Optimization method of coil parameters of eddy current sensor

Abstract

The present invention belongs to the field of metrology and calibration technology, and specifically discloses a method for optimizing coil parameters of an eddy current sensor, including: determining coil materials and basic size parameters; generating sample data between coil parameters and performance; constructing a BP neural network model, training and fitting the nonlinear relationship between coil parameters and measurement performance; optimizing the BP neural network model and coil parameters using a particle swarm algorithm; and designing and manufacturing an eddy current sensor based on the optimization results. The present invention uses a BP neural network to model the nonlinear mapping relationship between coil parameters and sensor performance, and combines the global optimization capability of a PSO algorithm to achieve intelligent optimization of coil parameters; the sensor coil designed through optimization has good performance and can achieve high-precision linear measurement within a range of 2 to 15 mm.

Description

Technical Field

[0001] The invention relates to the technical field of metrology and calibration, and in particular to a method for optimizing coil parameters of an eddy current sensor based on a particle swarm algorithm (PSO) and a BP neural network model. Background Art

[0002] Water-lubricated bearings are an important part of ship propulsion systems. Their lubrication characteristics, load-bearing performance, service life and dynamic characteristics have a significant impact on the safety and reliability of ship operation. However, when ships operate in complex marine environments, water-lubricated bearings are prone to wear, which in turn affects their load-bearing capacity and increases the risk of safety accidents. Therefore, real-time monitoring of water-lubricated bearing wear is particularly important so that timely warnings can be issued when bearings are abnormal to avoid potential dangers.

[0003] As an efficient monitoring method, eddy current sensors have the advantages of no need to change the bearing structure and adapt to complex environments compared with ultrasonic sensors and fiber optic sensors. They can meet the requirements of water-lubricated bearing wear measurement. The probe coil of the eddy current sensor is the core component for achieving high-performance measurement. The structure and size parameters of the coil directly affect the measurement accuracy and range of the sensor. Generally, the coil structure parameters include six main variables: number of layers, number of turns, inner diameter, line width, line height and spacing. Its design is crucial to the performance of the sensor. However, the current design of coil parameters mainly relies on experience and lacks a systematic optimization method. The design results have large uncertainties and are difficult to fully meet the needs of high-precision and large-range measurements.

[0004] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present invention, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention

[0005] The purpose of the present invention is to solve the technical problems existing in the background technology. To this end, a method for optimizing the coil parameters of an eddy current sensor based on a particle swarm algorithm (PSO) and a BP neural network model is provided. By constructing a nonlinear mapping relationship between sensor performance and coil parameters, an intelligent optimization algorithm is used to globally search for the optimal parameter combination, thereby significantly improving the coil design efficiency and sensor measurement performance. The eddy current sensor manufactured according to the optimization results can achieve a linear measurement range of up to 15 mm, meeting the actual needs of water-lubricated bearing wear monitoring.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0007] The eddy current sensor coil parameter optimization method based on the BP neural network model includes the following steps:

[0008] Step S1: first determine the coil material and basic size parameter range;

[0009] Step S2: Based on finite element simulation analysis, sample data between coil parameters and performance are generated (i.e., inductance, resistance and magnetic induction intensity data under different parameter combinations are generated);

[0010] Step S3: constructing a BP neural network model, training and fitting the nonlinear relationship between coil parameters and measurement performance;

[0011] Step S4: Optimizing the BP neural network model using a particle swarm algorithm to achieve an optimal combination of coil parameters;

[0012] Step S5: designing and manufacturing an eddy current sensor according to the optimization result.

[0013] The following is a technical solution further defined by the present invention, wherein the coil parameters include the number of turns, inner diameter, line width, line height and spacing.

[0014] The following is a technical solution further defined by the present invention, wherein the finite element simulation analysis is used to generate inductance, resistance and magnetic induction intensity data under different coil parameter combinations.

[0015] The following is a technical solution further defined by the present invention, wherein the BP neural network model includes an input layer, a hidden layer and an output layer, wherein: the input layer is used to receive coil parameters; the output layer outputs corresponding performance data (such as inductance, resistance, and magnetic induction intensity). A nonlinear mapping relationship between coil parameters and sensor performance is established, and the model is trained through training data so that it can accurately predict performance indicators.

