Prevention of molten iron/steel leak accidentsPrevention of major accidents,Stabilization of productionMaintenanceTemperature abnormalitiesInfrared thermographyData science

Data Science

Abnormality Judgment Logic Using Thermal Images

Improves the accuracy of abnormality detection by using temperature data from thermal images.

Solution Point

Possible to judge abnormality from only 1 thermal image.
Eliminates the effect of the vessel use condition by changing from threshold control to abnormality detection considering the temperature distribution.
Eliminates the effects of the refractory material and structure in the vessel by changing from threshold control to abnormality detection considering the temperature distribution.
Because thermal image data obtained by infrared thermography is used, simple and quick judgment is possible.

Features 01

Expected Problems

・There is room for improvement in abnormality judgments by a temperature threshold, as missed abnormalities detection and overdetection are still occurring due to changes in the condition of the process, differences in the refractory structure, etc., and the existence of hot spots depending on the measurement environment.

Features 02

Technology highlight

・Abnormality detection by instantaneous temperature distribution images captured by infrared thermography.

・Accuracy is improved in comparison with threshold control. In JFE's actual results, detection of molten iron leaks just before occurrence: 100% and overdetection: 0.01% were achieved.

Features 03

Proposed Solutions (Examples)

・Advice on installation position of thermal imaging device.

・Supply of logic for abnormality judgment from thermal image data.

Links 01