Synchronous motor condition monitoring recommendation system
Students Name: Soprun Oleh Viktorovych
Qualification Level: magister
Speciality: Systems and Methods of Decision Making
Institute: Institute of Computer Science and Information Technologies
Mode of Study: full
Academic Year: 2022-2023 н.р.
Language of Defence: ukrainian
Abstract: At a time when reducing energy consumption is a priority for end users and industrial facilities, permanent magnet synchronous motors are gaining a foothold in the electric motor market due to their efficiency. The popularity of this type of electric motor is associated with innovations, resulting in the prices of microprocessors and microcontrollers decreasing significantly, affecting the cost of electric motors. When they first appeared on the market, actuator control processors were expensive and lacked many features required for stable operation [1]. Regardless of the type, synchronous motors must be provided with a specific drive, also called control systems or controllers, that offer a wide range of functions to control the operation and condition of the electric motor. They can help the engine start in low voltage conditions, control several speeds or reverse, and protect against current overload while implementing a wide range of other functions. Some sophisticated motor control devices also help to effectively control the speed and torque of the motor. Even though these systems do well with monitoring the vibration level, frequency, and direction of rotation of the motor and its connections, some exceptions relate to the capabilities of these systems in controlling the temperature regime of synchronous motors. This issue is especially acute for synchronous motors with permanent magnets, since their traction electric drive design does not allow temperature measurements with sensors to be carried out with sufficient reliability, so temperature control requires additional temperature forecasting using other engine indicators and machine learning methods [2]. Within the frame of this master’s thesis, the problem of monitoring the condition of permanent magnet synchronous motors was described, as well as opportunities to improve the experience of using devices based on these electric motors for the end user. An analysis of systems that fully or partially solve the above problem was carried out, and the main advantages and disadvantages of these systems were determined. To solve this problem, a system analysis of the subject area was carried out using the method of the goal tree and the analytic hierarchy process, as a result of which the general purpose of the work and the type of the designed system were determined. To create a conceptual model of the system, a unified modeling language was used. Diagrams of use cases, classes, states, and activities, a sequence, and a deployment diagram were used[3]. The designed "Synchronous motor condition monitoring recommendation system" is designed to monitor the condition of the motor using the indicators recorded by the motor controller, as well as the value of the rotor temperature, which is predicted using the extreme gradient boosting algorithm [4]. Also, using the measured and predicted indicators, together with the k-nearest neighbor’s algorithm [5], the system will form recommendations for the user to increase his safety and improve the final experience of using a device with a synchronous motor. A set of sensor data collected from a synchronous motor placed on a test bench, which is a prototype model from the original equipment manufacturer and was collected by the LEA department of Paderborn University in Germany [6], was used to train mentioned machine learning algorithms. The Python programming language was used for the software implementation of the system, along with various modules for working with data and machine learning models [7]. The system database was implemented using the SQLite relational database management system [8]. The graphical user interface was created using the Qt platform [9]. As a result of achieving the general goal of the work, a synchronous motor condition monitoring recommendation system was created, which is capable of forecasting temperatures in real-time, using data measured by the sensors of the electric motor, as well as forming recommendations for the end user depending on the indicators of the current state of the motor, and on how the user interacts with it at the moment. Study object – the process of analyzing the condition of a synchronous motor to provide recommendations to the user of this motor device. Scope of research – methods, and means by which it is possible to analyze the condition of a synchronous motor and form recommendations for the use of a device with this motor. Goal of research: the creation of a synchronous motor condition monitoring recommendation system. The result of the research is a working synchronous motor condition monitoring recommendation system, which is designed to help users of synchronous motor devices, providing them with information about the condition of the motor and recommendations that are designed to help ensure stability, efficiency, and safety during motor operation. Keywords: recommender system, synchronous motor, monitoring, condition, machine learning algorithms, temperature forecasting. References. 1. The Advantages of Synchronous Motors. pumpsandsystems.com: веб-сайт. URL: https://www.pumpsandsystems.com/advantages-synchronous-motors (дата звернення: 01.11.2022). 2. Hongchang Ding, Xiaobin Gong, Yuchun Gong, "Estimation of Rotor Temperature of Permanent Magnet Synchronous Motor Based on Model Reference Fuzzy Adaptive Control", Mathematical Problems in Engineering, vol. 2020, Article ID 4183706, 11 pages, 2020. https://doi.org/10.1155/2020/4183706 3. Fowler M. UML Distilled: A Brief Guide to the Standard Object Modeling Language / M. Fowler, K. Scott., 2003. – 208 с. – (Addison-Wesley Professional). 4. XGBoost: Extreme Gradient Boosting — How to Improve on Regular Gradient Boosting?. towardsdatascience.com: веб-сайт. URL: https://towardsdatascience.com/xgboost-extreme-gradient-boosting-how-to-improve-on-regular-gradient-boosting-5c6acf66c70a (дата звернення: 01.11.2022). 5. K-Nearest Neighbor. medium.com: веб-сайт. URL: https://medium.com/swlh/k-nearest-neighbor-ca2593d7a3c4 (дата звернення: 01.11.2022). 6. Electric Motor Temperature. kaggle.com: веб-сайт. URL: https://www.kaggle.com/wkirgsn/electric-motor-temperature (дата звернення: 01.11.2022). 7. Why Use Python for AI and Machine Learning? steelkiwi.com: веб-сайт. URL: https://steelkiwi.com/blog/python-for-ai-and-machine-learning/ (дата звернення: 01.11.2022). 8. About SQLite. sqlite.org: веб-сайт. URL: https://www.sqlite.org/about.html (дата звернення: 01.11.2022). 9. About Qt. qt.io: веб-сайт. URL: https://wiki.qt.io/About_Qt (дата звернення: 01.11.2022).