ISA-ThermoLogycs

Our internally developed AI model predicts the final temperature characteristic curve of an arm with over 95% accuracy. This enables us for the first time to select suitable pairs before they are manufactured.

How it works:

Our software processes data from various sources, such as information on material composition or specific material properties. This data is collected and structured in a clear table. An intelligent algorithm then analyses this information, recognizing patterns and correlations as well as possible deviations. On this basis, a model is created that helps to optimize production processes and better predict future developments.

When precision starts at 0.1 °C

The AI analyzes the properties of the pins (e.g. 93.3 % copper, 1.6 % nickel, etc.) and predicts with over 95 % accuracy how the finished thermocouple will behave.
The graphic shows the suggested thermocouples from stock.

Recommendation:

• The gray dashed line shows the available stock (here only a part of the stock for clarity).
• The red dashed line shows the best fitting thermocouple pairing from stock.

Model validation in the tolerance framework: Measurement vs. estimation

The diagram illustrates how precisely the model reproduces real measured values and historical reference data within defined tolerances.

• Black dashed lines mark the tolerance limits within which the material should ideally be located. These limits define the permissible range for the characteristic under consideration.
• The orange solid line represents the actual measured values. It shows the actual progression of the characteristic over the distance or time under consideration.
• The orange dashed line represents the model-based estimation. This was created without knowledge of the actual measured values and shows how the model predicts the course of the characteristic.

The configuration interval, which is shown around the estimation, indicates the range in which 95% of the observed values were located for similar configurations in the past. It serves as a frame of reference for classifying the forecast.

The next step for ISA-ThermoLogycs

In manufacturing, every decision counts - especially when it comes to materials. What used to be a laborious and error-prone process is now efficient, reliable and well-thought-out thanks to intelligent software solutions. Our internally developed application supports planning and control in selecting exactly the right combination of pairs or arms from countless options.

But we are going one step further: in future, ISA-ThermoLogycs will be used at other key interfaces to make work processes even more efficient. In addition to the automated selection of optimal pairs of arms from stock, our employees will in future also receive targeted suggestions for the ideal production process - from the primary material to the end product.

These enhancements make the entire arm pair allocation process even more precise and almost completely automatable. The result? A system that thinks for itself, reduces the workload and ensures safety. Our customers notice this when the quality is right, the quantity is right and the delivery date is met.