Loading...
1 results
Search Results
Now showing 1 - 1 of 1
- Production Optimization Indexed to the Market Demand Through Neural NetworksPublication . Mateus, Balduíno Patrício César; Cardoso, António João Marques; Farinha, José TorresConnectivity, mobility and real-time data analytics are the prerequisites for a new model of intelligent production management that facilitates communication between machines, people and processes and uses technology as the main driver. Many works in the literature treat maintenance and production management in separate approaches, but there is a link between these areas, with maintenance and its actions aimed at ensuring the smooth operation of equipment to avoid unnecessary downtime in production. With the advent of technology, companies are rushing to solve their problems by resorting to technologies in order to fit into the most advanced technological concepts, such as industries 4.0 and 5.0, which are based on the principle of process automation. This approach brings together database technologies, making it possible to monitor the operation of equipment and have the opportunity to study patterns of data behavior that can alert us to possible failures. The present thesis intends to forecast the pulp production indexed to the stock market value.The forecast will be made by means of the pulp production variables of the presses and the stock exchange variables supported by artificial intelligence (AI) technologies, aiming to achieve an effective planning. To support the decision of efficient production management, in this thesis algorithms were developed and validated with from five pulp presses, as well as data from other sources, such as steel production and stock exchange, which were relevant to validate the robustness of the model. This thesis demonstrated the importance of data processing methods and that they have great relevance in the model input since they facilitate the process of training and testing the models. The chosen technologies demonstrated good efficiency and versatility in performing the prediction of the values of the variables of the equipment, also demonstrating robustness and optimization in computational processing. The thesis also presents proposals for future developments, namely in further exploration of these technologies, so that there are market variables that can calibrate production through forecasts supported on these same variables.