[0016] The following is a technical solution further defined by the present invention. In step S4, the input parameters (coil structure parameters) of the BP neural network are optimized by a particle swarm algorithm (PSO): a particle swarm (a group of particles) is initialized, where each particle represents a group of coil parameters; the fitness value of the particle is calculated, and the fitness function is based on the performance error predicted by the BP neural network model (i.e., the fitness function is defined according to the prediction result of the BP neural network); the position and velocity of the particle are updated, the global optimal solution is searched, and the optimal coil parameters are output according to the optimal fitness value (i.e., the position and velocity of the particle are iteratively updated to finally find the coil parameter combination with the best performance).

[0017] The following is a technical solution further defined by the present invention. According to the coil parameters optimized by PSO, actual processing and manufacturing are carried out to produce an eddy current sensor, and the performance of the sensor is verified through experimental calibration to ensure that it can achieve high-precision linear measurement within the measurement range of 2 to 15 mm.

[0018] The following is a technical solution further defined by the present invention, wherein the measurement error of the sensor does not exceed ±1%.

[0019] The following is a technical solution further limited by the present invention, wherein the coil material is made of copper material with high conductivity, and the number of coil layers is optimized and determined according to simulation results.

[0020] Compared with the prior art, the present invention has the following technical effects:

[0021] The present invention uses BP neural network to model the nonlinear mapping relationship between coil parameters and sensor performance, and combines the global optimization capability of PSO algorithm to realize intelligent optimization of coil parameters. The sensor coil designed by optimization has good performance and can realize high-precision linear measurement in the range of 2 to 15 mm with an error of less than ±1%.

[0022] The method of the present invention is efficient and simple in design, can significantly reduce the time cost of experimental design and testing, and avoid the complexity of multiple trials in traditional design. The optimization result is directly applied to sensor manufacturing or sensor calibration in equipment, and has a wide range of applications.

[0023] The present invention is further described below in conjunction with the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0025] Figure 1 is a flow chart of the method of the present invention;

[0026] Figure 2 It is a schematic diagram of the calibration results of the eddy current sensor of the present invention. DETAILED DESCRIPTION

[0027] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without violating the connotation of the present invention, so the present invention is not limited by the specific embodiments disclosed below.

[0028] like Figure 1As shown, this embodiment proposes a method for optimizing the parameters of eddy current sensor coils based on a particle swarm algorithm (PSO) and a BP neural network model. The method achieves efficient optimization design of eddy current coil parameters by combining an intelligent algorithm with a neural network.

[0029] The method specifically includes: first, select high-conductivity copper as the material of the coil. Copper material has excellent conductivity and can significantly improve the electromagnetic properties of the coil. Then, preliminarily determine the key parameter range of the coil design, including the number of coil turns, inner diameter, line width, line height and the spacing range between coil turns. Based on experience and design requirements, the initial parameter range is set as: number of turns 50-150, inner diameter 10-30mm, line width 0.2-1mm, line height 0.1-0.8mm, spacing 0.1-0.5mm. The determination of these parameter ranges provides a basis for subsequent simulation and optimization.

[0030] Next, the above coil parameter combinations were analyzed using ANSYS finite element simulation software: by establishing a three-dimensional simulation model of the eddy current coil in the software, the parameter combinations were changed in sequence and the corresponding inductance, resistance and magnetic induction intensity data were calculated. During the simulation process, actual working conditions, such as operating frequency, current density and magnetic field distribution, need to be considered to ensure that the simulation results are accurate and reliable. Through multiple combinations of parameters, 243 sets of simulation data were finally obtained, which fully reflected the impact of different coil parameters on the performance of the eddy current sensor. These sample data provide sufficient data support for the subsequent training and verification of the BP neural network.

[0031] Based on the above simulation results, a BP neural network model is constructed to establish a nonlinear mapping relationship between coil parameters and sensor performance. Specifically, the input of the BP neural network model is the coil parameters (number of turns, inner diameter, line width, line height, spacing), and the output is the simulated sensor performance (inductance, resistance, magnetic induction intensity). The network structure design includes an input layer, two hidden layers and an output layer, where the number of neurons in the hidden layer is set to 10 to 15, and the specific number is adjusted through cross-validation to achieve the best fit effect. During the model training process, the Levenberg-Marquardt algorithm is used to optimize the network weights and biases, and the mean square error (MSE) is selected as the loss function. The training process continues until the error converges or the maximum number of iterations (such as 1000 times) is reached. By dividing the sample data into a training set (80%) and a test set (20%), the trained BP neural network can accurately predict the performance indicators corresponding to any set of input coil parameters, and its prediction error is less than 3%.

