Failure analysis is an essential part of the maintenance process. That is, understanding why the failure occurred is crucial for addressing the cause and correcting the problem assertively, making the industry more efficient and productive.
According to NBR 5462 (Brazilian technical standard), failure is the termination of an item’s ability to perform the required function. This same standard has established the general principles of maintenance management since 1994, making it one of the most important references in the field.
When it comes to maintenance management, it is essential to investigate the cause of the failure. By identifying the root of the problem, repetition can be avoided, and a precise maintenance plan can be drawn up. To support this process, several tools are available that enable effective failure analysis in maintenance. Check it out!
Failure Analysis in Maintenance: 5 Key Tools
There are several tools that contribute to investigating what may have caused a failure. The choice among them (or their combination) depends on what best suits the objective. That is, the techniques will mobilize certain information and offer respective insights into the problem. Therefore, it is important to be clear about your goal and how the tool works to make the most from it.
Ishikawa Diagram
The Ishikawa Diagram is used to study and identify the causes of a failure. It provides an opportunity for correction or learning to prevent future occurrences. With this tool, you can optimize the resolution process by focusing on the solution to the failure, not just its consequences.
The method was created in 1943 by a Japanese engineer, Kaoru Ishikawa. In addition to the tool, Ishikawa published several books and is considered one of the greatest authors on quality management.
Over time, the diagram was refined, and the application of the cause-and-effect model became applicable to different businesses and sectors. Thus, the method became widely known and received other names, including: Cause and Effect Diagram, 6 Ms, Cause Tree, and Fishbone PDCA. Despite some specificities in one or another, the essence of the tool remains the same.
How to Apply the Ishikawa Diagram
Applying the method is quite simple, but there are some basic steps to achieve success with maintenance failure analysis.
- 1st Select the failure or process to be analyzed:
The first step is choosing what you intend to analyze. The Ishikawa Diagram helps understand the cause-and-effect relationship—that is, the set of factors that led to a particular result. Therefore, it can be useful for understanding failures, as well as inefficient processes or those with frequent errors.
- 2nd Draw the fishbone:
The next step involves drawing the analysis structure. To do this, place the problem to be investigated at the far right of the central line. From there, you need to write the possible primary and secondary causes on the transversal lines.
- 3rd Analyze cause and effect:
This is the moment to evaluate all the causes and list the most critical ones. This establishes prioritization for resolving each of them and, later, allows you to study the impacts of your changes.
- 4th Plan solutions:
After analyzing and defining priorities, it’s time to plan solutions. Here, it’s useful to design an action plan, starting from the most critical points and moving to those of lesser criticality. Additionally, it’s important to establish the paths for these solutions, identify necessary resources (if applicable), and, of course, define the execution timeline.
- 5th Validate the results:
Finally, it is essential to evaluate the results of the changes. This way, you can verify whether the method was performed correctly and if the identified causes were indeed the true ones. The analysis should be recurrent and, therefore, undergo adjustments if the expected improvements are not identified.
Analysis of the 5 Whys
Perhaps this is one of the simplest failure analysis methods, as the technique basically consists of constructing logical answers to a series of “why” questions. You start with the problem and keep asking “why” to uncover what caused it.
The tool emerged in the 1950s, developed by the architect of the Toyota production system, Taiichi Ohno. The methodology gained widespread adoption due to its efficiency and simplicity in solving internal problems.
However, if the team is unprepared or unaware of the most common failure modes, the analysis might remain superficial or point to incorrect causes. Additionally, for certain asset components, the “5 Whys” method is suitable, but for others with multiple possible causes, it might be insufficient.
For example, components with well-defined construction and maintenance parameters can benefit from the “5 Whys” analysis. These include seals, bearings, etc., as they have very characteristic and generally single-cause failure modes.
However, failures caused by human issues, organizational factors, or a combination of elements may not benefit as much from this technique. In such cases, it’s advisable to consider another investigation method or combine it with other tools.
How to Apply the 5 Whys Analysis
- 1st Describe the problem:
The first step is to describe the problem. That is, a bearing defect, strange noises in the hydraulic system, etc. It is essential to be objective and clear about defining the problem so that the root cause can be discovered.
- 2nd Ask “why”:
At this point, ask “why” the problem occurred. Then, ask “why” to that answer, and so on. The idea is to reach the root cause of the problem after approximately five “why” questions.
