Fault detection method and device of magnetic suspension bearing system, electronic equipment and medium

Abstract

The invention discloses a fault detection method and device for a magnetic suspension bearing system, electronic equipment and a medium. The method comprises the following steps: acquiring real-time data of a control system of the magnetic suspension bearing in a control process; determining a first fault detection result according to the real-time data based on a fault diagnosis rule; processing the real-time data to obtain input data; inputting the input data into a pre-trained fault detection model for processing, and outputting a second fault detection result; and fusing the first fault detection result and the second fault detection result, and determining a system fault detection result of the control system of the magnetic suspension bearing. According to the embodiment of the invention, faster fault detection and more accurate fault prediction are realized.

Description

Technical Field

[0001] The present invention relates to the field of bearings, and in particular to a fault detection method, device, electronic equipment and medium for a magnetic suspension bearing system. Background Art

[0002] As a new type of bearing technology, magnetic bearings have shown broad application potential in multiple high-tech fields with their advantages such as non-contact, wear-free, high speed and high precision.

[0003] Due to the complexity of the magnetic bearing system, its electrical part is prone to various faults, which directly affect the stability and safety of the system.

[0004] Traditional fault diagnosis methods often have difficulty in balancing speed and accuracy, and are not effective in detecting faults with fuzzy signal characteristics. Summary of the invention

[0005] The present invention provides a fault detection method, device, electronic equipment and medium for a magnetic bearing system, which realize faster fault detection and more accurate fault prediction.

[0006] According to one aspect of the present invention, a fault detection method for a magnetic bearing system is provided, characterized in that it comprises:

[0007] Obtain real-time data of the control system of the magnetic bearing during the control process;

[0008] Based on the fault diagnosis rule, determining a first fault detection result according to the real-time data;

[0009] Processing the real-time data to obtain input data;

[0010] Inputting the input data into a pre-trained fault detection model for processing, and outputting a second fault detection result;

[0011] The first fault detection result and the second fault detection result are integrated to determine a system fault detection result of the control system of the magnetic bearing.

[0012] According to another aspect of the present invention, there is provided a fault detection device for a magnetic bearing system, characterized in that it comprises:

[0013] A bearing data acquisition module is used to acquire real-time data of the control system of the magnetic suspension bearing during the control process;

[0014] A first fault detection module, configured to determine a first fault detection result according to the real-time data based on a fault diagnosis rule;

[0015] An input data processing module, used for processing the real-time data to obtain input data;

[0016] A second fault detection module, used for inputting the input data into a pre-trained fault detection model for processing, and outputting a second fault detection result;

[0017] The fault result fusion module is used to fuse the first fault detection result and the second fault detection result to determine the system fault detection result of the control system of the magnetic bearing.

[0018] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0019] at least one processor; and

[0020] a memory communicatively connected to the at least one processor; wherein,

[0021] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the fault detection method for the magnetic bearing system described in any embodiment of the present invention.

[0022] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the fault detection method for a magnetic bearing system described in any embodiment of the present invention when executed.

[0023] The technical solution of the embodiment of the present invention obtains the control data of the magnetic levitation bearing and adopts two methods to perform fault detection to obtain a first fault detection result and a second fault detection result. It can realize rule-based rapid fault diagnosis and model-based accurate fault diagnosis, taking into account the accuracy and speed of fault detection, reducing unnecessary maintenance and downtime, and reducing maintenance costs. The first fault detection result and the second fault detection result are merged to obtain a system fault detection result, achieving consistency between the two methods and enhancing the reliability of the fault detection results.

[0024] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. 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 creative work.

[0026] Figure 1 is a flow chart of a fault detection method for a magnetic bearing system provided according to an embodiment of the present invention;

[0027] Figure 2 This is a front view of a magnetic suspension bearing provided according to an embodiment of the present invention;

[0028] Figure 3 is a front cross-sectional view of a magnetic suspension bearing provided according to an embodiment of the present invention;

[0029] Figure 4 is a top view of a magnetic bearing provided according to an embodiment of the present invention;

[0030] Figure 5 is a top view cross-sectional view of a magnetic bearing provided according to an embodiment of the present invention;

[0031] Figure 6 is a structural schematic diagram of a control system provided according to an embodiment of the present invention;

[0032] Figure 7 is a schematic diagram of a decision tree provided according to an embodiment of the present invention;

[0033] Figure 8 is a structural diagram of a fault detection model provided according to an embodiment of the present invention;

[0034] Fig. 9 is a scene diagram of a fault detection method for a magnetic bearing system provided according to an embodiment of the present invention;

[0035] Fig.10 is a structural schematic diagram of a fault detection device for a magnetic bearing system provided according to an embodiment of the present invention;

[0036] Fig.11 It is a structural schematic diagram of an electronic device for implementing the fault detection method of the magnetic bearing system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0037] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0038] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0039] Figure 1 A flow chart of a fault detection method for a magnetic bearing system provided by an embodiment of the present invention. The embodiment of the present invention can be applied to the fault detection of a control system for controlling the suspension and rotation of a rotor in a magnetic bearing. The method can be executed by a fault detection device for a magnetic bearing system. The fault detection device for a magnetic bearing system can be implemented in the form of hardware and / or software. The fault detection device for a magnetic bearing system can be configured in an electronic device with a certain data computing capability, and the electronic device can be a server or a client.

