Abstract
Fouling is a permanent problem in process technology and is estimated to cost 0.25% of the gross national product. Evaporative cooling systems are especially susceptible to air-side fouling: as they work with untreated outside air, they are exposed to both natural (e.g., pollen) and human-made (e.g., industrial dust) contaminants. In addition, suspended solid particles and dissolved salts in the spray water are an issue. In this study we analyzed an approach for fouling detection based on a semi-physical (grey-box) cooling tower model which we calibrated with measurement data. A test series with reliable laboratory data indicates good applicability of the model. In three datasets, the performance decreases due to fouling (scaling, which was provoked intentionally) in the range of 5–11% were clearly detected. When applied to measurement data of two cooling towers in real applications, the model also proved to be well calibratable with relatively little data (two to four operating days). For two data sets, the model yielded reasonable results when applied to long term data: a cooling tower cleaning could be retraced and nominal operation was verified during the remaining time. During the analysis of a third data set a temporary performance deviation was found, which could not be explained with the recorded data. Thus, the approach turned out to be generally applicable but requires further verification and refinement in order to increase the robustness. If successful, it can be transferred to a commercial product for need-oriented maintenance in order to reduce cooling tower operating costs and environmental impact.