The motor warns you
months before it breaks.
DeepMaint measures voltage and current at the control cabinet and finds the fault while it is still incipient. No motor teardown, no production stop, no model to train.
Done. The next detection case lands in your inbox.
Three steps from the cabinet to your screen
Measured at the cabinet
Six Hall effect current sensors and three voltage fuses on DIN rail, next to the rest of the motor switchgear. Voltage measured directly up to 1 kV; above that, from the transformer secondary. Converters of 16 to 24 bits.
Electrical signature analyzed
Every fault injects its own harmonics into voltage and current. A single capture holds up to a million possible harmonics. The UPV patented algorithm pinpoints the ones that reveal a fault and tracks how they grow.
Read in the web app
Motor health, power, efficiency and power quality, with indicators refreshed on the server every 2.5 minutes. No prior training: the indicators are physics based, explainable and comparable across motors.
What the customer found inside the motor
Three detections on machines in production, with the indicator moving weeks or months before the failure.
The rotor asymmetry indicator rose steadily from February to September. After more than six months of growth, the customer pulled the motor out of the well.
The stator condition indicator started climbing with the motor running. One month after the first alarm the customer sent it to the repair shop.
The eccentricity indicator grew clearly from late September. The customer chose to keep the motor running for twenty more days.
One device, four jobs
Everything comes from the same voltage and current measurement taken at the cabinet.
Predictive maintenance
Detects incipient faults in the motor, the drive and the load. It is the only device based on voltage and current during normal operation able to see short circuits between a few stator turns, one of the most critical and fastest evolving faults.
Power measurement
Active, reactive and apparent power and power factor per IEEE 1459, traced with over 20,000 points per second. Captures fast transients such as a direct on line startup, with peaks of 3 to 10 times rated power.
Efficiency
Traces the motor efficiency curve across the speed range and places the current operating point on it. If the load curve is known, as with a hydraulic pump, the whole set is optimized and not just the motor.
Power quality
Symmetrical components and imbalance, voltage and current THD, TDD, crest factor, sags, swells and interruptions. Each indicator computed per its standard: IEEE 1159, IEC 61000-4-7.
Twelve years of research at the Universitat Politècnica de València
UPV is the leading engineering university in Spain according to the Shanghai ranking and the national leader in patents: 44 filed and 18 licences issued in 2024 alone. Its Ideas Institute has spun out more than 800 companies. DeepMaint is one of them, already running at water sector companies, with deployments planned in energy and oil & gas.
Why it replaces the maintenance round
Genuinely non invasive
Installed at the electrical cabinet and measuring during normal operation. Nothing is dismantled and the production process is untouched.
Reaches inaccessible motors
Submersible pumps in aquifers 50 to 500 metres down, or motors in hazardous atmospheres, monitored from surface measurements alone.
No training, no history needed
Indicators are physics based and predefined. No model to feed for months and no fault history required for the asset.
One avoided stoppage pays for it
The cost of an unplanned stoppage and a multi component repair far exceeds fixing a single incipient fault you already located.
One unplanned stoppage costs more than the whole system.
We send each detection case as we publish it: which indicator moved, how long before the failure, and what showed up inside the motor. No filler, no sales team behind it.
Done. The next detection case lands in your inbox.