top of page

Nimac Group

The Company

In 1977, the company NIMAC Group began the production of horizontal panel saws. The continuous effort for the improvement of the machines and the reliable service confirmed the company as one of the most important manufacturers of beam saws with a remarkable international presence. The gross percentage of exports, more than 50% of the production and the sales at the biggest furniture industries, are indicative of the company's success. Aiming always at the production of high quality and value-priced machines for the world wide marketplace.

The case


One of the most basic machines of production for Nimac is the Rapid Drill where all the basic processing of its products takes place. This particular machine works non-stop and is vital for the production.
Nimac was looking for a solution that would be able to alert before something goes wrong with the machine to arrange maintenance of the machine at a scheduled time.


The solution

Predictive Maintenance

The solution proposed by the Smiling Machines was the predictive maintenance of the basic axes of the machine. Basic axes of the rapid drill are more stressed and bring out the biggest problems. With this proposal there were several advantages

  • Low application cost

  • 85% fault coverage

  • Fault prediction up to 15 days

  • Reduced complexity of algorithms and data to be processed

  • 65% success rate forecast


Machine Simulation

The first step of the implementation was the creation of a digital replica of the rapid drill. In this way we achieved a better analysis of the problem and it allowed us to see the necessary measurements that we should monitor.


Advanced Solutions Using AI Sensors

After the digital replica was analyzed, the solution proposed to Nimac was the use of intelligent sensors for measuring vibration and temperature with active filters who are able to be adapted to the operation of the machine and send only the important data reducing the complexity of the application.

Deep Learning Analysis 

With the use of Deep Learning models we draw accurate results from large volumes of input data.
The result of the Deep Learning was the accurate predict of an issue very accurate even days before the issue become production problem.

For this specific application we perform fault diagnosis based on acceleration, vibration, sound and  temperature signals.

bottom of page