[0040] S101. Acquire real-time data of the control system of the magnetic bearing during the control process.

[0041] The magnetic bearing may be a three-degree-of-freedom magnetic bearing. The control system of the magnetic bearing is used to control the suspension and rotation of the rotor in the magnetic bearing. The control system uses physical information neural network predictive control to fine-tune the rotor so that the rotor remains suspended. The control process may refer to the process in which the control system adjusts the position of the rotor in the magnetic bearing. Real-time data may refer to data that can be collected by the control system during the control process, and the real-time data may include data of each module in the control system and / or data of the magnetic bearing.

[0042] In some embodiments, Figure 2 A front view of a magnetic bearing is provided. Figure 3 A front cross-sectional view of a magnetic bearing is provided. Figure 4 A top view of a magnetic bearing is provided. Figure 5 A top view cross-sectional view of a magnetic suspension bearing is provided. The magnetic suspension bearing is a bearing with three degrees of freedom.

[0043] like Figure 2-5 As shown, the magnetic bearing includes: an electromagnetic ring 5, a first end cover 4-A, a second end cover 4-B, a rotor 6, an X-axis electromagnetic coil 8, a Y-axis electromagnetic coil 9 and a Z-axis electromagnetic coil 7; the first end cover 4-A and the second end cover 4-B are arranged on the electromagnetic ring 5, and the first end cover 4-A and the second end cover 4-B are respectively arranged on both sides of the electromagnetic ring 5, and the first end cover 4-A and the second end cover 4-B are parallel; the rotor 6 is placed inside the electromagnetic ring 5; the X-axis electromagnetic coil 8 and the Y-axis electromagnetic coil 9 are arranged on the inner side of the electromagnetic ring 5; the Z-axis electromagnetic coil 7 is arranged on the inner side of the first end cover 4-A and the inner side of the second end cover 4-B.

[0044] The first end cover 4-A and the second end cover 4-B can be in a top-bottom relationship with each other. The X-axis electromagnetic coil 8, the Y-axis electromagnetic coil 9, the Z-axis electromagnetic coil 7 and the rotor 6 are non-contacting. The X-axis electromagnetic coil 8, the Y-axis electromagnetic coil 9 and the Z-axis electromagnetic coil 7 are used to generate a magnetic field when energized to control the movement of the rotor 6 in different directions, wherein the movement directions of the rotor 6 controlled by the X-axis electromagnetic coil 8, the Y-axis electromagnetic coil 9 and the Z-axis electromagnetic coil 7 are perpendicular to each other.

[0045] The X-axis electromagnetic coil 8 is installed on the inner side of the electromagnetic ring 5, responsible for controlling the position of the rotor 6 in the X-axis direction; the Y-axis electromagnetic coil 9 is installed on the inner side of the electromagnetic ring 5, responsible for controlling the position of the rotor 6 in the Y-axis direction; the Z-axis electromagnetic coil 7 is installed on the inner side of the first end cover 4-A and the second end cover 4-B, responsible for controlling the position of the rotor 6 in the Z-axis direction. In this way, by integrating the axial magnetic bearing with the radial magnetic bearing, a three-degree-of-freedom hybrid magnetic bearing with a compact structure is realized. In addition, except for a few parts such as the electromagnetic coil, the magnetic bearing is made of aluminum alloy as a whole, and the adjustment device and the bearing are designed as an integral whole, which reflects its advantages of miniaturization and lightweight compared with existing magnetic bearings.

[0046] In addition, each shaft electromagnetic coil may be provided with an anti-collision gasket at one end close to the rotor, and each shaft electromagnetic coil or a part of the electromagnetic coil may be provided with a Hall sensor.

[0047] It can be seen that this embodiment, by setting a three-degree-of-freedom magnetic bearing, can place the rotor in the space formed by the electromagnetic ring, the first end cover and the second end cover, so as to realize a compact bearing structure, reduce the volume of the magnetic bearing, simplify the manufacturing and installation difficulty, can be installed and applied in space-constrained scenarios, can meet the application requirements of small spaces, and can realize fault detection for the three-degree-of-freedom magnetic bearing, can adapt to general magnetic bearings for fault detection, increase the bearing range of fault detection, and increase the versatility of the fault detection method.

[0048] In some embodiments, Figure 6 A schematic diagram of a control system for a magnetic bearing is provided. Figure 6 As shown, the control system of the magnetic bearing includes: a magnetic bearing, a controller, a power amplifier and a displacement sensor; the displacement sensor is used to obtain the actual position of the rotor; the controller obtains the position deviation between the actual position and the expected position of the rotor, and outputs a pulse width modulation signal according to the position deviation; the power amplifier is used to output a control current according to the pulse width modulation signal; the electromagnetic coil is used to convert the control current into an electromagnetic thrust and adjust the position of the rotor.

[0049] The controller includes an X-axis controller, a Y-axis controller and a Z-axis controller, the power amplifier includes an X-axis power amplifier, a Y-axis power amplifier and a Z-axis power amplifier; the displacement sensor includes an X-axis displacement sensor, a Y-axis displacement sensor and a Z-axis displacement sensor. For each axial direction, the coaxial controller is connected to the power amplifier, the power amplifier is connected to the electromagnetic coil, the electromagnetic coil is connected to the rotor, the rotor is connected to the displacement sensor, the displacement sensor is connected to the subtractor, and the subtractor is connected to the controller.

