Failure prediction on a ship’s radar: use case

Context

Military ships are equipped with radars, and the shipbuilder wanted to ensure that the ships could operate at sea for months at a time, and that maintenance had to be carried out either hot on board, or cold on land. Data recovery by the manufacturer takes place on land, and he wanted to focus on a posteriori analysis of radar behavior.

Need

The customer (the radar manufacturer) needed a non-intrusive solution for on-board radars, and needed to provide its business experts with access to easily understandable information. The customer wanted to anticipate failures as far as possible, to avoid critical equipment failure at sea.

Solution

DiagFit, coupled with an on-board data acquisition solution, enabled the business experts to diagnose radar anomalies quickly, enabling the manufacturer to provide a better radar maintenance service to sailors. DiagFit also enabled the data to be analyzed in order to devise a continuous engineering process for the product range.

Results

The model learned without any indication of failures in the data (in blind mode) and helped the customer to discover periods of radar malfunction.

Thanks to DiagFit, the customer was able to focus on the specific behavior of radar failures, and better understand them.