This technology utilizes multivariate analysis on enormous amounts of data generated on a daily basis by equipment and sensors to detect signs of abnormality and equipment deterioration to help ensure stable operation of a plant.
In addition, we have developed a technique in which a mathematical model is created from multivariate data obtained during normal operation on a monthly or yearly basis. This realizes high-precision prediction and detection of signs of abnormality from a long-term perspective.
As a result of applying this technique to measurement data of an actual plant, we successfully caught signs of abnormality a few hours or days before failures occurred.
Application to abnormality sign analysis and deterioration diagnosis components of multivariate analysis technologies