[0050] Specifically, x, y, and z are the output voltages of the X-axis, Y-axis, and Z-axis sensors, respectively, and the voltages are proportional to the rotor position; i x 、i y 、i z They are the currents on the coils on the X-axis, Y-axis, and Z-axis; pwm x 、pwm y 、pwm z They are the output PWM signals of the X-axis, Y-axis, and Z-axis controllers; , , They are the low-frequency components of the PWM signals of the X-axis, Y-axis, and Z-axis controllers; v dc is the voltage of the bus power supply, i dc is the current of the bus power supply. Among them, x, y, z, ix, iy, iz, pwmx, pwmy, and pwmz come from the control system of the magnetic bearing, which is a closed-loop control.

[0051] Specifically, for the X-axis direction, the X-axis controller obtains the position deviation of the rotor in the direction of the corresponding coordinate axis, that is, x ref The X-axis controller outputs a PWM signal pwm according to the duty cycle of the control algorithm x ; pwm x Input to the X-axis power amplifier, the X-axis power amplifier adopts H-bridge topology, according to the signal pwm x Output corresponding current i x ; The X-axis electromagnetic coil receives and converts the current i x Converted into electromagnetic thrust to adjust the rotor position; the X-axis displacement sensor uses an eddy current displacement sensor to output a voltage signal x that is proportional to the rotor position. The subtraction operator calculates the reference voltage x ref The difference between .

[0052] For the Y-axis direction, the Y-axis controller obtains the position deviation of the rotor in the corresponding coordinate axis direction, that is, y ref The Y-axis controller outputs a PWM signal pwm according to the duty cycle of the control algorithm y ; pwm y Input to the Y-axis power amplifier, the Y-axis power amplifier adopts H-bridge topology, according to the signal pwm y Output corresponding current i y ; The Y-axis electromagnetic coil receives and converts the current i y Converted into electromagnetic thrust to adjust the rotor position; the Y-axis displacement sensor uses an eddy current displacement sensor to output a voltage signal y that is proportional to the rotor position. The subtraction operator calculates the reference voltage y ref The difference between y and y.

[0053] For the Z-axis direction, the Z-axis controller obtains the position deviation of the rotor in the corresponding coordinate axis direction, that is, z ref The Z-axis controller outputs a PWM signal pwm according to the duty cycle of the control algorithm z ; pwm z Input to the Z-axis power amplifier, the Z-axis power amplifier adopts H-bridge topology, according to the signal pwm z Output corresponding current i z ; The Z-axis electromagnetic coil receives and converts the current i z Converted into electromagnetic thrust to adjust the rotor position; the Z-axis displacement sensor uses an eddy current displacement sensor to output a voltage signal z that is proportional to the rotor position. The subtraction operator calculates the reference voltage z ref The difference between .

[0054] Through the control system, high-precision control of the magnetic bearing can be achieved, the safety of the magnetic bearing can be improved, and noise and vibration can be reduced.

[0055] S102: Based on a fault diagnosis rule, determine a first fault detection result according to the real-time data.

[0056] The fault diagnosis rule can be a preset standard that can be used to detect faults in the magnetic suspension bearing and the control system. Multiple fault diagnosis rules can be preset, based on which the real-time data satisfies the fault diagnosis rule, and based on the satisfied fault diagnosis rule, it is determined whether there is a fault, and a first fault detection result is obtained.

[0057] S103: Process the real-time data to obtain input data.

[0058] The real-time data may be determined as the input data, or the real-time data may be processed to obtain the input data. In some embodiments, key data is extracted from the real-time data, and the key data to be processed is pre-processed to obtain the input data.

[0059] S104: input the input data into a pre-trained fault detection model for processing, and output a second fault detection result.

[0060] The fault detection model is used to process the input data and output the second fault detection result. The input data is data determined based on the control data of the control system. The fault detection model can obtain the operating data of the magnetic bearing and the operating data of the control system from the input data, obtain the information of the fault whose signal characteristics are not obvious, and process and detect to obtain the second fault detection result. In this way, the second fault detection result is equivalent to a supplement to the first fault detection result, which can break through the limitations of rule-based fault diagnosis and thus improve the accuracy of the overall fault detection result.

[0061] S105: Fusing the first fault detection result and the second fault detection result to determine a system fault detection result of the control system of the magnetic bearing.

[0062] The first fault detection result and the second fault detection result are integrated, and a consistent fault detection result can be selected as the system fault detection result; or if the first fault detection result and the second fault detection result are inconsistent, the first fault detection result and the second fault detection result are spliced ​​to obtain the system fault detection result, or the fault detection result with a high priority is determined as the system fault detection result.

[0063] The first fault detection result obtained by rule-based fault detection is actually a fault detection result determined for a normal, simple signal with obvious fault characteristics; the second fault detection result obtained by model-based fault detection is actually a fault detection result determined for an abnormal, complex signal with vague fault characteristics. The fusion of the first fault detection result and the second fault detection can take into account both normal, simple signals with obvious fault characteristics and abnormal, complex signals with vague fault characteristics for fault detection, enriching the application scenarios applicable to fault detection; at the same time, rule-based fault detection is fast and low-cost, while model-based fault detection is highly accurate. The fusion of the first fault detection result and the second fault detection can effectively improve the efficiency and accuracy of fault detection.

