Multi-target parameter optimization method for disc type permanent magnet motor of oil pumping unit

Abstract

The invention discloses a pumping unit disc type permanent magnet motor multi-objective parameter optimization method, which relates to the technical field of motor design and optimization, and comprises the following steps: pre-establishing a multi-physical field coupling simulation model, constructing a disc type permanent magnet motor electromagnetic and thermal performance simulation model based on Motor-CAD and Maxwell software, and completing no-load, load and thermal performance tests of the motor. According to the method, the multi-physical field coupling simulation model is established in advance, so that comprehensive coupling analysis of electromagnetic and thermal performance of the disc type permanent magnet motor is realized, and high-precision data support is provided for multi-objective optimization; key structure parameters such as air gap length, slot opening width and the like are scientifically selected as optimization objects, maximization of motor efficiency and minimization of cogging torque are taken as core targets, and meanwhile, constraint requirements of output torque, back electromotive force and structure size are strictly adherent to each other, so that the pain point that multiple performances are difficult to consider at the same time in traditional optimization is effectively solved.

Description

Technical Field

[0001] This invention relates to the field of motor design and optimization technology, and more specifically, to a multi-objective parameter optimization method for a disc-type permanent magnet motor for oil pumping units. Background Technology

[0002] Traditional oil pumping units employ a multi-stage transmission system consisting of a motor, belt, gearbox, and four-bar linkage, resulting in high energy loss, high equipment failure rate, and high maintenance costs. Actual measurement data shows that belt drive efficiency is only around 85%, gearbox efficiency is approximately 92%, and four-bar linkage efficiency is 94%, leading to low overall energy efficiency due to the multi-stage transmission. Furthermore, traditional oil pumping units use excessively high-power three-phase asynchronous motors, which operate under light loads in most conditions, creating a "large horse pulling a small cart" phenomenon, further reducing motor efficiency and the power factor of the power grid. To address these issues, the industry has adopted permanent magnet synchronous motors combined with semi-direct drive technology, simplifying the transmission structure, reducing mechanical losses and maintenance costs, and has already been widely applied in several oilfields, including Daqing Oilfield and Shengli Oilfield.

[0003] However, there is still room for optimization in existing disc-type permanent magnet motors: the efficiency of the motor itself needs to be improved in some designs, and the excessive cogging torque affects the smoothness of operation; traditional optimization methods based on finite element simulation require a large number of iterative calculations, resulting in long optimization cycles and high computational costs. Existing multi-objective optimization algorithms such as NSGA-II are computationally inefficient when directly combined with finite element simulation; and a single optimization objective is difficult to simultaneously meet multiple performance requirements such as motor efficiency and smooth operation.

[0004] Therefore, there is an urgent need for an efficient and accurate multi-objective parameter optimization method to achieve synergistic optimization of maximizing motor efficiency and minimizing cogging torque, while meeting the structural constraints and operational requirements of the pumping unit. Summary of the Invention

[0005] To address the problems in related technologies, this invention proposes a multi-objective parameter optimization method for disc-type permanent magnet motors in oil pumping units. This method aims to solve the problems of low optimization efficiency and difficulty in simultaneously achieving multiple performance indicators in existing permanent magnet motor optimization for oil pumping units, thereby realizing synergistic optimization of motor efficiency improvement and cogging torque suppression, while ensuring the high efficiency and engineering feasibility of the optimization process.

[0006] The technical solution of this invention is implemented as follows: A method for optimizing multi-objective parameters of a disc-type permanent magnet motor in an oil pumping unit includes the following steps: A multiphysics coupling simulation model was pre-established, and an electromagnetic and thermal performance simulation model of a disc permanent magnet motor was constructed based on Motor-CAD and Maxwell software. The motor's no-load, load, and thermal performance tests were completed. The optimization parameters and target constraints are determined. The air gap length, slot opening width, slot depth, tooth width, permanent magnet width and pole arc coefficient are selected as optimization parameters. The optimization objectives are to maximize motor efficiency and minimize cogging torque. The output torque, back electromotive force and structural dimensions are the constraints. A Kriging surrogate model was constructed, and a sample set was generated using Latin hypercube sampling. Sample performance data was obtained through simulation, and the Kriging surrogate model was trained and validated. Multi-objective optimization is performed, based on the NSGA-II algorithm and the Kriging surrogate model, to generate the Pareto optimal frontier and select the optimal parameter combination; The optimization scheme is verified by using finite element simulation to verify whether the performance indicators of the optimal scheme meet the constraints.

[0007] Furthermore, the construction of the electromagnetic and thermal performance simulation model of the disc-type permanent magnet motor includes: setting up a permanent magnet synchronous motor topology in Motor-CAD that is adapted to the high torque and low speed output of the pumping unit; inputting the stator and rotor structural geometry; selecting the core and permanent magnet materials; and setting the number of winding turns and branches to complete the motor model definition; by setting the rated current and rated speed parameters, performing static, no-load, and load performance analyses in sequence; using the two-dimensional finite element method to calculate the output torque, power, efficiency, and various losses and importing them into Maxwell for accurate electromagnetic simulation; simultaneously defining the motor cooling environment and heat dissipation method; performing steady-state thermal calculations and transient thermal model analyses; and obtaining the temperature and temperature rise data of each key component.

