Modeling of high voltage induction motor cooling system using linear regression mathematical models

Abstract

Achieving reliable power efficiency from a high voltage induction motor (HVIM) is a great challenge, as the rigorous control strategy is susceptible to unexpected failure. External cooling is commonly used in an HVIM cooling system, and it is a vital part of the motor that is responsible for keeping the motor at the proper operating temperature. A malfunctioning cooling system component can cause motor overheating, which can destroy the motor and cause the entire plant to shut down. As a result, creating a dynamic model of the motor cooling system for quality performance, failure diagnosis, and prediction is critical. However, the external motor cooling system design in HVIM is limited and separately done in the past. With this issue in mind, this paper proposes a combined modeling approach to the HVIM cooling system which consists of integrating the electrical, thermal, and cooler model using the mathematical model for thermal performance improvement. Firstly, the development of an electrical model using an established mathematical model. Subsequently, the development of a thermal model using combined mathematical and linear regression models to produce motor temperature. Then, a modified cooler model is developed to provide cold air temperature for cooling monitoring. All validated models are integrated into a single model called the HVIM cooling system as the actual setup of the HVIM. Ultimately, the core of this modeling approach is integrating all models to accurately represent the actual signals of the motor cooler temperature. Then, the actual signals are used to validate the whole structure of the model using Mean Absolute Percentage Error (MAPE) and Root Mean Square Error (RMSE) analysis. The results demonstrate the high accuracy of the HVIM cooling system representation with less than 1% error tolerance based on the industrial plant experts. Thus, it will be helpful for future utilization in quality maintenance, fault identification and prediction study.

Extended Kalman Filter design for sensorless sliding mode predictive control of induction motors without weighting factor: An experimental investigation

“…The 3-phase feed voltages ( v g ) are linked to the rectifier’s entirely regulated bridge of power transistors through the filter’s inductances ( L g ) and resistances ( R g ). The vector formula indicates the incoming current in the αβ frame [ 20 , 36 ]. where , , represents the incoming I’s vector, V’s supply line, and V’s produced from the converter, respectively.…”

An integrated method for the leakage fault mode diagnosis and life prediction of the reactor coolant pump

“…The main steps of data-driven methods are data preprocessing, feature extraction, feature selection, model selection, and model evaluation [ 15 – 17 ]. Data-driven model-based methods mainly includes methods based on statistical regression [ 18 ], similarity [ 19 ], and the stochastic process [ 5 , 20 ]. Zhang first used the bootstrap method to resample the raw data, subsequently trained neural networks with each subset of the obtained data, and finally comprehensively processed the diagnostic results of these neural networks [ 21 ].…”

Automated determination of optimal component design for a binary solvent for absorption-based acid gas removal

“…Multivariate regression modelling is one of the most common methodologies for this purpose and can be used to improve energy efficiency (EE) [8]. Statistical modelling helps understand the most significant factors for optimizing energy systems which can be seen in previous studies [9][10][11][12][13][14][15][16][17][18].…”