[0064] The technical solution of the embodiment of the present invention obtains the control data of the magnetic levitation bearing and adopts two methods to perform fault detection to obtain a first fault detection result and a second fault detection result. It can realize rule-based rapid fault diagnosis and model-based accurate fault diagnosis, taking into account the accuracy and speed of fault detection, reducing unnecessary maintenance and downtime, and reducing maintenance costs. The first fault detection result and the second fault detection result are merged to obtain a system fault detection result, achieving consistency between the two methods and enhancing the reliability of the fault detection results.

[0065] In some embodiments, the fault diagnosis rule-based method determines the first fault detection result according to the real-time data, including: detecting whether the electromagnetic coil current in the real-time data is within a reasonable current range; when the electromagnetic coil current is greater than the reasonable current range, determining the first fault detection result according to a pulse width modulation signal; when the electromagnetic coil current is less than the reasonable current range, determining the first fault detection result according to the DC voltage, DC current, coil current and pulse width modulation signal; determining the first fault detection result according to the signal of the displacement sensor in the control system and the change of the coil current.

[0066] Based on the fault diagnosis rules, a decision tree can be used to determine whether the magnetic bearing system has a fault through the input and output signals of multiple links of the magnetic bearing closed-loop control system. In order to improve the detection algorithm, an FPGA (Field-Programmable Gate Array) chip can be used to implement the fault detection based on the fault diagnosis rules.

[0067] Figure 7 A schematic diagram of a fault diagnosis decision tree is provided, such as Figure 7As shown, it is possible to determine whether each fault diagnosis rule is satisfied based on coil current, PWM, DC voltage, DC current and displacement sensor signal. The reasonable current range can be a current range obtained by experimentally counting the current values ​​in the historical operation data of the coil current of the magnetic bearing. The reasonable current range is used to distinguish the current in normal operation from the current in the event of a fault. When the coil current in the real-time data is greater than or less than the reasonable current range, it indicates that the coil current is abnormal and there is a corresponding fault, and fault diagnosis can be performed in different subdivided situations. When the coil current in the real-time data is within the reasonable current range, it indicates that the coil current is fault-free, and other rules need to be judged, for example, detecting the output signal of the only sensor, detecting the change in the signal output by the displacement sensor and the change in the coil current, and judging whether there is a fault based on the change in the displacement sensor and the change in the coil current.

[0068] If the coil current is greater than the reasonable current range, make a judgment based on the PWM signal. Determine whether the characteristic value of the PWM signal is within the normal range. Specifically, if the amplitude, duty cycle and frequency of the corresponding PWM signal are within the normal range (the corresponding normal range can be obtained based on historical operation data statistics for each parameter item), then output a coil short circuit fault signal. Determine the duty cycle of the PWM signal. Specifically, if the corresponding PWM signal duty cycle is close to or equal to 1, output a controller fault signal. Determine the consistency of the PWM signal. Specifically, if the corresponding power amplifier has the same bridge arm PWM signal that is 1, output a controller fault signal. If none of the above situations occur, output other fault signals.

[0069] If the coil current is less than the reasonable current range, judge the DC voltage and DC current. Specifically, if the DC bus voltage and current are both too small, output the bus power supply fault signal; if the DC bus voltage is too small and the DC bus current is too large, output the switch tube disconnection signal. Determine whether other currents are within the normal range. Specifically, if other currents (other coil currents and DC bus currents) are within the reasonable range, output the current sensor fault signal. Determine whether the PWM signal is normal and whether the coil current changes. Specifically, if the corresponding PWM signal amplitude is too small or equal to 0, output the controller fault signal; if the corresponding PWM signal is normal and the coil current is sampled multiple times in a switching cycle and is 0, output the coil disconnection signal; if the corresponding PWM signal is normal, there is a certain difference in the multiple sampling values ​​of the coil current in a switching cycle, output the switch tube disconnection signal. If none of the above situations occur, output other fault signals.

[0070] When the coil current in the real-time data is within a reasonable current range, if the output signal of the displacement sensor remains unchanged for a period of time, but there is a certain change in the coil current during the same period, or the displacement sensor output is fully biased or equal to 0, a displacement sensor fault signal is output.

[0071] According to the above rules, it is detected whether a fault signal is output. When a fault signal is output, the first fault detection result is determined to be the presence of a fault and the type of the existing fault; when a fault signal is not output, the first fault detection result is determined to be the absence of a fault.

[0072] In addition, if a fault signal is output, the fault signal is determined as the first fault detection result, and at the same time, all PWM signals are blocked and the DC bus power supply is turned off.

[0073] By adopting multiple fault diagnosis rules to perform fault detection of the control system, fault detection can be quickly realized based on real-time data, and fault signals with obvious characteristics can be diagnosed quickly and accurately.

[0074] In some embodiments, the processing of the real-time data to obtain input data includes: extracting a rotor position voltage from the real-time data; extracting a low-frequency component from a pulse width modulation signal in the real-time data to obtain a low-frequency modulation signal; extracting an electromagnetic coil current from the real-time data; and determining the rotor position voltage, the low-frequency modulation signal, and the electromagnetic coil current as input data.

