Determination of a fault status of a boiler

A method and model for performing diagnostics or fault determination of the operation of a boiler or water heater. A computer model is trained using electrical energy time series data and outlet temperature time series data. The model identifies a cold flow events 304 in the outlet temperature data, such as the temperature being below an expected threshold. A time window coinciding with the cold event is selected. Electrical energy data and outlet temperature data within the time window is analysed to identify a fault status 305. A level of dependence, or relationship, between the identified fault and the time window of electrical energy data is derived. A computer model is created based on the level of dependence, the model can determine the presence of a fault in a boiler using an electrical energy data recorded in a time window preceding a cold event. Preferably, an alert 307, such as an audio/visual signal or electronic notification is generated when a fault status is identified 306. The model may utilise mathematical methods, such as comparison of shape functions and classification algorithms including K-nearest neighbours (KNN).