Universal monitor and fault detector in fielded generators and method

A method of monitoring and determining a fault condition in an engine driven generator, the method comprising:

• retrofitting additional sensors to the engine driven generator;
• sampling original equipment sensor data from original equipment sensors;
• sampling additional sensors data from additional sensors, said additional sensors including:
• an oil pressure sensor;
• an intake pressure sensor;
• an acceleration sensor;
• a hall effect sensor;
• an A phase voltage;
• a B phase voltage;
• a C phase voltage;
• a neutral voltage;
• an A phase current;
• a B phase current;
• a C phase current;
• a neutral current;
• and an ambient temperature sensor;
• performing a learning algorithm on the sampled original equipment data and on the sampled additional sensors data;
• processing a vibration data of the additional acceleration sensor;
• decomposing frequency bands of a discrete wavelet transformation of the vibration data into decomposition frequency bands;
• analyzing the decomposed frequency bands for normal values;
• using peak values of the decomposed frequency bands and a respective known proximity to the additional acceleration sensor, determining a number of firing cylinders and a firing order;
• determining a set of normal operating values for the engine driven generator from the performed learning algorithm on the sampled original equipment data and on the sampled additional sensors data;
• exiting the learning algorithm;
• resampling the original equipment sensor data from the original equipment sensors;
• resampling the additional sensors data from the additional sensors;
• generating a vibration fault indicator from analyzed decomposed frequency bands peak event values when the determined normal values are exceeded;
• and
• providing the processed vibration data and the generated vibration fault indicator to a user.