[0075] The rotor position voltage may refer to the voltage corresponding to the position reached by the rotor. The rotor position voltage may refer to the voltage output by the displacement sensor that is proportional to the rotor position. In fact, the neural network model can learn the key content of the low-frequency modulation signal, and can extract the low-frequency component of the pulse width modulation signal to obtain a low-frequency modulation signal. Replacing the pulse width modulation signal with a low-frequency modulation signal can reduce the amount of data of the input signal, thereby reducing the amount of data processed by the model and improving the efficiency of the model in predicting fault detection results. The infinite impulse response filtering algorithm can be applied to obtain the low-frequency component of the PWM signal ( , , ). The electromagnetic coil current may include a current input to the X-axis electromagnetic coil, a current input to the Y-axis electromagnetic coil, and a current input to the Z-axis electromagnetic coil.

[0076] By extracting the rotor position voltage, low-frequency modulation signal and electromagnetic coil current from real-time data, a small amount of input can be used to extract fuzzy fault features from them to form input data for model fault detection, which can improve the efficiency of model fault detection. At the same time, it can assist the first fault detection result based on the fault diagnosis rule judgment, supplement the limitations of the rule-based fault detection method, and thus improve the accuracy of the overall fault detection.

[0077] In some embodiments, the fault detection model is trained as follows: obtain signal data of the control system of the magnetic bearing under normal operation and signal data under various fault types, and generate a training sample set; use the training sample set to train the fault detection model to obtain a trained fault detection model.

[0078] For faults with clear signal characteristics and relatively serious faults, the above-mentioned fast fault diagnosis method based on fault diagnosis rules can be used. For other faults with vague signal characteristics, the effect of the above-mentioned fault diagnosis method based on fault diagnosis rules is not ideal. For example, the gain inaccuracy and linearity degradation of the position sensor caused by external interference, and the large coil current ripple caused by changes in system parameters (partial short circuit of the electromagnetic coil, deviation in the attenuation ratio of the current sensor, and partial short circuit of the feedback resistor, etc.).

[0079] In order to improve the accuracy of magnetic bearing fault detection, an embodiment of the present invention proposes a fault detection algorithm based on a fault detection model. The fault detection model can be a physical information neural network (Physics-informed Neural Networks, PINNs). The fault detection model combines the laws of physics with neural networks. By minimizing the error between the prediction and the true value and considering the constraints of the laws of physics, the model accuracy and reliability are improved. The advantage of PINNs is that it can use a small amount of data to learn a model with strong generalization ability. It is suitable for complex systems, such as magnetic bearing fault detection, and effectively improves the detection efficiency and accuracy. The fault detection model involves a large number of mathematical operations. In order to balance the operation speed and system cost, it is chosen to be placed in a DSP (Digital Signal Processing) chip for operation.

[0080] Figure 8 A structural diagram of a fault detection model is provided, such as Figure 8 As shown, the process of training a fault detection model may include:

[0081] 1. Data collection and preprocessing: Collect signal data of various states (normal or faulty). The signal data and fault detection results of normal operation form at least one training sample for no fault, where the signal data is used as input data and the content of the fault detection result is used as the true value detection result. The signal data and fault detection results of a fault type form a training sample for the fault of that fault type, where the signal data is used as input data and the content of the fault detection result is used as the true value detection result. A training sample includes input data and true value detection results. The input data is generated based on the acquired signal data.

[0082] The operating data that needs to be collected include: the output voltage x, y, z of the X-axis, Y-axis, and Z-axis position sensors; the current i on the X-axis, Y-axis, and Z-axis coils; x 、i y 、i z , low-frequency components of the PWM signals of the X-axis, Y-axis, and Z-axis controllers , , .

[0083] The data sources include historical operating data and factory test data of magnetic bearings.

[0084] The input data can be manually calibrated, that is, to determine whether the input data has labels of faults and fault types.

[0085] 2. Establish a physical information neural network model: Determine the corresponding physical equations based on the physical characteristics and failure mechanism of the magnetic bearing. Build a neural network architecture and incorporate physical information as constraints into the loss function to ensure that the model learns both data features and complies with physical laws.

[0086] The physical information neural network model consists of three layers: input layer, hidden layer and output layer. The input layer contains 9 nodes, namely x, y, z, i x 、i y 、i z , , , .

[0087] The hidden layer contains 9 nodes. The output layer contains 4 nodes, which represent the four faults: the position sensor is disturbed (F1), the electromagnetic coil is partially short-circuited (F2), the current sensor attenuation ratio is deviated (F3), and the feedback resistor is partially short-circuited (F4). The input-output relationship of the hidden layer is shown in Equation 1.

[0088] (1)

[0089] In the formula, represents the weight coefficient of the hidden layer, f1 represents the activation function, which is a hyperbolic tangent function. O2 represents the output signal of the neurons in the second layer (hidden layer) of the neural network, and net2 is the linear part of the neuron model in this layer.

[0090] The input-output relationship of the output layer is shown in Equation 2.

[0091] (2)

[0092] In the formula, represents the weight coefficient of the output layer, f2 represents the activation function, which is the ReLU function. O3 represents the output signal of the neurons in the third layer (output layer) of the neural network, and net3 is the linear part of the neuron model in this layer.

[0093] 3. Model training: Use the preprocessed data set to train the PINNs model. Calculate the output through forward propagation and compare it with the actual fault state. Use the back propagation algorithm to adjust the network weights and biases to minimize the sum of the prediction error and the physical constraint error.

[0094] The input data in the training sample is input into the PINNs model to obtain the prediction result, and the difference between the prediction result and the true value detection result in the training sample is detected. According to the difference corresponding to each training sample, the loss function is calculated, and the PINNs model is adjusted according to the loss function until the value of the loss function is minimized or converged. The training is determined to be completed, and the PINNs model at the current moment is determined as the trained PINNs model.

