IIOT World Manufacturing & Supply Chain Days, Dec 2022
It was a pleasure to be invited to join this distinguished panel at the IIOT World Manufacturing Day.
A couple of the highlights of the discussion were:
- Predictive Analytics to predict manufacturing equipment failures
- Benefits of Predictive Maintenance
- How AI/ML can improve Manufacturing results
- Why a CEO should care about Manufacturing data
Let’s explore those topics a little more in this brief blog post.
Predictive analytics to predict manufacturing equipment failures
Traditional scheduled maintenance, e.g. once a week, or every x hours of usage, can still lead to downtime if parts fail earlier than expected, or wastage due to replacing parts too early. For example, a bearing life could normally be 5,000 hours, but it could fail after 3,000 or last 10,000 hours.
When a machine failure occurs there are usually clues amongst the data that provide an early warning, however, finding these clues can be difficult!
The use of AI/ML in this area provides the ability to identify difficult to find patterns that take place leading up to a failure across 100’s or thousands of data items. These patterns enable early prediction of future problems, allowing action to be taken in a planned and controlled manner, avoiding catastrophic failures.
What are the benefits of Predictive Maintenance over traditional maintenance?
- Avoid downtime: an hour of downtime can cost hundreds of thousands of dollars, this is bottom line impact to your business and profitability
- Reduce maintenance costs: Conducting unnecessary maintenance, such as replacing parts too early is a considerable waste of resource, and cost of parts
Production Optimisation
Operations will often talk about finding the sweet spot, the optimal settings to deliver the best production output, both volume and quality. This will vary based on raw materials, the product being produced that day, and even the weather.
AI/ML is one option to identify the best settings to optimize production – by building up a library of data including machine settings, volume produced, output quality, temperature and humidity, AI/ML can identify the best process settings for the product being produced, given data on the raw material.
In the past this would have been done by expert operators with years of experience, but now it has been demonstrated that AI can often make better decisions.
Real-time Quality Inspection
Traditionally quality inspections are usually undertaken by sending samples to the lab for review or testing. This process takes time and can mean that large batches of product are on hold until the quality is approved. If a problem is found then a whole batch might need to be scrapped.
While this approach is still useful in many circumstances there are alternatives…….
AI powered vision systems are capable of analyzing and inspecting products while they are moving through the production process. This means that errors or problems can be detected early in the process and removed from the line, and corrections can be made on the fly to avoid large volumes of waste.
Why should your CEO care about AI/ML, IIoT and collecting Manufacturing data?
The short answer……Reduced downtime, improved OEE! Downtime is wasted opportunity and less production, quality defects are lost production. In addition, AI can recommend the speed of the line to maximize output. All of which impact your profitability.
Improved transparency of business performance – this data provides greater insights into what is really happening within your production lines.
Higher revenue – use of these approaches maximize your current investment in equipment, providing increased capacity, with minimal extra investment.
Bottom line: Improved profitability – these approaches reduce unit costs, remove wastage, and improve overhead absorption.
To see the full recording of this session, please click this link.
………..
Thank you for reading this article, feel free to browse our other posts.
Industry 4.0 / IIoT 