Predictive Maintenance for a Ventilator Using LSTM Algorithm

A ventilator is a device or machine that supports and assists breathing. Ventilators are needed by patients who have respiratory problems, either due to an illness or injury. The ventilator pumps air in a few seconds to deliver oxygen to the patient’s lungs. This study intends to apply the Long Short-Term Memory (LSTM) method as a maintenance prediction on the ventilator system to simplify and predict failure that would occur. By using predictive maintenance, it can be carried out at the right time before equipment failure occurs. The LSTM scheme is one of the algorithms that can be used to predict the treatment of the ventilator that needs to be done in the future. The highest result was obtained; the probability of failure in 50 cycles is 98.4% with an accuracy of 82%.