[0095] The loss function L is shown in Formula 3:

[0096] (3)

[0097] In the formula, is the vacuum magnetic permeability. Ar is the projected area of ​​the radial magnetic pole, which can be obtained through ANSYS software simulation analysis. cx1 , Φ cx2 , Φ cy1 , Φ cy2 and Φ cz are the controlled magnetic flux of the magnetic bearing in the X-axis direction, Y-axis direction and Z-axis direction, as shown in Equation 4; Φ px1 , Φ px2 , Φ py1 , Φ py2 , Φ pz1 and Φ pz2 are the magnetic flux of the bias magnetic field of the magnetic bearing in the X-axis direction and the Y-axis direction, as shown in Formula 5-7.

[0098] (4)

[0099] (5)

[0100] (6)

[0101] (7)

[0102] In formula 4-7, N r is the number of turns of radial excitation coil, N zis the number of turns of the axial excitation coil, k i Represents the magnetic resistance of each part of the radial bearing, R y is the magnetic resistance of the stator yoke, R iz is the soft magnetic reluctance, R1~R6 are the reluctance of each axial part, R m is the permanent magnet reluctance, R s is the leakage magnetic resistance. scz is the equivalent magnetic resistance of the leakage magnetic circuit in the radial control magnetic field, R ib is the soft magnetic reluctance of the bias magnetic circuit. m is the magnetomotive force generated by the inner ring of the permanent magnetic bearing. All the above-mentioned magnetic resistances are constants. The axial and radial three-dimensional magnetic fields can be equivalent to a two-dimensional equivalent model through the ANSYS axisymmetric model. After simulation, the values ​​of the above parameters can be obtained.

[0103] In the embodiment of the present invention, the loss function L shown in Formula 3 embeds the mathematical model of the magnetic bearing (Formula 4-Formula 7). The mathematical model in the loss function provides guidance for the parameter optimization direction of the neural network model. The ultimate desired effect is that under the same input, the neural network model and the mathematical model have similar outputs. The differential result of the loss function L is the gradient of L relative to the model parameters, which is used for parameter update in the back propagation stage. During the back propagation process:

[0104] (1) Define functions and variables: variables are the parameters of each neuron; the function is a composite function, the outer function is the loss function L, and the inner function is the forward propagation function of the neural network model (that is, the neural network model itself).

[0105] (2) Constructing a computational graph: In a neural network, each operation is considered a node in a computational graph. The computational graph records all sequences of operations from input to output.

[0106] (3) Initialize the gradient: At the end of the computational graph, that is, at the output node, the gradient is initialized to a vector of all 1s. This represents the rate of change of the output with respect to itself.

[0107] (4) Backpropagation: Starting from the output node, the gradient is propagated backward (toward the input direction) along the computational graph. At each node, the chain rule is applied to calculate and update the gradient.

[0108] (5) Calculate the gradient: At each node, the gradient is updated using the derivative of the operation corresponding to the node. This process continues until the input node of the computational graph is reached.

[0109] (6) Storing gradients: Store the calculated gradients at each node. For input nodes, these gradients are the derivatives of the function with respect to the input variables (model parameters).

[0110] (7) Reading gradients: Once all nodes in the computational graph have calculated and stored the gradients, the derivatives can be read from the input nodes.

[0111] In the embodiment of the present invention, through the deep learning framework, when a model is built and the back propagation function is called, the gradient of each node in the entire calculation graph is automatically calculated, and these gradients are used to update the parameters of the model.

[0112] Add a parameter update function as shown below.

[0113]

[0114] This formula introduces a momentum term, which means that when updating parameters, not only the current gradient is considered, but also the previous gradient direction and magnitude, so as to accumulate momentum during the optimization process. This can reduce random oscillations in the flat area of ​​the objective function, and can accelerate the learning process when the gradient direction is consistent, and reduce the update amplitude when the gradient direction is inconsistent.

[0115] In the formula, θ is the parameter of the model; vt is the momentum term, as shown in the following formula.

[0116]

[0117] Where γ is the momentum coefficient (determined by experiments to be 0.5), η is the learning rate (determined by experiments to be 0.02), is the gradient of the loss function L with respect to the parameter θ.

[0118] 4. Fault detection and classification: After training, the PINNs model is determined as a fault detection model and applied to the new magnetic bearing vibration signal data. The fault detection model output will indicate the health status of the bearing or the specific fault type, achieving real-time and accurate fault detection and classification.

[0119] In the magnetic bearing fault detection algorithm, after training with the physical information neural network model, the accuracy of the model was verified through the simulation platform and the experimental platform. On the simulation platform, MATLAB was used to build a dynamic model of the magnetic bearing system, and different types of fault signals were injected, such as displacement sensor failure and electromagnet failure. The fault detection model had an accurate recognition rate of more than 95%. On the experimental platform, a magnetic bearing experimental device was built to collect operating data in real time. The fault detection model had a classification accuracy of 92% for actual faults, which effectively verified the efficiency and accuracy of the physical information neural network model in magnetic bearing fault detection. These results demonstrate the potential and value of the algorithm in practical engineering applications. At the same time, the fault detection results can break through the limitations of rule-based fault detection results, enrich the detectable fault range, and improve the accuracy of fault detection.

