In this paper the implementation of a convolutional neural network (CNN) as identification method for an industrial process is presented, also this identification is pretended to work as a sensorless application where data are missing from the sensor due to any kind of failure. For this matter, first the CNN has to be trained to identify process in order to used it later as the sensorless system. To validated the identification stage the Mean square error metric is used. This is essential because knowledge of the system is an important factor in order to realize further actions such as control techniques, which is a very important step in most industrial processes also allowing the system to keep working even if sensors are not responding correctly.