect a deviation from the standard behavior. Rely on systems with continuous monitoring connected to a cloud platform, where real-time information is consolidated and analyzed. In this way, the response times of your predictions will be faster and more accurate.

03.

The intelligence behind the predictions

The amount of data that a predictive maintenance platform can manage is immense, and therefore it has to have tools of Artificial intelligence That is able to perform analysis automatically. It is equally important that behind the algorithms of these platforms there is a good team of data scientists, with experience and knowledge in the nature of the industrial sector. It is also important that the platform has learning mechanisms and feedback systems that will allow it to improve its predictions with external inputs.

PREDICTIVE MAINTENANCE GUIDE

From predictive to prescriptive, a step

Prescriptive maintenance is the natural evolution of a technological model that allows predictive maintenance

It is common agreement that predictive maintenance is an important qualitative leap in the face of traditional industrial maintenance strategies, since it makes available to experts, quantifiable data on the behaviour of machinery that allows Avoid many of the unforeseen stops in an installation. Predictive maintenance allows, among other things, to increase the availability of machinery, reduce unforeseen stops and reduce costs associated with maintenance by making use of industry technologies 4.0.

You may be interested in: What are the main technologies in the 4.0 industry?


Data scientists and machine learning

However, predictive maintenance depends largely on the human team, the so-called data scientists,who interpret the data and develop the algorithms capable of identifying such patterns of behavior. To date, technology has focused on providing data scientists with tools to understand the universe of Big Data, so that they can become interpreters of that data and develop an artificial intelligence. But the human dependence on the detection of patterns is not the ideal situation, because the computing capacity of the machines allows the detection of patterns and deviations much more quickly and accurately. For this reason, the natural technological evolution is that the machines develop their own intelligence for the detection of patterns with the minimum human intervention, what is known as machine learning.


The difference lies in the actionable information

The introduction of machine learning strategies in predictive maintenance platforms, coupled with data crossings from different industrial sources (MES, human resources, environmental conditions, etc.) would allow a thorough knowledge of the Operation of the machinery in relation to its use and real behavior, making the system able not only to predict a breakdown, but to give instructions, ie, prescribe specific actions to solve the possible failure. For example, there is a dramatic difference between the prediction of a possible cortocicuito in a motor to a message that recommends to the maintenance team to check the connection between the motor and the power supply for possible short circuit risk.

Prescriptive maintenance is the natural evolution of a technological model that allows predictive maintenance. Companies that adopt predictive maintenance models today will have a technological base that will allow them to easily migrate to prescriptive maintenance strategies and benefit from all the advantages that this implies.