MANAGER FOR PREDICTION HEALTH STATUS OF PRODUCT LINE

摘要

The device (102) -central control unit for monitoring of the production line (100) has a new data acquisition product test results from the final stage of production (204) (phase testing of finished products) and based on the collected data creates models (103) for prediction and decisions of the product line (100). Each production line (1, ... M) (100) is characterized by the input parameters (200), which are needed to set up the working process and a certain set of processes (201,202,203 and 204), which describes the realization of production flow. Each process is characterized by a certain set of machines (1, ... N) to perform the procedural steps and lead the production process to the final stage (204). Manager (102) of the product line (100) collects the data from the final stage(204) of the production process and compares them with reference values and then creates a prediction model (103) and decision models (103) the results of which goes back to the very product line and / or the individual process (201,202,203 and 204). The database contains information concerning the history of the quality of the data and finalized product manufacturing history of each process separately. Manager (102) of the product line (100) based on the created model predictions can determine whether the fault of software and / or hardware nature, then it is human nature, then if there is a fault on one of the machines of the production process (201,202,203 and 204), and whether the whole process suffered certain anomalies et al. Manager (102) of product line (100) can be used in the processing of neural networks, eg. multi-layer feed forward neural networks.