[0120] In some embodiments, the first fault detection result and the second fault detection result are merged to determine the system fault detection result of the control system of the magnetic bearing, including: when the first fault detection result and the second fault detection result are different, determining that the system fault detection result of the control system of the magnetic bearing is the first fault detection result.

[0121] When the first fault detection result and the second fault detection result are consistent, any fault detection result is determined as the system fault detection result of the control system of the magnetic bearing; when the first fault detection result and the second fault detection result are inconsistent, a fault detection result with a higher priority can be selected according to the priority and determined as the system fault detection result of the control system of the magnetic bearing. Generally, the priority of the fault detection result determined based on the fault diagnosis rule is higher than the fault detection result output based on the fault detection model, that is, the priority of the first fault detection result is higher than the second fault detection result.

[0122] Fig. 9 A scenario diagram of a fault detection method for a magnetic bearing system is provided, such as Fig. 9 As shown, both the first fault detection result and the second fault detection result can be input into the gate 3 for fusing the first fault detection result and the second fault detection result. The gate 3 is responsible for summarizing the outputs of the fault detection method 1 based on the fault diagnosis rule and the fault detection method 2 based on the fault detection model, and giving the final detection result. Among them, the fault detection method 1 based on the fault diagnosis rule has a higher priority. When the two fault detection results are inconsistent, the first fault detection result will overwrite the second fault detection result.

[0123] By determining that the system fault detection result of the control system of the magnetic bearing is the first fault detection result when the first fault detection result and the second fault detection result are different, the fault detection results can be integrated, fault-tolerant control can be implemented, and the reliability of the fault detection system can be improved.

[0124] The embodiments of the present invention have the following advantages:

[0125] (1) Integration of fault diagnosis methods: The embodiment of the present invention proposes an integrated fault diagnosis method, which combines FPGA-based fast fault diagnosis (a fault detection method based on fault diagnosis rules) and DSP-based physical information neural network (PINNs) model (a fault detection method based on a fault detection model) for fault detection of three-degree-of-freedom magnetic bearings.

[0126] (2) Fast fault diagnosis algorithm: The fast fault diagnosis algorithm running on the FPGA in the embodiment of the present invention adopts a decision tree method and can quickly identify the fault type based on real-time data.

[0127] (3) Physical Information Neural Network Model: In the embodiment of the present invention, a PINNs-based fault prediction algorithm runs on a DSP. This algorithm combines physical laws with neural networks. It improves the accuracy and reliability of the model by minimizing the error between the predicted result output by the fault detection model and the true value and taking into account the constraints of physical laws.

[0128] (4) Selector: In the embodiment of the present invention, the selector combines the two fault detection results and gives the final detection result. When the results of the two algorithms are inconsistent, the fast fault diagnosis algorithm has a higher priority.

[0129] The above content can improve the accuracy and real-time performance of magnetic bearing fault detection, while reducing the dependence on a large amount of historical data and improving the system's adaptability to complex working conditions. The embodiment of the present invention adopts a more advanced physical information neural network (PINNs) model, which not only analyzes current characteristics, but also combines knowledge from multiple physical fields such as electromagnetism and mechanics, thereby improving the accuracy and generalization of fault diagnosis; the PINNs model adopted at the same time further combines the laws of physics, optimizes the network training process through physical constraints, and thus can obtain better diagnostic results even with less data; and through the combination of real-time data and physical models, faster fault detection and more accurate fault prediction are achieved, especially in the processing of unbalanced samples and complex working conditions. This integrated fault diagnosis method not only includes fault-tolerant control, but also covers fault detection and prediction. Through the collaborative work of FPGA and DSP, a more comprehensive fault diagnosis solution is achieved.

[0130] In summary, the embodiments of the present invention not only improve the accuracy and speed of fault diagnosis, but also enhance the reliability and adaptability of the system by integrating advanced PINNs model and hardware acceleration technology, especially showing obvious advantages in processing complex working conditions and unbalanced samples.

[0131] In one application scenario, an embodiment of the present invention provides a specific application solution of a magnetic bearing and its control system in a high-speed motor.

[0132] With the development of industrial automation and precision manufacturing technology, the demand for high-speed motors in various fields is growing. These motors require extremely high precision and reliability during operation, which puts higher requirements on the stability and fault detection capabilities of the bearing system. Magnetic bearings are an ideal choice for high-speed motors due to their non-contact and frictionless characteristics.

[0133] The high-speed motor integrated with the magnetic bearing provided by the embodiment of the present invention has higher performance and reliability. The motor is designed for extreme environments and high load conditions, and the speed can reach 100,000 revolutions per minute.

[0134] Fault detection of magnetic bearing control system:

[0135] 1. Data collection and preprocessing: The operating data of the magnetic bearing under different speeds and load conditions were collected, including voltage, current and PWM signals.

[0136] 2. Model building and training: Using this data, a physical information neural network model is constructed as a fault detection model, which can learn the behavioral characteristics of the bearing in healthy and faulty states.

[0137] 3. Fault detection and classification: During the operation of the motor, the model monitors the status of the magnetic bearing in real time, and immediately classifies and alarms if there are signs of fault.

[0138] The final effects achieved by the motor include:

[0139] 1. Improve detection speed: Through the fast fault diagnosis algorithm on the FPGA chip, the speed of fault detection has been greatly improved, and it can respond immediately at the early stage of the fault.