- 3rd Plan and implement corrections:
Once the cause of the problem is discovered, it’s possible to think about possible resolutions. For example, if you discover that the strange noise in the hydraulic system was due to wear on the seal, you can then plan for replacement and implement a maintenance plan, if necessary.
- 4th Evaluate and Monitor:
If the resolution eliminated the problem, the method was effective. Otherwise, either the measure was not effective, or the correct cause was not identified. Therefore, it’s important to monitor the process after executing the action to evaluate its effectiveness.
Fault Tree Analysis (FTA)
The FTA is a tool for mapping possible scenarios. It aids asset management by documenting potential failures and mapping the probability of occurrence. Thus, combined with the concept of criticality, the manager can make decisions regarding the choice of maintenance model, for example, among a series of other issues.
The method emerged around 1960 and continues to be used in various industrial sectors. The tool allows for the calculation of the probability of basic events occurring. Based on this information, the diagram can be constructed following a logical sequence:
How to Apply Failure Tree Analysis (FTA)
- 1st Define the problem to be analyzed:
At the top of the tree, you should place the problem or failure to be analyzed. It should be objective and concise, making it possible to delineate the logical sequence that led to it. For example, a water circulation pump shutdown.
- 2nd List intermediate causal events:
At this stage, define the possible intermediate events that could indicate a failure. In this example, the pump shutdown can occur if any of the basic events are met. That is, in a pump shutdown case, the cause could have been cavitation, corrosion, or scaling.
- 3rd Identify main causal events:
A breakdown or second level of the Failure Tree should point to the root causes — those that led to the intermediate events. For example, in the case of cavitation, the cause is insufficient suction pressure at the pump or the liquid transforming into vapor. In this way, air bubbles create stress on the pump, causing mechanical damage.
Read more about Fault Tree Analysis here.
Pareto Analysis
Pareto analysis gave rise to the 80-20 rule idea. That is, 80% of problems originate from 20% of the causes. And, therefore, the remaining 80% would cause 20% of the problems. Thus, the method allows for understanding and quantifying the causality of events.
The tool emerged in 1941, developed by Vilfredo Pareto in Italy. Pareto was an economist who observed that 80% of his country’s wealth was concentrated in the hands of 20% of the population, causing economic imbalances.
In the case of maintenance, the principle is the same: a few causes generate the vast majority of failures. For example, almost 80% of failures in belt conveyors have practically only three causes.
How to Apply Pareto Analysis
- 1st List the problem and causes:
The first step is to identify the problems and their causes by collecting and organizing data. You can select a key problem, such as a failure in an electrical transformer. From there, identify the possible causes: manufacturing defect, short circuit, aging, corrosion, etc.
- 2nd Order the list of possible causes:
Organize the causes from most common to least common. This way, you can identify those that are most recurrent and deserve greater attention or correction as quickly as possible.
- 3rd Organize the data:
The third step is to organize the data into a bar chart that relates problems vs. frequency. This will serve as a basis for maintenance planning and for minimizing or resolving the problem.
Failure Analysis in Maintenance: Automated Detection
The four tools above are manual and involve a high degree of subjectivity. Maintenance decisions are often based on technical knowledge, past experiences, or even intuition.
Now, professionals who prefer to make decisions with greater agility, assertiveness, and reliability tend to choose tools that offer higher accuracy. One such recommendation is Dynamox’s automated detection tool (DynaDetect).
Although its objective isn’t to analyze the scenario as a whole or evaluate the factors that generated the failure, DynaDetect is valuable for precisely pinpointing the problem. Thus, professionals who desire a macro analysis can use it in conjunction with other tools to increase the assertiveness and speed of diagnosis.
DynaDetect operates as a specialized tool, analyzing sensor data and quickly diagnosing various failure modes. It generates alerts categorized by criticality level, notifying technicians of anomalies. In addition to diagnostics, it produces reports to streamline maintenance planning.
Automated Failure Identification
This tool enables the identification of various types of failures and provides detailed detection insights. As a result, decision-making becomes faster and more accurate. It is a powerful ally in failure analysis for maintenance.
Among the automatically detected failures are:
The goal, therefore, is to assist and increase the reliability and accuracy of technical analysis and support the manager in asset maintenance planning and decision-making. In this context of thousands of collected data points, it is essential that the analyst can focus on the machines and measurements that require the specialist’s attention.