[0008] Furthermore, the Maxwell performs accurate electromagnetic simulations, including: simulating magnetic field and heat distribution using a combination of analytical models and finite element analysis; describing two-dimensional static or quasi-static magnetic field problems using equations related to magnetic vector potential, permeability, and current density; calculating electromagnetic torque using the Maxwell stress tensor method based on radial and tangential magnetic flux density components, radius, and vacuum permeability at the air gap; and performing Motor-CAD thermal simulation based on steady-state or transient heat conduction equations, establishing convective heat transfer boundary conditions through empirical fluid dynamics formulas, calculating one-dimensional steady-state heat conduction thermal resistance according to Fourier's law, and calculating convective heat transfer thermal resistance through convective heat transfer coefficient and heat transfer area.

[0009] Furthermore, the range of values ​​for the optimized parameters includes: air gap length 0.8-1.6mm, slot opening width 5-15mm, slot depth 40-50mm, tooth width 12-20mm, permanent magnet width 7-15mm, and pole arc coefficient 0.67-0.87; and the constraints are that the output torque is not lower than the rated value of 2208Nm, the effective value of the back electromotive force does not exceed 340V, the radial profile of the motor meets 1200mm×220mm, and the total axial length does not exceed 1060mm.

[0010] Furthermore, the training and validation of the Kriging agent model includes the following steps: To optimize the range of motor size parameters, a dataset was built using Latin hypercube sampling. Samples were generated by stratified sampling of each variable and random combination. Based on the definition of a Gaussian static stochastic process using the Kriging model, the response value of unknown points is predicted by linearly weighting and summing the response values ​​of known sample points, and the weighting coefficients are solved to construct the model. The training and test sets were divided in an 8:2 ratio, and the correlation coefficient R0 was used. 2 Evaluate model accuracy and ensure R 2 A value of ≥0.95 is used to achieve accurate fitting of the complex nonlinear performance relationship of the motor.

[0011] Furthermore, the dataset constructed using Latin hypercube sampling includes: the i-th variable, whose value range is... The k-th sample is represented as: ; in, It is the randomized interval index. is a random number within the interval, and N is the number of interval divisions.

[0012] Furthermore, the correlation coefficient R is used. 2 Evaluating model accuracy includes: calibrating R... 2 The closer the merit value is to 1, the higher the accuracy of the surrogate model. 2 Represented as: ; in, It is a finite element simulation calculation; These are the predicted values ​​from the proxy model; This is the average value.

[0013] Furthermore, the multi-objective optimization includes the following steps: Initialize a size of parental population And generate a size of through crossover and mutation operators. offspring population ; Merging parent populations With offspring population , forming a scale of 2 N temporary population ; For the merged population Perform a quick non-dominated sort to divide the system into different non-dominated levels. Next, calculate each level. The crowding distance between individuals is determined, and the optimal distance is selected based on the non-dominance level and crowding degree. Individuals form a new parental population. ; Repeat the above iterative process until the preset termination condition is met.

