How to conduct a failure mode and effects analysis (FMEA) of your industrial asset?

Determining your critical failure modes is the first step toward maximizing uptime

Every day we make about 30.000 decisions. This aligns with the average amount of thoughts you have daily. With our lives being busy enough as it is, we make choices which decisions require more consideration. What do we do now? And what do we postpone, or not do at all? In predictive maintenance making the right choices can be more challenging. With a  Failure Mode and Effect Analysis (FMEA) you can make the right decisions and prioritize the right actions. Eventually, you know which failure modes are worth addressing and which aren’t.

smart maintenance_Sensorfy_engineer working on analytics

What is a failure modes and effects analysis?

So, what exactly is FMEA? In short, it’s a risk management tool to systematically identify potential failure modes in an asset, system or process. For each component in an asset or system, or for each step in the process, you determine each possible failure. Typically the more complex the asset is, the greater the number of failure modes. 

The next step is classifying each failure mode by likelihood of occurrence and ability to detect as well as on severity. Performing an FMEA enables you to determine which failure modes require preventive or predictive measures and which require no action at all. The results of an FMEA help you build an effective maintenance strategy with just the right corrective, preventive, and predictive elements.

engineer working from data_tech talent shortage_Sensorfy

Types of FMEAs

There are three types of FMEAs: a design (DFMEA), a process (PFMEA) and a service FMEA. Let’s zoom in on what each FMEA entails.

Design FMEA

This methodology explores the possibility of product or design malfunctions. It is used in product development to improve quality and reduce potential risks of failure. This rings especially true for new, updated, or modified product designs which can unintentionally introduce design failures. Using a DFMEA, you can identify failure modes early in the design process and make changes to the design. Even before going into production, leading to significant cost savings compared to countermeasures in later phases. After all, prevention is better than cure.

Service FMEA

A DFMEA is based on theoretical assumptions. So, there’s always a possibility that those assumptions do not fully represent reality. Moreover, sometimes a DFMEA is not created during the design or is not available to the asset owner. This is where a service FMEA is of value. It helps to address failures that occur during the life cycle despite the efforts made in the design phase. Failures that happen during service can be the consequence of design flaws but equally because of improper installation, -operation, -maintenance, and -repair. Although service FMEAs are typically associated with assets or systems, the same principles can be applied to a process. 

The purpose of a service FMEA is to ensure service tools will perform as required, reach improvement goals, customers experience fewer failures throughout the product’s life cycle, and the asset owner reduces costs by deferring planned maintenance. It’s the first step towards an effective maintenance strategy, as occurrence and severity are based on experience.

Process FMEA

The structured methodology known as Process FMEA (PFMEA) is used to identify probable process failures, including failure that impacts product quality, reduced reliability of the process, customer dissatisfaction, and safety or environmental hazards. PFMEA is more concerned with identifying failures brought on by process modifications than new or updated product designs. For example, this is particularly helpful when a new process or technology is launched. But also when an existing process is introduced into a new operating environment

Training your data models

Why you should implement FMEA

Regardless of the type of FMEA the primary purpose is always to identify and prioritise corrective, preventative or predictive maintenance actions.  However, when you set out to improve the performance of your asset, system, or process, the possibilities can be daunting. There are many technologies and techniques available to monitor and prevent failures. An FMEA is a simple and effective tool to determine which measures are most valuable to implement. These could be, for example, scheduling maintenance activities based on running hours and calendar time or planning a regular inspection. Often, however, to truly optimise the performance of your asset, or process, predictive maintenance is part of the solution. 

An FMEA will likely provide you with quite a list of potential failures. But do you need to address them all? Probably not. And that’s another upside of performing an FMEA; it allows you to choose which risks are worth accepting and which aren’t. Moreover, provided the severity and occurrence are properly assessed, you can calculate the lifecycle costs of your asset or system, for example using a Monte Carlo analysis. Then you can make informed choices about how much maintenance and the type of maintenance you want. 

When the fleet of (similar) assets is particularly large the FMEA is often implemented in the monitoring system so that the asset can generate error codes to mobilize engineers for maintenance efficiently. In this manner, condition monitoring of your assets will feed the database with information on the actual occurrence of failure modes, their effect, and their impact. This helps you to transition from a transactional business into a service-based, data-enabled business that aligns closely with benefits realized by the end customer. 

Download the free ebook!
What to skyrocket your company valuation?
Download the ebook to learn how to disrupt your business models

From transactional to service-based

For original equipment manufacturers (OEMs), shifting to a service-based model offers a huge opportunity. It allows them to optimize their assets and predict their lifetime value, with which they can create a more viable business model. For example, currently, your asset has an expected lifetime of about 10 years. That’s the guaranteed warranty you have on it. So you can sell that product once for 100 euros and again after 10 years when it needs replacement. Or, you redirect it to an outcome-based model, meaning you charge 15 euros per year while using proper analyses and predictive maintenance to ensure the assets last as long as possible. Eventually, it saves the asset owner investments and enhances the OEMs’ profits.

