Thermal management analysis of induction motors by combining an air-cooled system and an integrated water-cooling system

Thank you for visiting You are using a browser version with limited CSS support. For the best experience, we recommend that you use an updated browser (or disable Compatibility Mode in Internet Explorer). In the meantime, to ensure ongoing support, we are showing the site without styles and JavaScript.
Due to the operating costs and longevity of the engine, a proper engine thermal management strategy is extremely important. This article has developed a thermal management strategy for induction motors to provide better durability and improve efficiency. In addition, an extensive review of the literature on engine cooling methods was performed. As the main result, a thermal calculation of a high-power air-cooled asynchronous motor is given, taking into account the well-known problem of heat distribution. In addition, this study proposes an integrated approach with two or more cooling strategies to meet current needs. A numerical study of a model of a 100 kW air-cooled asynchronous motor and an improved thermal management model of the same motor, where a significant increase in motor efficiency is achieved through a combination of air cooling and an integrated water cooling system, has been carried out. An integrated air-cooled and water-cooled system was studied using SolidWorks 2017 and ANSYS Fluent 2021 versions. Three different water flows (5 L/min, 10 L/min, and 15 L/min) were analyzed against conventional air-cooled induction motors and verified using available published resources. The analysis shows that for different flow rates (5 L/min, 10 L/min and 15 L/min respectively) we obtained corresponding temperature reductions of 2.94%, 4.79% and 7.69%. Therefore, the results show that the embedded induction motor can effectively reduce the temperature compared to the air-cooled induction motor.
The electric motor is one of the key inventions of modern engineering science. Electric motors are used in everything from household appliances to vehicles, including the automotive and aerospace industries. In recent years, the popularity of induction motors (AM) has increased due to their high starting torque, good speed control and moderate overload capacity (Fig. 1). Induction motors not only make your light bulbs glow, they power most of the gadgets in your home, from your toothbrush to your Tesla. Mechanical energy in IM is created by the contact of the magnetic field of the stator and rotor windings. In addition, IM is a viable option due to the limited supply of rare earth metals. However, the main disadvantage of ADs is that their lifetime and efficiency are very sensitive to temperature. Induction motors consume about 40% of the world’s electricity, which should lead us to think that managing the power consumption of these machines is critical.
The Arrhenius equation states that for every 10°C rise in operating temperature, the life of the entire engine is halved. Therefore, to ensure the reliability and increase the productivity of the machine, it is necessary to pay attention to the thermal control of blood pressure. In the past, thermal analysis has been neglected and motor designers have considered the problem only at the periphery, based on design experience or other dimensional variables such as winding current density, etc. These approaches lead to the application of large safety margins for worst-case heating conditions, resulting in an increase in machine size and therefore an increase in cost.
There are two types of thermal analysis: lumped circuit analysis and numerical methods. The main advantage of analytical methods is the ability to perform calculations quickly and accurately. However, considerable effort must be made to define circuits with sufficient accuracy to simulate thermal paths. On the other hand, numerical methods are roughly divided into computational fluid dynamics (CFD) and structural thermal analysis (STA), both of which use finite element analysis (FEA). The advantage of numerical analysis is that it allows you to model the geometry of the device. However, system setup and calculations can sometimes be difficult. The scientific articles discussed below are selected examples of thermal and electromagnetic analysis of various modern induction motors. These articles prompted the authors to study thermal phenomena in asynchronous motors and methods for their cooling.
Pil-Wan Han1 was engaged in thermal and electromagnetic analysis of MI. The lumped circuit analysis method is used for thermal analysis, and the time-varying magnetic finite element method is used for electromagnetic analysis. In order to properly provide thermal overload protection in any industrial application, the temperature of the stator winding must be reliably estimated. Ahmed et al.2 proposed a higher order heat network model based on deep thermal and thermodynamic considerations. The development of thermal modeling methods for industrial thermal protection purposes benefits from analytical solutions and consideration of thermal parameters.
Nair et al.3 used a combined analysis of a 39 kW IM and a 3D numerical thermal analysis to predict the thermal distribution in an electrical machine. Ying et al.4 analyzed fan-cooled fully enclosed (TEFC) IMs with 3D temperature estimation. Moon et al. 5 studied the heat flow properties of IM TEFC using CFD. The LPTN motor transition model was given by Todd et al.6. Experimental temperature data are used along with calculated temperatures derived from the proposed LPTN model. Peter et al.7 used CFD to study the air flow that affects the thermal behavior of electric motors.
Cabral et al8 proposed a simple IM thermal model in which the machine temperature was obtained by applying the cylinder heat diffusion equation. Nategh et al.9 studied a self-ventilated traction motor system using CFD to test the accuracy of optimized components. Thus, numerical and experimental studies can be used to simulate the thermal analysis of induction motors, see fig. 2.