[0014] The beneficial effects of this invention are: This invention utilizes Motor-CAD and Maxwell software simulation modeling to achieve a comprehensive coupled analysis of the electromagnetic and thermal performance of a disc permanent magnet motor, providing high-precision data support for multi-objective optimization. By scientifically selecting key structural parameters such as air gap length and slot opening width as optimization objects, and focusing on maximizing motor efficiency and minimizing cogging torque as core objectives, while strictly adhering to constraints on output torque and back electromotive force, it effectively solves the pain point of traditional optimization methods where a single objective cannot simultaneously consider multiple performance aspects. By using Latin hypercube sampling to create a sample set and establishing a surrogate model through a Kriging surrogate model, the number of iterations in finite element simulation is significantly reduced, significantly improving optimization computational efficiency and overcoming the shortcomings of traditional optimization methods, such as long cycles and high costs. Based on NSG... The A-II algorithm generates the Pareto optimal front, which can quickly screen out the parameter combination that balances efficiency and cogging torque. Simulation verification shows that after optimization, the motor efficiency is improved to over 94.6%, the cogging torque is suppressed to below 17.86 Nm, the output torque and back EMF both meet the rated operating conditions, and the temperature of each key component of the motor is within the safe operating range. At the same time, the optimized motor is adapted to the semi-direct drive transmission scheme of the pumping unit. It has a compact structure and strong adaptability. It not only simplifies the multi-stage transmission structure of the traditional pumping unit, reduces mechanical losses and maintenance costs, but also effectively improves the "oversized motor for a small load" phenomenon and improves the power factor of the power grid. In actual oilfield applications, it can achieve the dual benefits of significantly reduced energy consumption and greatly improved operational stability, and has outstanding engineering practicality and economic value. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart illustrating a multi-objective parameter optimization method for a disc-type permanent magnet motor in an oil pumping unit according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the semi-direct drive permanent magnet motor drive mode of a pumping unit according to an embodiment of the present invention, which is a method for optimizing multi-objective parameters of a disc-type permanent magnet motor in a pumping unit. Figure 3 This is a schematic diagram of the motor structure of a multi-objective parameter optimization method for a disc-type permanent magnet motor of an oil pumping unit according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the winding connection of a multi-objective parameter optimization method for a disc-type permanent magnet motor of an oil pumping unit according to an embodiment of the present invention. Figure 5 This is a schematic diagram of the no-load magnetic flux density distribution of a multi-objective parameter optimization method for a disc-type permanent magnet motor of an oil pumping unit according to an embodiment of the present invention. Figure 6 This is a schematic diagram of the no-load motor performance of a multi-objective parameter optimization method for a disc-type permanent magnet motor of an oil pumping unit according to an embodiment of the present invention. Figure 7 This is a schematic diagram of the load magnetic flux distribution of a multi-objective parameter optimization method for a disc-type permanent magnet motor of an oil pumping unit according to an embodiment of the present invention. Figure 8 This is a schematic diagram of the load motor performance of a multi-objective parameter optimization method for a disc-type permanent magnet motor of an oil pumping unit according to an embodiment of the present invention. Figure 9 This is a schematic diagram of the rated load temperature of a disc-type permanent magnet motor for an oil pumping unit according to an embodiment of the present invention. Figure 10 This is a schematic diagram of the rated load temperature rise of a disc-type permanent magnet motor for an oil pumping unit according to an embodiment of the present invention. Figure 11 This is a schematic diagram of the correlation coefficients between the optimization objectives and constraints of a multi-objective parameter optimization method for a disc-type permanent magnet motor of an oil pumping unit according to an embodiment of the present invention. Figure 12 This is a schematic diagram of the sensitivity analysis of the optimization parameters of a multi-objective parameter optimization method for a disc-type permanent magnet motor of an oil pumping unit according to an embodiment of the present invention; Figure 13 This is a schematic diagram illustrating the optimization objective iteration trend of a multi-objective parameter optimization method for a disc-type permanent magnet motor of an oil pumping unit according to an embodiment of the present invention. Figure 14 This is a Pareto front plot of a multi-objective parameter optimization method for a disc-type permanent magnet motor of an oil pumping unit according to an embodiment of the present invention; Figure 15 This is a comparison of simulation results of a multi-objective parameter optimization method for a disc-type permanent magnet motor in an oil pumping unit according to an embodiment of the present invention; Figure 16 This is a MAP diagram of motor efficiency for a disc-type permanent magnet motor in an oil pumping unit, according to an embodiment of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention are within the scope of protection of the present invention.

[0018] According to an embodiment of the present invention, a method for optimizing multi-objective parameters of a disc-type permanent magnet motor for an oil pumping unit is provided.

[0019] like Figure 1 As shown, the multi-objective parameter optimization method for a disc-type permanent magnet motor in an oil pumping unit according to an embodiment of the present invention includes the following steps: Step S1: First, establish a multi-physics coupling simulation model in advance. Then, construct an electromagnetic and thermal performance simulation model of the disc permanent magnet motor based on Motor-CAD and Maxwell software, and complete the no-load, load and thermal performance tests of the motor. This technical solution involves setting up a permanent magnet synchronous motor topology in Motor-CAD suitable for the high torque and low speed output of an oil pumping unit; inputting the geometric dimensions of the stator, rotor, and other motor structures, and selecting the core and permanent magnet materials; and defining a motor model that meets the load requirements of the oil pumping unit by setting parameters such as the number of winding turns and the number of branches.

[0020] Once the model is established, electromagnetic performance calculations and evaluations can be performed. Motor performance is assessed by setting parameters such as rated current and rated speed. The first step is a no-load test to calculate key inherent characteristics, such as cogging torque and back EMF waveforms. Next is load performance analysis, using the two-dimensional finite element method to calculate the motor's output torque, power, efficiency, and various losses, including copper losses, iron losses, and permanent magnet eddy current losses. These results are then imported into Maxwell for more accurate electromagnetic simulations.

[0021] Define the motor's cooling environment and heat dissipation method within Motor-CAD, selecting modes such as natural air cooling or water jacket cooling, and setting the corresponding environmental parameters. After setting, perform steady-state thermal calculations. By solving the lumped-parameter thermal network, calculate the final temperature of each key component inside the motor after continuous operation and reaching thermal equilibrium. Next, establish a transient thermal model, set the duration, and calculate the temperature rise of each component of the motor.

[0022] Maxwell's simulation principle is based on the theory of electromagnetism. By combining analytical models with finite element analysis (FEA), it can realize the rapid simulation of the thermal and magnetic field distribution of motors under different operating conditions.