So, apart from saving unnecessary expenses, decreasing unplanned downtime also makes you money

In our Ebook ‘How to 3x your company valuation, we cover different business models for transitioning from a transactional to a service-based one.

All in all, completing the failure mode and effects analysis template can be your gateway to effective predictive maintenance. When implementing it, you further professionalize and improve the prediction of failure modes plus the prioritization of them, doubling down on the benefits mentioned above. Even better, in a world where technology talent is scarce, preventing failure and prioritizing maintenance gives you the advantage of allocating your team as efficiently as possible – making the most out of your resources.

8 Steps to execute failure modes and effects analysis

Now you know what an FMEA is and how you can benefit from applying it, let’s dive into how to perform a study like this one. 

Step 1: Identify and map different components of the assets

The first thing you do is identify each component of your asset and describe its intended function. Say you are a  rail manufacturer. Defining different components of the asset type, such as rubber attachments, shock absorbers, and metal mounting material, is vital.

Step 2: Define the failure modes

Use the information gathered in the first step to determine potential failures for each component. Although some failure modes are component-specific, malfunctions in one interconnected system can impact and result in additional consequences on other subsystems. For this reason, it is crucial to establish the asset hierarchy in step one and identify all failure scenarios in step two. For example, say that one of the essential aspects of a rail asset is that it needs to be electrically insulated. Then you’ll define it as a failure mode if that’s not the case.

Step 3: Identify each failure effect

For every failure mode, you define accompanying failure effects. They describe the consequences of a failure. This is not limited to one effect because a single failure mode can cause several failure effects. Failure impacts can be used to characterize the repercussions of a failure on safety, the environment, and the business (or production).

Step 4: Grade severity

This step is where you add grades to the severity of the failure consequences as felt by the customer. Rank them on a 1 through 10 scale, with 1 having little to no effect and 10 presenting a severe risk. When we were referring to the lack of isolation as a failure mode in step 2, this is the step in which you grade how much of a risk little to no isolation would present.

P-F curve_Audible noise and thermography analysis

Step 5: Rank the occurrence

Failures are ranked on a scale of 1 to 10 in the occurrence ranking, just like they are in the severity ranking. It measures the likelihood of failure. So, for example, where 10 might be that a failure occurs every couple of days, 1 means it only happens every 10 years.

Step 6: Evaluate and assign detection grades

The detection ranking represents the chance that a failure will be found before it affects the customer. For example, grade 1 signifies that a failure is nearly certainly discovered, whereas 10 suggests that the failure cannot be detected. For a rail component, for which the primary function is isolation, measurement is necessary to tell whether or not it’s possible to detect isolation failure modes. This is where predictive maintenance techniques, such as measuring insulation values, oil analysis, and vibration analysis, can help.

Step 7: Determine RPN

The variables in steps 4, 5, and 6 help you to determine the Risk Priority Number (RPN). This number ranks the risks from highest to lowest. Doing so enables teams to prioritize assets in need of additional quality planning. The formula is as follows:

RPN = severity x occurrence x detection

Step 8: Perform a small-scale pilot

Once you have determined the RPN and have insights into failure modes, you want to keep up the momentum. An FMEA is a singular analysis you can use as a starting point for continuous maintenance. This is where predictive maintenance comes into play. Based on the FMEA results, Sensorfy can design a preliminary predictive maintenance system to detect the most critical failure modes early on and improve the availability of your asset. Once this custom IoT solution is implemented, we can collect data to validate key assumptions and determine the solution’s viability during a proof of concept.

From FMEA to predictive maintenance

Any stage of a project’s lifecycle, including concept creation, design, production, testing, installation, maintenance, repair, and disposal, can benefit from using FMEAs. An FMEA should be seen as a systematic technique that offers a framework for identifying important concerns connected to product dependability and quality rather than as a stand-alone instrument. It’s a methodology that helps you define the next steps of your optimization process. Above all, it provides a foundation for implementing predictive maintenance that will eventually help you achieve more with fewer resources, prevent failures, and dramatically minimize your downtime. 

So, what are you waiting for? Start with your FMEA today by downloading our (free) failure modes and effects analysis template! After filling this out, our experts are ready to guide you in the further steps!

Download your (free) FMEA template

This blog was written by industry experts

predictive maintenance experts_Gijs Meuleman_Frank Velthuis

We empower forward-thinking industrials.

Let's work together.