Yinye et al.10 proposed a design to improve thermal management by exploiting the common thermal properties of standard materials and common sources of machine part loss. Marco et al.11 presented criteria for designing cooling systems and water jackets for machine components using CFD and LPTN models. Yaohui et al.12 provide various guidelines for selecting an appropriate cooling method and evaluating performance early in the design process. Nell et al.13 proposed to use models for coupled electromagnetic-thermal simulation for a given range of values, level of detail and computational power for a multiphysics problem. Jean et al.14 and Kim et al.15 studied the temperature distribution of an air-cooled induction motor using a 3D coupled FEM field. Calculate input data using 3D eddy current field analysis to find Joule losses and use them for thermal analysis.
Michel et al.16 compared conventional centrifugal cooling fans with axial fans of various designs through simulations and experiments. One of these designs achieved small but significant improvements in engine efficiency while maintaining the same operating temperature.
Lu et al.17 used the equivalent magnetic circuit method in combination with the Boglietti model to estimate iron losses on the shaft of an induction motor. The authors assume that the distribution of magnetic flux density in any cross section inside the spindle motor is uniform. They compared their method with the results of finite element analysis and experimental models. This method can be used for express analysis of MI, but its accuracy is limited.
18 presents various methods for analyzing the electromagnetic field of linear induction motors. Among them, methods for estimating power losses in reactive rails and methods for predicting the temperature rise of traction linear induction motors are described. These methods can be used to improve the energy conversion efficiency of linear induction motors.
Zabdur et al. 19 investigated the performance of cooling jackets using a three-dimensional numerical method. The cooling jacket uses water as the main source of coolant for the three-phase IM, which is important for the power and maximum temperatures required for pumping. Rippel et al. 20 have patented a new approach to liquid cooling systems called transverse laminated cooling, in which the refrigerant flows transversely through narrow regions formed by holes in each other magnetic lamination. Deriszade et al. 21 experimentally investigated the cooling of traction motors in the automotive industry using a mixture of ethylene glycol and water. Evaluate the performance of various mixtures with CFD and 3D turbulent fluid analysis. A simulation study by Boopathi et al.22 showed that the temperature range for water-cooled engines (17-124°C) is significantly smaller than for air-cooled engines (104-250°C). The maximum temperature of the aluminum water-cooled motor is reduced by 50.4%, and the maximum temperature of the PA6GF30 water-cooled motor is reduced by 48.4%. Bezukov et al.23 evaluated the effect of scale formation on the thermal conductivity of the engine wall with a liquid cooling system. Studies have shown that a 1.5 mm thick oxide film reduces heat transfer by 30%, increases fuel consumption and reduces engine power.
Tanguy et al.24 conducted experiments with various flow rates, oil temperatures, rotational speeds and injection modes for electric motors using lubricating oil as a coolant. A strong relationship has been established between flow rate and overall cooling efficiency. Ha et al.25 suggested using drip nozzles as nozzles to evenly distribute the oil film and maximize engine cooling efficiency.
Nandi et al.26 analyzed the effect of L-shaped flat heat pipes on engine performance and thermal management. The heat pipe evaporator part is installed in the motor casing or buried in the motor shaft, and the condenser part is installed and cooled by circulating liquid or air. Bellettre et al. 27 studied a PCM solid-liquid cooling system for a transient motor stator. The PCM impregnates the winding heads, lowering the hot spot temperature by storing latent thermal energy.
Thus, motor performance and temperature are evaluated using different cooling strategies, see fig. 3. These cooling circuits are designed to control the temperature of windings, plates, winding heads, magnets, carcass and end plates.
Liquid cooling systems are known for their efficient heat transfer. However, pumping coolant around the engine consumes a lot of energy, which reduces the engine’s effective power output. Air cooling systems, on the other hand, are a widely used method due to their low cost and ease of upgrade. However, it is still less efficient than liquid cooling systems. An integrated approach is needed that can combine the high heat transfer performance of a liquid-cooled system with the low cost of an air-cooled system without consuming additional energy.
This article lists and analyzes heat losses in AD. The mechanism of this problem, as well as the heating and cooling of induction motors, is explained in the Heat Loss in Induction Motors section through Cooling Strategies. The heat loss of the core of an induction motor is converted into heat. Therefore, this article discusses the mechanism of heat transfer inside the engine by conduction and forced convection. Thermal modeling of IM using continuity equations, Navier-Stokes/momentum equations and energy equations is reported. The researchers performed analytical and numerical thermal studies of IM to estimate the temperature of the stator windings for the sole purpose of controlling the thermal regime of the electric motor. This article focuses on thermal analysis of air-cooled IMs and thermal analysis of integrated air-cooled and water-cooled IMs using CAD modeling and ANSYS Fluent simulation. And the thermal advantages of the integrated improved model of air-cooled and water-cooled systems are deeply analyzed. As mentioned above, the documents listed here are not a summary of the state of the art in the field of thermal phenomena and cooling of induction motors, but they indicate many problems that need to be solved in order to ensure the reliable operation of induction motors.
