| |
| Acadêmico(a): Leandro Starke |
| Título: RITA: UTILIZANDO MACHINE LEARNING NO PROCESSO DE PUMP SIZING EM BOMBAS DE CAVIDADES PROGRESSIVAS |
| |
| Abstract: |
| This work presents the development of a prototype to define the ideal interference for the calculation of the rotor helicoid dimension in Progressive Cavity Pumps (BCP). The data used to prepare this prototype were provided by Netzsch do Brasil where they were organized and structured before being used. The development of the prototype was divided into three stages: data import in a structured way, classification and construction of the Dataframe with the characteristics and data needed to solve the proposed problem and, finally, the application of machine learning algorithms to perform the predictions. In the first step, the data used in the current process, stored in Excel spreadsheets, is converted and imported into a relational database. Then, in the second step, the data is selected from the database and imported into a Dataframe format. In this step, calculations are performed to define the interference. Finally, in the third step, the complete Dataframe is used to apply machine learning algorithms, more specifically, K-Nearest Neighbor (KNN) and Polynomial Regression algorithms. The results show that the prototype is able to partially perform satisfactory predictions from training data that are coherent and meet the input requirements. |
| |
|
