Design and Development of an Industrial Edge System to Continuously Monitor Metalworking Fluids

Every year 400 million liters of oils are used for metalworking fluids in the US alone. A large portion of that oil is disposed of earlier than necessary due to insufficient maintenance and care. By keeping parameters like concentration and pH within an optimal range, the fluid lifetime is expected to be extended by about 2-4 times. To achieve this goal it is needed to monitor fluid characteristics continuously and accurately. Therefore, the main focus of this study is to develop an industrial system to continuously monitor lubricants health parameters. The system was projected to follow an Edge-Cloud architecture, where the Edge side is responsible of communicating with the sensors, control the actuators, collecting, processing and sending the data to Cloud. While the Cloud stores the data coming from all the Edge systems and provide a user interface to visualize the collected data. EdgeX framework was chosen to develop the Edge side of the architecture given its great capabilities and extended usage in the company. During the project, there were built two prototypes. The first to study the operation of an innovative concentration sensor developed by the Research Institute of Sweden (RISE), and a second one to test an automatic cleaning system and perform data analysis to validate the possibility to correlate the lubricant viscosity with the impedance and phase measured by the concentration sensor. The results obtained from the analysis are promising, the model created highlights how the idea of estimating viscosity from impedance, phase and temperature is viable and feasible, but the dataset with which the model was trained needs to be improved and updated with correct data, as it was found that the viscosity sensor was returning conflicting results during different measurements. This paper shows the whole process followed during development, from the requirements gathering phase to the testing phase. In the end, future improvements are proposed to solve the problems found and to add new features.