Heat loss is usually divided into copper loss, iron loss and friction/mechanical loss.
Copper losses are the result of Joule heating due to the resistivity of the conductor and can be quantified as 10.28:
where q̇g is the heat generated, I and Ve are the nominal current and voltage, respectively, and Re is the copper resistance.
Iron loss, also known as parasitic loss, is the second main type of loss that causes hysteresis and eddy current losses in AM, mainly caused by the time-varying magnetic field. They are quantified by the extended Steinmetz equation, whose coefficients can be considered constant or variable depending on operating conditions10,28,29.
where Khn is the hysteresis loss factor derived from the core loss diagram, Ken is the eddy current loss factor, N is the harmonic index, Bn and f are the peak flux density and frequency of the non-sinusoidal excitation, respectively. The above equation can be further simplified as follows10,29:
Among them, K1 and K2 are the core loss factor and eddy current loss (qec), hysteresis loss (qh), and excess loss (qex), respectively.
Wind load and friction losses are the two main causes of mechanical losses in IM. Wind and friction losses are 10,
In the formula, n is the rotational speed, Kfb is the coefficient of friction losses, D is the outer diameter of the rotor, l is the length of the rotor, G is the weight of the rotor 10.
The primary mechanism for heat transfer within the engine is via conduction and internal heating, as determined by the Poisson equation30 applied to this example:
During operation, after a certain point in time when the motor reaches steady state, the heat generated can be approximated by a constant heating of the surface heat flux. Therefore, it can be assumed that the conduction inside the engine is carried out with the release of internal heat.
The heat transfer between the fins and the surrounding atmosphere is considered forced convection, when the fluid is forced to move in a certain direction by an external force. Convection can be expressed as 30:
where h is the heat transfer coefficient (W/m2 K), A is the surface area, and ΔT is the temperature difference between the heat transfer surface and the refrigerant perpendicular to the surface. The Nusselt number (Nu) is a measure of the ratio of convective and conductive heat transfer perpendicular to the boundary and is chosen based on the characteristics of laminar and turbulent flow. According to the empirical method, the Nusselt number of turbulent flow is usually associated with the Reynolds number and the Prandtl number, expressed as 30:
where h is the convective heat transfer coefficient (W/m2 K), l is the characteristic length, λ is the thermal conductivity of the fluid (W/m K), and the Prandtl number (Pr) is a measure of the ratio of the momentum diffusion coefficient to the thermal diffusivity (or velocity and relative thickness of the thermal boundary layer), defined as 30:
where k and cp are the thermal conductivity and specific heat capacity of the liquid, respectively. In general, air and water are the most common coolants for electric motors. The liquid properties of air and water at ambient temperature are shown in Table 1.
IM thermal modeling is based on the following assumptions: 3D steady state, turbulent flow, air is an ideal gas, negligible radiation, Newtonian fluid, incompressible fluid, no-slip condition, and constant properties. Therefore, the following equations are used to fulfill the laws of conservation of mass, momentum, and energy in the liquid region.
In the general case, the mass conservation equation is equal to the net mass flow into the cell with liquid, determined by the formula:
According to Newton’s second law, the rate of change of the momentum of a liquid particle is equal to the sum of the forces acting on it, and the general momentum conservation equation can be written in vector form as:
The terms ∇p, ∇∙τij, and ρg in the above equation represent pressure, viscosity, and gravity, respectively. Cooling media (air, water, oil, etc.) used as coolants in machines are generally considered to be Newtonian. The equations shown here only include a linear relationship between shear stress and a velocity gradient (strain rate) perpendicular to the shear direction. Considering constant viscosity and steady flow, equation (12) can be changed to 31:
According to the first law of thermodynamics, the rate of change in the energy of a liquid particle is equal to the sum of the net heat generated by the liquid particle and the net power produced by the liquid particle. For a Newtonian compressible viscous flow, the energy conservation equation can be expressed as31:
where Cp is the heat capacity at constant pressure, and the term ∇ ∙ (k∇T) is related to the thermal conductivity through the liquid cell boundary, where k denotes the thermal conductivity. The conversion of mechanical energy into heat is considered in terms of \(\varnothing\) (i.e., the viscous dissipation function) and is defined as:
Where \(\rho\) is the density of the liquid, \(\mu\) is the viscosity of the liquid, u, v and w are the potential of the direction x, y, z of the liquid velocity, respectively. This term describes the conversion of mechanical energy into thermal energy and can be ignored because it is only important when the viscosity of the fluid is very high and the velocity gradient of the fluid is very large. In the case of steady flow, constant specific heat and thermal conductivity, the energy equation is modified as follows:
These basic equations are solved for laminar flow in the Cartesian coordinate system. However, like many other technical problems, the operation of electrical machines is primarily associated with turbulent flows. Therefore, these equations are modified to form the Reynolds Navier-Stokes (RANS) averaging method for turbulence modeling.
