This project aims to create a digital twin of the CAES system for fault detection. A data acquisition model (DAQ) was created to collect data from the CAES system. The DAQ model consists of three one-wire temperature sensors, two voltage sensors, a vibration sensor, a current sensor, a pressure sensor, and a real-time clock for a timestamp. Additionally, an acoustic sensor was used along with four thermocouples. Data was collected and inserted into MATLAB, which was used to create a predictive model using supervised machine learning, specifically classification with bagged trees. The model is used to classify faults into four different classes of system health: healthy system (HS), fault 1 (F1), fault 2 (F2), and fault 3 (F3).