[0023] For two-dimensional static or quasi-static magnetic field problems, the expression is as follows: ; in, Represents the magnetic vector potential at z Components in the axial direction, The magnetic permeability represents the magnetic conductivity of a material and reflects its ability to conduct magnetic energy. This represents the current density along the z-axis.

[0024] In calculating electromagnetic torque, Maxwell typically employs the Maxwell stress tensor method, which uses the distribution of the magnetic field on the air gap surface to determine the average electromagnetic torque, expressed as:

[0025] in, and These are the radial and tangential magnetic flux density components at the air gap, respectively. For radius, The value is the vacuum permeability. This method can reflect detailed characteristics such as cogging effect and torque fluctuation.

[0026] The Motor-CAD thermal simulation part is based on steady-state or transient heat conduction equations, expressed as follows: ; in, For density, For specific heat capacity, Thermal conductivity, This refers to the heat source power density per unit volume. The heat sources in an electric motor mainly come from copper losses, iron losses, mechanical losses, and additional losses.

[0027] For fluid cooling, Motor-CAD establishes convective heat transfer boundary conditions using empirical fluid dynamics formulas, expressed as: ; in, Where A is the convective heat transfer coefficient and A is the heat transfer area. For surface temperature, This refers to the fluid temperature. By coupling the cooling channel geometry with the fluid flow rate, it can be dynamically adjusted. The distribution of the data is used to evaluate the effectiveness of different cooling strategies.

[0028] Among them, the one-dimensional steady-state thermal conduction resistance, based on Fourier's law, is used to calculate the resistance encountered when heat is conducted in a single direction in the solid components of the motor, and is expressed as: ; in, It is the thermal resistance of conduction. It is the path length of heat conduction. It is the thermal conductivity of a material, representing the material's ability to conduct heat. It is the cross-sectional area perpendicular to the direction of heat flow.

[0029] The convective heat transfer thermal resistance equation is used to calculate the resistance encountered when heat is transferred from the solid surface of the motor to the surrounding fluid, and is expressed as: ; in, It is convective thermal resistance. It is the convective heat transfer coefficient. It refers to the surface area of ​​the solid undergoing convective heat transfer. The main losses in a motor are due to core losses. Copper loss Eddy current loss Mechanical wear It consists of several parts; mechanical losses are composed of bearing friction and wind resistance losses; motor efficiency... , is represented as , ; in, This refers to the output power.

[0030] Step S2: Determine the optimization parameters and target constraints. Select the air gap length, slot opening width, slot depth, tooth width, permanent magnet width, and pole arc coefficient as optimization parameters. Set the optimization objectives as maximizing motor efficiency and minimizing cogging torque. Set the output torque, back electromotive force, and structural dimensions as constraints. This technical solution selects air gap length, slot opening width, slot depth, tooth width, permanent magnet width, and pole arc coefficient as optimization parameters, and clarifies the value range of each parameter. The optimization objective is to maximize motor efficiency and minimize cogging torque; the constraints are that the output torque is not less than the rated value of 2208 N·m, and the effective value of the back electromotive force does not exceed 340 V.

[0031] Step S3: Construct the Kriging surrogate model. Generate a sample set using Latin hypercube sampling, obtain sample performance data through simulation, and train and validate the Kriging surrogate model. This includes the following steps: Step S301: Optimize the range of motor size parameters in advance and establish a dataset of motor geometric parameters. Use LHS (Latin Hypercube Sampling) to establish the dataset. By performing stratified sampling and random combination of each variable, Latin Hypercube Sampling can cover the entire parameter space uniformly with a small number of samples with extremely high efficiency.

[0032] The principle of Latin hypercube sampling (LHS) is to uniformly divide the range of values ​​for each variable into N intervals, randomly select a point in each interval, and then randomly pair intervals of different variables to form N sets of samples. This ensures that each variable is uniformly distributed across all samples, guaranteeing the homogeneity of the samples in the multidimensional space. For the i-th variable, the value interval is... The k-th sample is represented as: ; in, It is the randomized interval index. It is a random number within the interval.

[0033] Step S302: The Kriging model effectively handles the complex nonlinear relationships between motor performance parameters, while also considering the spatial correlation between samples, exhibiting high prediction accuracy and robustness. The Kriging model is an interpolation model that predicts the response value at any unknown point in the design space by linearly weighting and summing the response values ​​of known sample points. The model's predicted values... It can be calculated using the following formula: ; in, It is the predicted value of the unknown point. It is the response value of the i-th known sample point, and These are the corresponding weighting coefficients.

[0034] In order to solve for the weighting coefficients, the Kriging model uses the unknown function... Defined as a Gaussian static random process, its mathematical expression is: ; in, It is the mathematical expectation of the function, representing the global trend of the model. It is a function with a mean of 0 and a variance of . For a static stochastic process, the covariance of these random variables in the design space is expressed as: ; in, The number of design variables; Indicates the first The degree to which each variable contributes to the objective; Represents sample points and The distance between them.

