How To Perform Root Cause Analysis In Automation Systems

Root cause analysis (RCA) is an important process in many automation systems. It can be a powerful tool for manufacturers, as it goes right to the heart of manufacturing errors and defects. By identifying and fixing the most fundamental – ie, root – causes of these problems, you’ll be able to enhance your product quality to a finer degree. This will improve customer satisfaction and trust, reduce warranty claims and cut the costs of reworking faulty products. 

Best practice in traditional RCA requires that you conduct the following steps: 

• Identify a problem

• Collect available data

• Identify contributory causes

• Pinpoint and prioritise the root cause(s)

• Implement and monitor a solution

Identifying A Problem

First, you have to recognise that a problem exists and then identify exactly what that problem is. You also need to consider what effect the problem might have on production and whether this will affect customer needs.

Collect Available Data

Next, you need to gather all relevant data relating to an incident or defect. Make a list of its defining characteristics and parameters. Include things like what factors may have contributed to the problem, when or how often it occurred and what its impact was.

Identify Contributory Causes

It’s a good idea to create a timeline of events to help identify any contributory causes of the error, defect or problematic event. A project team should brainstorm the issue, implementing exploratory techniques such as ‘The 5 Whys’ method or a causal graph like a Pareto chart.

Pinpoint The Root Cause(s)

Using the above techniques, the analysis team can narrow down the potential root cause(s) of the issue and what factors might have contributed to it. If more than one fundamental cause of the problem is identified, you need to prioritise them in order of impact magnitude and number of contributory factors.

Implementing And Monitoring A Solution

The final step in traditional RCA is to find and implement a solution. Another brainstorming session might help. Extend your investigation to include gathering input from other people involved, such as machine operators, line managers or stakeholders. Collect ideas and recommendations from as many sources as you can and devise a solution that works for everyone affected by the problem. Once a solution has been implemented, its results should be monitored to ensure that the problem is effectively and permanently solved.

What Kind Of Root Cause Analysis Can You Perform?

RCA currently falls into two major categories: manual and automated. 

Manual RCA means that you determine the root causes of any manufacturing process failure by investigation or manual data analysis. This may include a variety of approaches, such as: 

• Gathering data manually from different sources 

• Brainstorming sessions 

• Investigative methods like Failure Mode & Effects Analysis (FMEA) 

• Visualisations like diagrams or charts 

• Excel data spreadsheets

• Data analysis tools 

Automated RCA does the same job with machine learning and/or advanced statistical methods. This is usually done using a single software solution that analyses tests and processes data and can be used as a supplement to manual methods.

Automating RCA Systems

Many manufacturers are still using manual RCA methods, but there’s a substantial argument that automating the process is more effective for the following reasons:

1. Automated RCA saves time and money on data collection and analysis and makes it less likely that production needs to be halted to investigate a critical defect. A software solution connects all manufacturing and test data, which can be accessed at any point for real-time analysis. 

2. Automated RCA software contains all the necessary tools such as big data analytics and machine learning, so engineers can handle defect investigations directly from the shop floor. 

3. Automated RCA can increase uptime, as production doesn’t necessarily have to slow down or stop for a problem to be investigated. Modern software solutions are capable of processing vast amounts of data in real-time, rapidly analysing the frequently convoluted relationships that exist between various processes, components and manufacturing parameters. 

4. Automated RCA can employ AI to identify and analyse multiple factors contributing to failure or malfunction, including mechanical defects, human error or unexpected variations in source materials or components. With a data-driven approach to RCA, quality controllers and engineers have immediate access to the data analysis of the most likely contributing factors.

5. Automated RCA factors out human error, often caused by stressful and unrealistic deadlines. It relies more on pure objective data rather than incomplete, biased, conflicting or inaccurate information gathered from machine operators or line managers. It will be as accurate as the manufacturing data supplied and will always be faster. 

6. Automated RCA can scale as the manufacturing landscape develops greater complexity. Platforms adapt effortlessly to new innovative technologies, different data sources, and changing manufacturing methods. This means that you can be sure of continuous and seamless improvement, regardless of how tools, technology or staff members evolve. 

Accelerating Root Cause Analysis

RCA is a vital part of the manufacturing process and makes a significant contribution to process optimisation. You can use it to identify and reduce inefficiencies in your operation, solve process bottlenecks and address part quality concerns. It promotes continuous improvement, allowing you greater freedom to innovate and keep abreast of industry developments. This is especially relevant in the world of Industry 4.0, as the industry shifts towards new technologies and newly innovative components. Manufacturing such recently developed products carries a higher risk of defects, as designs and processes are still evolving. 

You can speed up RCA in manufacturing and solve shop floor issues more quickly if you invest in predictive quality analytics software. A typical solution will begin with the targeted defect or failure and analyse all upstream data gathered from different parts of the process. In a few hours, the module will probably have identified those factors most likely to have contributed to the issue. Automated RCA systems are sufficiently flexible to incorporate skilled input from engineers to improve the end results.