[0032] On the basis of the BP neural network model, the particle swarm algorithm (PSO) is further used to globally optimize the coil parameters. The PSO algorithm inputs the five parameters of the coil, namely the number of turns, inner diameter, line width, line height, and spacing. Then the trained BP neural network is called to predict the inductance, resistance, and magnetic induction intensity. The predicted data is returned to the PSO algorithm, and the value of the fitness function is calculated at this time. Then, it is compared and updated. Through continuous iteration, the optimal coil parameters are found. Through the iterative process of PSO, particles continuously update their position and speed in the search space, and search for the global optimal solution based on the fitness function. After about 100 iterations, it finally converges to the optimal coil parameter combination.

[0033] According to the results of PSO optimization, the eddy current coil is actually manufactured. After manufacturing, the coil is assembled with other parts of the eddy current sensor and the performance is tested. The test results show that the optimized eddy current sensor has excellent linearity and measurement range. Within the measurement range of 2 to 15 mm, the linear error of the sensor is less than ±1%. Figure 2 As shown in the figure, it meets the requirements of high-precision measurement and has better performance than the coil structure designed by traditional experience. Through this design process, the efficiency and performance of eddy current sensor design are greatly improved, providing reliable technical support for practical applications.

[0034] It should be noted that the sensor parameter optimization method of this embodiment can be applied to both the design of a single eddy current sensor and the performance improvement of sensors installed in the equipment, and the method has strong versatility.

[0035] In summary, the present invention uses the number of turns, inner diameter, line width, line height and spacing of the eddy current coil as optimization parameters, combines the global search capability of the PSO algorithm with the nonlinear fitting capability of the BP neural network, constructs a mapping model between sensor performance and coil parameters, and realizes intelligent optimization of the coil structure. The eddy current sensor manufactured by optimizing the coil structure parameters can achieve high-precision linear measurement within the range of 2 to 15 mm.

[0036] The above is only a preferred embodiment of the present invention, and does not limit the present invention in any form. Any technician familiar with the art can make many possible changes and modifications to the technical solution of the present invention by using the above disclosed methods and technical contents without departing from the scope of the technical solution of the present invention, or modify it into an equivalent embodiment of equivalent changes. Therefore, all equivalent changes made according to the shape, structure and principle of the present invention without departing from the content of the technical solution of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for optimizing coil parameters of an eddy current sensor, characterized in that: The following steps are involved: Step S1: Determine coil material and basic size parameters; Step S2: Generate sample data between coil parameters and performance based on finite element simulation analysis; Step S3: constructing a BP neural network model, training and fitting the nonlinear relationship between coil parameters and measurement performance; Step S4: Optimizing the BP neural network model using a particle swarm algorithm to achieve an optimal combination of coil parameters; Step S5: designing and manufacturing an eddy current sensor according to the optimization result.

2. The eddy current sensor coil parameter optimization method according to claim 1, characterized in that: The coil parameters include the number of turns, inner diameter, line width, line height and spacing.

3. The eddy current sensor coil parameter optimization method according to claim 1, characterized in that: The finite element simulation analysis is used to generate inductance, resistance and magnetic induction intensity data under different coil parameter combinations.

4. The eddy current sensor coil parameter optimization method according to claim 1, characterized in that: The BP neural network model includes an input layer, a hidden layer and an output layer, wherein: the input layer is used to receive coil parameters; the output layer outputs corresponding inductance, resistance and magnetic induction intensity.

5. The eddy current sensor coil parameter optimization method according to claim 1, characterized in that: In step S4, the particle swarm algorithm optimizes the coil parameters by the following steps: Step S41: Initialize a particle group, where each particle represents a set of coil parameters; Step S42: Calculate the fitness value of the particle, where the fitness function is based on the performance error predicted by the BP neural network model; Step S43: Update the position and velocity of the particle and search for the global optimal solution; Step S44: outputting optimal coil parameters according to the optimal fitness value.

6. The eddy current sensor coil parameter optimization method according to claim 1, characterized in that: The coil structure of the sensor manufactured in step S5 realizes linear measurement within a measurement range of 2 to 15 mm.

7. The method for optimizing coil parameters of an eddy current sensor according to claim 6, characterized in that: The measurement error of the sensor does not exceed ±1%.

8. The method for optimizing coil parameters of an eddy current sensor according to claim 6, wherein: The coil material is made of copper material with high conductivity, and the number of coil layers is optimized and determined according to simulation results.

Images