[0035] To ensure that the model still has good predictive ability on unknown sample points and to effectively prevent the model from overfitting, it is usually necessary to divide the sample data into training and test sets in a reasonable way.

[0036] Step S303, using the correlation coefficient Evaluate the accuracy of the proxy model. The closer the merit value is to 1, the higher the accuracy of the surrogate model. Represented as ; in, It is a finite element simulation calculation; These are the predicted values ​​from the proxy model; This is the average value.

[0037] Using the above scheme, a Latin hypercube sampling (LHS) method was employed to generate 120 sample sets, covering the entire optimization parameter space. Performance data for each sample set was obtained through a multiphysics coupled simulation model. The training and test sets were divided in an 8:2 ratio. A Kriging surrogate model was trained based on the training set, and the correlation coefficient R0 was used to perform the simulation. 2 Evaluate model accuracy to ensure This enables accurate fitting of the complex nonlinear performance relationships of motors.

[0038] Step S4: Perform multi-objective optimization, generate the Pareto optimal front based on the NSGA-II algorithm and the Kriging surrogate model, and select the optimal parameter combination; This technical solution, based on a previously trained Kriging surrogate model, constructs a multi-objective optimization problem containing optimization objectives and constraints. It iteratively generates the Pareto front representing all optimal solution sets using the NSGA-II algorithm, including the following steps: Initialize a size of parental population And generate a size of through crossover and mutation operators. offspring population ; Merging parent populations With offspring population , forming a scale of 2 N temporary population ; For the merged population Perform a quick non-dominated sort to divide the system into different non-dominated levels. Next, calculate each level. The crowding distance between individuals is determined, and the optimal distance is selected based on the non-dominance level and crowding degree. Individuals form a new parental population. ; Repeat the above iterative process until the preset termination condition is met.

[0039] Using the above approach, a multi-objective optimization problem is constructed based on the trained Kriging surrogate model. The NSGA-II algorithm is used for global optimization, with an initial population size of 200, a crossover coefficient of 0.8, a mutation probability of 40%, and 200 iterations. Through fast non-dominated sorting and crowding calculation, a Pareto optimal front is generated. The optimal solution for balancing efficiency and cogging torque is selected from the front, ensuring that the optimized motor efficiency is ≥94.6% and the cogging torque is ≤17.86 Nm.

[0040] Step S5: Verify the optimization scheme by verifying whether the performance indicators of the optimal scheme meet the constraint requirements through finite element simulation.

[0041] Specifically, a finite element simulation model is established based on the optimal parameter combination, and electromagnetic and thermal performance simulations are conducted to confirm that the output torque, back electromotive force, and other indicators meet the constraint requirements, and that the temperature of each component of the motor is within the safe operating range.

[0042] Using the aforementioned technical solution, a motor reference model was established based on the design parameters of the pumping unit motor in the Kong-XX well of the Dagang Oilfield during implementation. The semi-direct drive transmission scheme adopted involves direct connection between the motor and the gearbox, integrated into the original pulley mounting location, such as... Figure 2 As shown in Table 1, due to the constraints of the flattened space, the motor shape needs to be adapted to a compact layout, with its radial profile within 1200mm × 220mm and its total axial length not exceeding 1060mm. The parameters of the permanent magnet motor for this well are shown in Table 1.

[0043] Table 1 Motor Parameters Parameters value Parameters value Rated Power / KW 37 Stator OD / mm 990 Rated voltage / V 380 Stator ID / mm 850 Rated current / A 63 Number of stator slots 72 Rated frequency / Hz 80 Poles 60 Rated Torque / Nm 2208 Bore diameter / mm 90 Efficiency / % 94 Shaft diameter / mm 215 Considering the periodic symmetry of the permanent magnet motor in the circumferential direction, and to balance computational accuracy and efficiency, this invention uses a 1 / 12 scale two-dimensional model of the motor for electromagnetic field analysis, such as... Figure 3 As shown.

[0044] In addition, to ensure stable operation and high efficiency of the motor, the stator adopts a parallel slotted tooth structure, and the stator and rotor cores are made of 50W470 grade non-oriented silicon steel sheets. The permanent magnets adopt a surface-mount topology and use N45UH grade neodymium iron boron (NdFeB) rare earth permanent magnet material, which has excellent anti-demagnetization ability. The winding section adopts a three-phase double-layer concentrated winding with a pitch of 1 and 54 turns per coil, with 6 parallel branches per phase. The most significant advantage of this structure is the extremely short winding ends, which can effectively reduce copper losses and shorten the axial dimension of the motor. The winding connection is as follows: Figure 4 As shown.

[0045] The simulation results are analyzed using the above scheme, as detailed below: 1) Perform electromagnetic analysis, as follows: Electromagnetic analysis of permanent magnet motors is divided into two main operating conditions: no-load and load. No-load analysis evaluates the inherent design characteristics of the motor, while load analysis evaluates its actual operating performance.

