Development of a Soft Sensor for Detecting Overpitched Green Anodes

Increase in coke variability over the past few years has led to frequent changes in the optimum pitch demand (OPD) defined as the amount of pitch that yields the best anode properties for a given dry aggregate mix. Exceeding the OPD may cause post-baking anode sticking problems, thereby requiring additional operator interventions when unloading the baking furnaces and increasing losses due to scrapped anodes. This paper describes a new soft sensor developed for early detection of anodes with the potential to have sticking problems based on electrical resistivity measurements. Five resistivities measured at different points on approximately 120,000 individual anodes were used to build a Principal Component Analysis (PCA) model. A correlation was found between the anode sticking potential and the electrical resistivity data. It was shown that a breakdown in this correlation, as measured by the statistics of the Squared Prediction Error (SPE), may be used to detect post-baking sticking problems.