"SELF-PREDICTIVE GVA FUNDS" GVA I4.0 - SELF-PREDICTIVE - AUTOMATION OF PROCESSES THROUGH TECHNOLOGIES OF ANALYSIS, PREDICTION AND INTEGRATION OF APPLICATIONS AND SYSTEMS DESCRIPTION One of the strategic lines of action that AIDIMME carries out in the field of research and development, consists in the development and application of different elements (physical devices, software, methodologies, etc.) based on enabling technologies of Industry 4.0. These elements facilitate the transition of manufacturing companies in our environment towards smart manufacturing. The PREDICTIVE AUTOMATION project pursues the design of a predictive process control system, whose function is to warn of the possibility of anomalies, from the capture of data in real time, its analysis and modeling. The project consists of a part of study and analysis of the bibliography and previous experiences that have been developed in this field, and another part that experiments with data from different manufacturing processes to define to what extent the behavior of the same and use these results to reduce human intervention in the process. OBJECTIVES The ultimate goal is to achieve autonomous processes that are capable of managing the manufacturing flow with minimal external intervention. For this, you must previously know what is happening in the processes and why it is happening, for which you must know the status of the process in real time. Once the appropriate operating conditions have been established, the process will be predicted as the relevant parameters are modified due to the variability associated with the process itself. And finally, when all of these conditions are met, the predictive system can be autonomously managed by the process. The specific objectives of the project for 2020 are the following: Develop an intelligent application in a robotic cell Generate a methodology to approach intelligent process automation. • Validate the methodology in a controlled environment using a reduced-scale manufacturing line Project Number: 22000005 Grant Agreement: IMAMCC / 2020/1 Duration: From 01/01/2020 to 31/12/2020 Coordinated in AIDIMME by: SÁNCHEZ ASINS, JOSÉ LUIS R&D Line: INDUSTRY 4.0 EXPLANATORY VIDEO RESULTS OBTAINED Year 2021: The conclusions of the experiment carried out to validate the hypothesis are: - The modeling of a process that tries to reproduce a change of state in it, from state A to state B, should not be carried out by reproducing the conditions of both states separately, but rather it must reproduce the change of state. Otherwise there is not enough information to reliably detect when the state change will occur. That is, you cannot act preventively since the generated algorithm will only be able to identify one of the two states but will not be able to predict the change. In the algorithm trained in the experiment carried out, precisely this situation occurs, since it acts only as a state classifier. - Depending on the problem to be solved that has been prioritized in point E of the methodology, it is possible that data collection will take a long time. Such is the case of the loss of sharpness of wood drill bits, since the useful life of this tool before it begins to drill in poor conditions can range between two and six months depending on the workload. Other wear or malfunction processes can be much longer, so when you decide to approach predictive automation of a specific process, you must evaluate the time required for state changes to occur, and act accordingly. - If the experimentation time is considered excessively long, you can act as in the experiment carried out, extracting data from the two states separately as long as they can be reproduced. Once the classifier algorithm is available, it must be validated in the real process, estimating a percentage of “bad” results that indicate that a state change will occur soon, which must also be contrasted experimentally. It is possible that this contrast also requires a long period of testing, as has happened with the case studied. Año 2021: Year 2020: Results 2020 Deliverables:    PUBLISHED NEWS general broadcast https://actualidad.aidimme.es/2022/04/19/mejora-eficienc (...) https://www.interempresas.net/Robotica/Articulos/385646- (...) https://actualidad.aidimme.es/2020/09/28/aidimme-formato (...) https://actualidad.aidimme.es/2020/12/30/especial-difusi (...) http://actualidad.aidimme.es/2020/07/29/aidimme-concepto (...) http://actualidad.aidimme.es/2020/11/26/empresas-partici (...) http://actualidad.aidimme.es/2020/07/01/proyectos-apoyo- (...) http://actualidad.aidimme.es/2019/12/18/automatizacion-p (...) GRANT Total Agreement: € 2.803.920,44 TARGET AUDIENCE AND IMPACT MEASUREMENT 2665 Access to the project website 3924 Access to news published on own websites. Total Accesses: 6589 Objective Sectors by CNAE CNAE: 3101 - Manufacture of office furniture and e - Number of target companies: 574 (115 in the Valencian Community) CNAE: 3102 - Manufacture of kitchen furniture - Number of target companies: 666 (155 from the Valencian Community) CNAE: 3109 - Manufacture of other furniture - No. of target companies: 5579 (1327 in the Valencian Community) TRACTOR COMPANIES Thanks to its support and signature of the "declaration of participation" the project has been funded. ROYO SPAIN, SL MICUNA, SLU COMPANIES DIFFUSION R&D PROJECTS They want to know first-hand the evolution of the project, and its progress to the final result. COMPANIES TRANSFER KNOWLEDGE They will implement technologies, develop strategies or look for new models based on the results.