Neuron soundware developed a solution for machine health monitoring in the chemical industry and explosive zones. The monitoring technology successfully detected faults and wrong machine parameters on a paint recycling machine. These faults would be difficult to discover by other diagnostics methods otherwise.
Every such fault causes production break losses, the need for machine cleaning and material reprocessing, and in the worst-case scenario machine critical damage with tens of thousands of euros of repair costs.
The challenge: eliminate paint recycling machine outages
For our customer, a leading chemical company, manufacturer of packaging and graphic solutions, color and display technologies, and other products for the automotive and healthcare industries, the key requirement is smooth production and elimination of sudden outages
To increase production and maintenance efficiency the manufacturer implemented a Neuron soundware solution for machine health monitoring of paint recycling machines that were so far maintained using the approach “run to a failure”.
The situation: recycling machine outages impact smooth production
Neuron soundware solution monitors paint/lack recycling machine that processes old material with the addition of a special chemical. It is running in the loop for 4-8 hours to grind and mix the material back into a usable product (paint). The most important part is the grinder with a set of porcelain pins inside – if these pins break, the maintenance costs are in thousands of Euros.
The problem when the machines stop means not only that the customer has to repair the machine, clean it, take out the material, and put it again to processing, but sometimes when the machine stops working in the middle of the production cycle, the material can no longer be processed and must be discarded. Another problem appears when the machine produces the material with the wrong settings (viscosity) – then the customer has to discard it as well
Early detection of incoming failure or wrong machine settings is very beneficial for the company as until now the maintenance team waited for the machine to fail or finish the production cycle with insufficient settings.
To protect the uptime of the machine, the customer needs to understand machine health in real-time including receipt of alert once the machine condition starts to worsen. In that way maintenance and repair can be planned, unnecessary breaks in production are eliminated and machine park usage is optimized as regards primary (eg downtime cost) or secondary (eg end-customer reputation cost) impact.
The solution: AI-based remote monitoring
Reflecting the customer need, Neuron soundware implemented the sound-based predictive maintenance solution using 5 sensors for collecting the machine sound data. These sensors have been connected to the nEdge (IIoT device), where the data is captured and processed.
The complex Machine Learning algorithm in nEdge has been analyzing the collected data and raising a real-time alert upon detection of anomalous behavior compared to machines´ nominal. In the case of regular behavior, the online report has been showing standard performance in detail, confirming the OK status of the machine.
nEdge has been connected to the internet, and the machine has been monitored remotely through the Neuron soundware cloud portal. The project started with the recording of nominal sound and vibration data. Soon after, the end-to-end anomaly detection service was activated. So now, in case of any anomalous sound, the system alerts the operator after 30 seconds to avoid further damage or material losses as soon as possible. The installation is ATEX certified.
Examples of detected anomalies
The clogged filter. On 19.10.2021 an anomaly was detected by the NSW system. The maintenance team checked what happened during that time and found that the product had not reached the machine. After this investigation, the maintenance team opened the filter of the machine and it was all clogged (see photo below). The filter has been cleaned and started further production without problems. The NSW model worked correctly and correctly registered the anomaly.
Machine misconfiguration. On 23.12.2021 NSW AI algorithm detected another anomaly. When the maintenance team checked the machine, they found out that the system had transferred the product with the process parameters of the cleaning. The model has detected the differences in sound between cleaning with the solvent and cleaning with a product (completely different viscosity). This is a good result because the system that the customer is using nowadays didn’t detect this. After the failure, the material had to be unloaded and the machine needed the actual cleaning with solvent.
If not detected on time, both these anomalies would have most likely resulted in unscheduled downtime and costly repairs. Thanks to our solution´s high sensitivity resulting in early alerts, the customer maintenance team resolved both anomalies very fast without additional downtime and additional costs.
- Saving maintenance costs and increasing maintenance efficiency
- Preventing damage of the machine – notification sent directly to customers system when NSW solution detects suspicious sound
- Time & material savings (e.g. wrong settings of the machine detected)
- Increasing production efficiency
- Unsolvable problem by conventional monitoring systems is covered by Neuron soundware
Conclusion: Neuron soundware machine health solution enables cost-effective maintenance
Thanks to the Neuron soundware solution the maintenance team has real-time, online information about the machine condition and material settings and can check on our friendly visual dashboard the status of the machine remotely from any device with internet access. The operator or maintenance team gets warned when the machine condition deteriorates.
The maintenance specialist can prioritize the assets inspection at the company level. Such a solution means higher efficiency of employee time, minimized costs associated with failures, and asset life extension. Our solution has also proven to detect issues beyond the reach of standard methods in this case.