In this work, the ANSYS FLUENT 2021 program for CFD modeling with the corresponding boundary conditions was chosen, such as the model considered: an asynchronous engine with an air cooling with a capacity of 100 kW, the diameter of the rotor 80.80 mm, the diameter of the stator 83.56 mm (internal) and 190 mm (external), an air gap of 1.38 mm, the total length of 234 mm, the amount , the thickness of the ribs 3 mm. .
The SolidWorks air-cooled engine model is then imported into ANSYS Fluent and simulated. In addition, the results obtained are checked to ensure the accuracy of the simulation performed. In addition, an integrated air- and water-cooled IM was modeled using SolidWorks 2017 software and simulated using ANSYS Fluent 2021 software (Figure 4).
The design and dimensions of this model are inspired by the Siemens 1LA9 aluminum series and modeled in SolidWorks 2017. The model has been slightly modified to suit the needs of the simulation software. Modify CAD models by removing unwanted parts, removing fillets, chamfers, and more when modeling with ANSYS Workbench 2021.
A design innovation is the water jacket, the length of which was determined from the simulation results of the first model. Some changes have been made to the water jacket simulation to get the best results when using the waist in ANSYS. Various parts of the IM are shown in fig. 5a–f.
(A). Rotor core and IM shaft. (b) IM stator core. (c) IM stator winding. (d) External frame of the MI. (e) IM water jacket. f) combination of air and water cooled IM models.
The shaft-mounted fan provides a constant air flow of 10 m/s and a temperature of 30 °C on the surface of the fins. The value of the rate is chosen randomly depending on the capacity of the blood pressure analyzed in this article, which is greater than that indicated in the literature. The hot zone includes the rotor, stator, stator windings and rotor cage bars. The materials of the stator and rotor are steel, the windings and cage rods are copper, the frame and ribs are aluminum. The heat generated in these areas is due to electromagnetic phenomena, such as Joule heating when an external current is passed through a copper coil, as well as changes in the magnetic field. The heat release rates of the various components were taken from various literature available for a 100 kW IM.
Integrated air-cooled and water-cooled IMs, in addition to the above conditions, also included a water jacket, in which the heat transfer capabilities and pump power requirements were analyzed for various water flow rates (5 l/min, 10 l/min and 15 l/min). This valve was chosen as the minimum valve, since the results did not change significantly for flows below 5 L/min. In addition, a flow rate of 15 L/min was chosen as the maximum value, since the pumping power increased significantly despite the fact that the temperature continued to fall.
Various IM models were imported into ANSYS Fluent and further edited using ANSYS Design Modeler. Further, a box-shaped casing with dimensions of 0.3 × 0.3 × 0.5 m was built around the AD to analyze the movement of air around the engine and study the removal of heat into the atmosphere. Similar analyzes were performed for integrated air- and water-cooled IMs.
The IM model is modeled using CFD and FEM numerical methods. Meshes are built in CFD to divide a domain into a certain number of components in order to find a solution. Tetrahedral meshes with appropriate element sizes are used for general complex geometry of engine components. All interfaces were filled with 10 layers to obtain accurate surface heat transfer results. The grid geometry of two MI models is shown in Fig. . 6a, b.
The energy equation allows you to study heat transfer in various areas of the engine. The K-epsilon turbulence model with standard wall functions was chosen to model turbulence around the outer surface. The model takes into account kinetic energy (Ek) and turbulent dissipation (epsilon). Copper, aluminium, steel, air and water were selected for their standard properties for use in their respective applications. Heat dissipation rates (see Table 2) are given as inputs, and different battery zone conditions are set to 15, 17, 28, 32. The air speed over the motor case was set to 10 m/s for both motor models, and in addition, three different water rates were taken into account for the water jacket (5 l/min, 10 l/min and 15 l/min). For greater accuracy, the residuals for all equations were set equal to 1 × 10–6. Select the SIMPLE (Semi-Implicit Method for Pressure Equations) algorithm to solve the Navier Prime (NS) equations. After hybrid initialization is complete, the setup will run 500 iterations, as shown in Figure 7.


Post time: Jul-24-2023