[0046] First, analyze the motor under no-load conditions. Figure 5 The diagram shows the magnetic flux density distribution under no-load conditions. The magnetic flux density amplitude in the stator teeth is 1.762T, the magnetic flux density amplitude in the air gap is 1.095T, and the magnetic flux density in the stator yoke is the highest, reaching 2.174T.

[0047] Cogging torque not only affects motor performance but also reduces the lifespan of equipment. Cogging torque, for example... Figure 6 As shown in (a), the motor's peak cogging torque reaches 57 Nm, which is relatively high, potentially indicating a weakness in low-speed operation smoothness. No-load back EMF is one of the important indicators of motor performance, such as... Figure 6 As shown in (b), its back EMF waveform exhibits good symmetry and a near-ideal sinusoidal distribution, achieving high stability. In summary, in subsequent optimization designs, effectively suppressing cogging torque while maintaining the advantage of back EMF will be of paramount importance.

[0048] This includes analyzing the motor load. Figure 7 The load magnetic flux density distribution diagram shows that the magnetic flux density amplitude in the stator teeth is 1.71T and the magnetic flux density amplitude in the air gap is 1.109T.

[0049] Under rated operating conditions, the average torque of the motor is as follows: Figure 8 As shown, the motor can provide an average torque of 2240Nm to 2390Nm, with an output torque of 2218Nm, which meets the rated torque.

[0050] Simulation yielded relevant motor losses and output power: core losses. 0.8365KW, copper loss 0.8425KW, eddy current loss 0.295KW, mechanical loss At 0.48KW, the system efficiency is 93.769%.

[0051] Compared to the field-measured motor data (efficiency approximately 94%, cogging torque approximately 20 Nm), the benchmark model's efficiency is lower than the field level, while the cogging torque is significantly higher. This indicates a significant deficiency in the motor's operational smoothness, suggesting substantial room for optimization. Therefore, it is necessary to optimize the motor's key parameters to reduce cogging torque and further improve efficiency.

[0052] 2) Temperature Analysis: Beam pumping units are deployed in outdoor oilfield environments for extended periods, subject to natural conditions such as diurnal temperature variations and wind erosion. The motors typically employ natural cooling, which is highly dependent on ambient wind speed and the convective heat transfer characteristics of the unit surface. This can easily lead to heat accumulation in critical components such as the stator windings, rotor core, and permanent magnets. The temperature diagram of the motor under rated operating conditions until it reaches thermal equilibrium is shown below. Figure 9 As shown, under long-term output conditions, the average stator winding temperature stabilizes at 116.3℃, and the magnet temperature reaches 121℃.

[0053] Transient thermal simulation of the motor yields the temperature distribution of various parts of the motor, as well as the temperature change trend over time, such as... Figure 10 As shown in the figure, the temperature rise of each component is observed during half an hour of operation at an ambient temperature of 25℃. The average temperature of the winding is 82.2℃, and the average temperature of the permanent magnet is 41.7℃.

[0054] The optimization results were analyzed as follows: 1) Determination of Optimization Objectives and Parameters: The optimization objectives were to maximize motor efficiency and minimize cogging torque, while constraining the output torque to be greater than 2208 Nm and the effective value of the back electromotive force to be less than 340 V, with sufficient safety thresholds. Key structural parameters were optimized, and their ranges are shown in Table 2. Sample data were generated using the Latin hypercube sampling (LHS) method, resulting in 120 sample points.

[0055] Table 2 Optimization Parameter Range Parameters Initial value Range of variation Air gap / mm 1.2 [0.8,1.6] Channel opening / mm 10 [5,15] Slot depth / mm 45 [40,50] Tooth width / mm 16 [12,20] PM width / mm 11 [7,15] Pole arc coefficient 0.75 [0.67,0.87] 2) Proxy Prediction Model Analysis: The simulation sample data was divided into training and test sets in an 8:2 ratio. For model performance evaluation, R... 2 As a primary evaluation metric, it is used to measure the accuracy of the fit between the model's predictions and the actual observed values. Figure 11 It can be seen that the fitting accuracy of the surrogate models for both the optimization objective and constraints is higher than 0.95. R is generally considered to be... 2 A value greater than 0.9 indicates that the surrogate model has high accuracy. Therefore, the surrogate model established in this invention can replace the original finite element simulation model and be applied to subsequent optimization design. Figure 12 The figure shows the sensitivity of each optimization variable to the optimization objective, reflecting the degree of influence of the optimization variable on the optimization objective. As can be seen from the figure, tooth width has a significant impact on torque and back electromotive force, while slot opening has a significant impact on efficiency and cogging torque. Therefore, when adjusting these highly sensitive parameters, it is important to select appropriate variable values ​​to improve the overall performance of the motor.

