Elaboration and Analysis of Manufacturing Execution System data

The research work was carried out within the IT unit of Cornaglia Goup S.p.A., which develops cutting-edge technological solutions for the automotive world and is one of the most important Italian component companies, supplying the most prestigious customers on the European market. The study was born as an integration to an IT system already present in the company, the Manufacturing Execution System (MES). MES refers to a computerized system that has the main function of managing and controlling the productivity of a company. The company goal is to transform data into fundamental information to make the best strategic choices for the aforementioned company. Big Data, in fact, have enormous information potential that can help companies get to know themselves better (by optimizing their level of internal efficiency) and their reference market, with the possibility of anticipating developments and finding the best business strategies. In fact, at an operational and strategic level it is important to adopt a data driven approach to achieve business objectives. For this reason and for the need to identify behaviors, strengths and criticalities of the production process, it was decided to consider, as the basis of Tesi's work, the tables of the company relational database, coming from a process of manual data collection of company machinery, inserted within a SQL Server, which include records regarding the entire production process and the details of each operator on all productions. In reference to this, the framework is composed of several parts: the first phase of the study is the definition of what the problem is and the objectives, based on this the exploration of the data took place and the choice of which were useful or not to the final purpose. After having collected the useful data, they have been processed and organized so as not to be incongruent, incomplete or untrue.. Once this was done, the main step was to identify relevant characteristics that most of all influence the data and consequently future decisions on the production process, better known as Key Performance Indicators (KPI). The next step was to compare the available data and the calculated indicators to show them graphically so that they are more easily readable and understandable by any person, for example with histograms or pie charts. The main objective of this step is to give information and knowledge to those who will then go to make decisions, interpreting the results to make the best one. For the last step of this Thesis work, we focused on an important aspect of the company that causes most of the losses in terms of time and money, namely machine downtime caused by breakdowns. With the use of a Python script, the procedure involved the creation of an exponential distribution model that better captures the probability of a machine failing, based solely on the increase in production hours from last failure occurred and on a fundamental factor to prevent failures, namely the Mean Time Between Failures (MTBF). With this model, the probability of failure after each production record collected was calculated for each single company production machinery. The results of this thesis study are only theoretical and there is no real confirmation of the advantages that this solution has brought in terms of business benefits, it will become such once the recommended changes are made.