A hot water supplier monitoring system includes a monitoring server monitoring hot water suppliers via a communication network. The hot water supplier supplies hot water by adjusting temperature of hot water heated by a heat exchange unit using combustion heat from combustion operation. The monitoring server accumulates operation data of combustion operation transmitted from the suppliers, and predicts and notifies malfunction occurrence for each supplier based on trend of change in the data over time. The data includes number of flame quenchings of the combustion unit in combustion operation. The monitoring server divides the suppliers into groups according to installation areas, compares the number of quenchings of same period among the groups, compares with previous number of quenchings of same group, and, for a specific group determined having an abnormal number of quenchings, excludes the data determined having an abnormal number of quenchings and predicts malfunction occurrence for each supplier.

Techniques for electrical submersible pump equipment fault analysis include training, using first training data, one or more machine learning models to categorize previously unseen images of electrical submersible pump equipment into one or more categories of a plurality of categories. The techniques further include training, using second training data, the one or more machine learning models to generate captions for the previously unseen images of electrical submersible pump equipment.