[0056] 3) Multi-objective parameter optimization: The NSGA-II algorithm is used to optimize the motor structure parameters. The initial population size is defined as 200, the crossover coefficient is 0.8, the mutation probability is 40%, and the iteration is 200 times. The graph shows the change of the optimization objective value with the number of iterations. Figure 13 As shown, the cogging torque exhibits a continuous and significant decreasing trend, eventually converging to approximately 20 N·m around the 150th generation. The efficiency exhibits a continuous and significant increasing trend, ultimately reaching approximately 94.6%. Furthermore, the optimization trend stabilizes after approximately the 150th generation, achieving the optimization target.

[0057] Additionally, Pareto optimal frontier plots, such as... Figure 14 As shown in the figure, pursuing higher efficiency requires accepting a significant increase in cogging torque; while pursuing the lowest possible cogging torque leads to a significant decrease in efficiency.

[0058] After comprehensive consideration, the optimal solution was selected with an efficiency of 94.6% and a cogging torque of 17.86 Nm. This solution maintains a high level of efficiency while keeping the cogging torque relatively low, achieving a good balance between the two optimization objectives. The optimization parameters are shown in Table 3.

[0059] Table 3 Parameter optimization results Parameters Initial value Optimized value Air gap / mm 1.2 1.3 Channel opening / mm 10 8.64 Slot depth / mm 45 45.44 Tooth width / mm 16 15.44 PM width / mm 11 9.96 Pole arc coefficient 0.75 0.7538 The simulation was performed using the optimized parameters, and the results were compared with those before optimization. The comparison is shown in the following figure. Figure 15 As shown. From Figure 15 As can be seen, the motor output torque waveform is smoother, and torque fluctuations are significantly reduced. The cogging torque waveform amplitude of 13.21 Nm is significantly reduced compared to before optimization, and the waveform is more regular. This indicates that the harmonic components of the optimized system are effectively suppressed, and the back electromotive force changes little, still exhibiting an ideal sine wave. The optimized motor efficiency is 94.642%, a significant improvement in efficiency and a significant reduction in energy consumption. The output torque is 2230.3 Nm, meeting the rated operating conditions. The effective value of the back electromotive force is 336.6 V, meeting the safety threshold. Simultaneously, the motor can maintain high efficiency while meeting the operating characteristics of a beam pumping unit, such as... Figure 16 As shown, the motor efficiency remains at a high level within the typical operating range of the pumping unit, with most operating points in the high-efficiency range of 91% to 94%, especially in the range of 100 to 150 rpm and 1500 to 2750 Nm, where the efficiency is close to the maximum value of 94.7%.

[0060] In summary, by utilizing the technical solutions described above, and through innovative integration of Motor-CAD and Maxwell software in collaborative simulation modeling, a comprehensive coupled analysis of the electromagnetic and thermal performance of a disc permanent magnet motor is achieved, providing high-precision data support for multi-objective optimization. By scientifically selecting key structural parameters such as air gap length and slot opening width as optimization objects, with the core objectives of maximizing motor efficiency and minimizing cogging torque, while strictly adhering to constraints on output torque, back electromotive force, and structural dimensions, the pain point of traditional optimization—the inability to simultaneously consider multiple performance aspects with a single objective—is effectively addressed. Furthermore, the organic combination of Latin hypercube sampling and the Kriging surrogate model significantly reduces the number of iterations in finite element simulation, substantially improving optimization computational efficiency and overcoming the long cycle and high cost of traditional optimization methods. The optimized motor overcomes the shortcomings of traditional pumping units. Based on the NSGA-II algorithm, a Pareto optimal front is generated, which can quickly select the optimal parameter combination for efficiency and cogging torque balance. Simulation verification shows that after optimization, the motor efficiency is increased to over 94.6%, the cogging torque is suppressed to below 17.86 Nm, and the output torque and back EMF both meet the rated operating conditions. Furthermore, the temperatures of all key motor components are within safe operating ranges. Simultaneously, the optimized motor is compatible with the semi-direct drive transmission scheme of pumping units, featuring a compact structure and strong adaptability. This not only simplifies the multi-stage transmission structure of traditional pumping units, reducing mechanical losses and maintenance costs, but also effectively improves the "oversized motor for a small load" phenomenon and enhances the power factor of the power grid. In practical oilfield applications, it can achieve the dual benefits of significantly reduced energy consumption and greatly improved operational stability, demonstrating outstanding engineering practicality and economic value.

[0061] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Those skilled in the art, upon considering the disclosure in the specification and embodiments, will readily conceive of other embodiments of this disclosure. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims.

[0062] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. A method for optimizing multi-objective parameters of a disc-type permanent magnet motor for an oil pumping unit, characterized in that, Includes the following steps: A multiphysics coupling simulation model was pre-established, and an electromagnetic and thermal performance simulation model of a disc permanent magnet motor was constructed based on Motor-CAD and Maxwell software. The motor's no-load, load, and thermal performance tests were completed. The optimization parameters and target constraints are determined. The air gap length, slot opening width, slot depth, tooth width, permanent magnet width and pole arc coefficient are selected as optimization parameters. The optimization objectives are to maximize motor efficiency and minimize cogging torque. The output torque, back electromotive force and structural dimensions are the constraints. A Kriging surrogate model was constructed, and a sample set was generated using Latin hypercube sampling. Sample performance data was obtained through simulation, and the Kriging surrogate model was trained and validated. Multi-objective optimization is performed, based on the NSGA-II algorithm and the Kriging surrogate model, to generate the Pareto optimal frontier and select the optimal parameter combination; The optimization scheme is verified by using finite element simulation to verify whether the performance indicators of the optimal scheme meet the constraints.

2. The multi-objective parameter optimization method for a disc-type permanent magnet motor in an oil pumping unit according to claim 1, characterized in that, The construction of the electromagnetic and thermal performance simulation model of the disc permanent magnet motor includes: setting up a permanent magnet synchronous motor topology in Motor-CAD to adapt to the high torque and low speed output of the pumping unit; inputting the stator and rotor structural geometry; selecting the core and permanent magnet materials; and setting the number of winding turns and branches to complete the motor model definition; by setting the rated current and rated speed parameters, performing static, no-load, and load performance analyses in sequence; using the two-dimensional finite element method to calculate the output torque, power, efficiency, and various losses, and importing the results into Maxwell for accurate electromagnetic simulation; simultaneously defining the motor cooling environment and heat dissipation method; performing steady-state thermal calculations and transient thermal model analyses; and obtaining the temperature and temperature rise data of each key component.

3. The multi-objective parameter optimization method for a disc-type permanent magnet motor in an oil pumping unit according to claim 2, characterized in that, The Maxwell simulation is performed to achieve accurate electromagnetic simulation, including: simulating magnetic field and heat distribution using a combination of analytical models and finite element analysis; describing two-dimensional static or quasi-static magnetic field problems using equations related to magnetic vector potential, permeability, and current density; calculating electromagnetic torque using the Maxwell stress tensor method based on radial and tangential magnetic flux density components, radius, and vacuum permeability at the air gap; and performing Motor-CAD thermal simulation based on steady-state or transient heat conduction equations, establishing convective heat transfer boundary conditions using empirical fluid dynamics formulas, calculating one-dimensional steady-state heat conduction thermal resistance based on Fourier's law, and calculating convective heat transfer thermal resistance using the convective heat transfer coefficient and heat transfer area.

4. The multi-objective parameter optimization method for a disc-type permanent magnet motor in an oil pumping unit according to claim 1, characterized in that, The range of values ​​for the optimized parameters includes: air gap length 0.8-1.6mm, slot opening width 5-15mm, slot depth 40-50mm, tooth width 12-20mm, permanent magnet width 7-15mm, and pole arc coefficient 0.67-0.87; and the constraints are that the output torque is not less than the rated value of 2208Nm, the effective value of the back electromotive force does not exceed 340V, the radial profile of the motor meets 1200mm×220mm, and the total axial length does not exceed 1060mm.

5. The multi-objective parameter optimization method for a disc-type permanent magnet motor in an oil pumping unit according to claim 1, characterized in that, The training and validation of the Kriging proxy model includes the following steps: To optimize the range of motor size parameters, a dataset was built using Latin hypercube sampling. Samples were generated by stratified sampling of each variable and random combination. Based on the definition of a Gaussian static stochastic process using the Kriging model, the response value of unknown points is predicted by linearly weighting and summing the response values ​​of known sample points, and the weighting coefficients are solved to construct the model. The training and test sets were divided in an 8:2 ratio, and the correlation coefficient R0 was used. 2 Evaluate model accuracy and ensure R 2 A value of ≥0.95 is used to achieve accurate fitting of the complex nonlinear performance relationship of the motor.

6. The multi-objective parameter optimization method for a disc-type permanent magnet motor in an oil pumping unit according to claim 5, characterized in that, The method of establishing the dataset using Latin hypercube sampling includes: labeling the i-th variable, with a value range of... The k-th sample is represented as: ； in, It is the randomized interval index. is a random number within the interval, and N is the number of interval divisions.

7. The multi-objective parameter optimization method for a disc-type permanent magnet motor in an oil pumping unit according to claim 6, characterized in that, The correlation coefficient R is used 2 Evaluating model accuracy includes: calibrating R... 2 The closer the merit value is to 1, the higher the accuracy of the surrogate model. 2 Represented as: ； in, It is a finite element simulation calculation; These are the predicted values ​​from the proxy model; This is the average value.

8. The multi-objective parameter optimization method for a disc-type permanent magnet motor in an oil pumping unit according to claim 1, characterized in that, The multi-objective optimization includes the following steps: Initialize a size of parental population And generate a size of through crossover and mutation operators. offspring population ; Merging parent populations With offspring population , forming a scale of 2 N temporary population ; For the merged population Perform a quick non-dominated sort to divide the system into different non-dominated levels. Next, calculate each level. The crowding distance between individuals is determined, and the optimal distance is selected based on the non-dominance level and crowding degree. Individuals form a new parental population. ; Repeat the above iterative process until the preset termination condition is met.

Images