[0140] 2. Improved accuracy: The fault prediction algorithm based on PINNs improves the accuracy of fault prediction, especially when there is less data or noise.

[0141] 3. Reduced maintenance costs: Due to the improved accuracy of fault detection, unnecessary maintenance and downtime are reduced, thereby reducing maintenance costs.

[0142] 4. Enhanced system reliability: The design and application of the gate ensures consistent decisions between the fast fault diagnosis algorithm and the PINNs model, enhancing the reliability of the entire system.

[0143] In an actual operation, a subtle displacement sensor failure occurred in the magnetic bearing while the motor was running under high load conditions. The FPGA-based fast fault diagnosis algorithm first detected the abnormal current signal and triggered a preliminary alarm. Subsequently, the PINNs model further analyzed the fault characteristics, confirmed the problem with the displacement sensor, and predicted the serious consequences that might result if no action was taken. The gate combined the outputs of the two algorithms, confirmed the fault, and initiated the emergency shutdown procedure for the motor, avoiding possible equipment damage and production interruption.

[0144] The application of the embodiments of the present invention in high-speed motors not only improves the speed and accuracy of fault detection, but also reduces maintenance costs and enhances system reliability, which proves the potential and value of the embodiments of the present invention in practical engineering applications.

[0145] Fig.10 A schematic diagram of the structure of a fault detection device for a magnetic bearing system provided by an embodiment of the present invention. The embodiment of the present invention can be applied to the case of controlling a magnetic bearing, the device can execute a fault detection method for a magnetic bearing system, the device can be implemented in the form of hardware and / or software, and the device can be configured in an electronic device that carries a function with certain data computing capabilities.

[0146] See also Fig.10 The fault detection device of the magnetic bearing system shown comprises:

[0147] The bearing data acquisition module 1010 is used to acquire the real-time data of the control system of the magnetic suspension bearing during the control process;

[0148] A first fault detection module 1020, configured to determine a first fault detection result according to the real-time data based on a fault diagnosis rule;

[0149] An input data processing module 1030 is used to process the real-time data to obtain input data;

[0150] A second fault detection module 1040, configured to input the input data into a pre-trained fault detection model for processing and output a second fault detection result;

[0151] The fault result fusion module 1050 is used to fuse the first fault detection result and the second fault detection result to determine the system fault detection result of the control system of the magnetic bearing.

[0152] In some embodiments, the magnetic bearing includes: an electromagnetic ring, a first end cover, a second end cover, a rotor, an X-axis electromagnetic coil, a Y-axis electromagnetic coil and a Z-axis electromagnetic coil; the first end cover and the second end cover are arranged on the electromagnetic ring, the first end cover and the second end cover are respectively arranged on both sides of the electromagnetic ring, and the first end cover and the second end cover are parallel; the rotor is placed inside the electromagnetic ring; the X-axis electromagnetic coil and the Y-axis electromagnetic coil are arranged on the inner side of the electromagnetic ring; the Z-axis electromagnetic coil is arranged on the inner side of the first end cover and the inner side of the second end cover.

[0153] In some embodiments, the control system of the magnetic bearing comprises: the magnetic bearing, a controller, a power amplifier and a displacement sensor;

[0154] The displacement sensor is used to obtain the actual position of the rotor;

[0155] The controller obtains a position deviation between an actual position of the rotor and a desired position, and outputs a pulse width modulation signal according to the position deviation;

[0156] The power amplifier is used to output a control current according to the pulse width modulation signal;

[0157] The electromagnetic coil is used to convert the control current into electromagnetic thrust and adjust the position of the rotor.

[0158] In some embodiments, the first fault detection module 1020 is specifically configured to:

[0159] Detecting whether the electromagnetic coil current in the real-time data is within a reasonable current range;

[0160] When the electromagnetic coil current is greater than a reasonable current range, determining a first fault detection result according to the pulse width modulation signal;

[0161] When the electromagnetic coil current is less than the reasonable current range, determining a first fault detection result according to the DC voltage, the DC current, the coil current and the pulse width modulation signal;

[0162] A first fault detection result is determined according to a signal of a displacement sensor in the control system and a change in coil current.

[0163] In some embodiments, the input data processing module 1030 is specifically used to:

[0164] extracting rotor position voltage from the real-time data;

[0165] Extracting a low-frequency component from a pulse width modulation signal in the real-time data to obtain a low-frequency modulation signal;

[0166] extracting electromagnetic coil current from the real-time data;

[0167] The rotor position voltage, the low frequency modulation signal and the electromagnetic coil current are determined as input data.

[0168] In some embodiments, the fault detection model is trained as follows:

[0169] Acquire signal data of the control system of the magnetic bearing in normal operation and signal data under various fault types, and generate a training sample set;

[0170] The fault detection model is trained using the training sample set to obtain a trained fault detection model.

[0171] In some embodiments, the fault result fusion module 1050 is specifically used to:

[0172] When the first fault detection result and the second fault detection result are different, it is determined that the system fault detection result of the control system of the magnetic bearing is the first fault detection result.

[0173] The fault detection device for a magnetic bearing system provided in an embodiment of the present invention can execute the fault detection method for a magnetic bearing system provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0174] In the technical solutions of the embodiments of the present invention, the acquisition, storage and application of network traffic etc. involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0175] Fig.11 A schematic diagram of the structure of an electronic device 1100 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0176] like Fig.11 As shown, the electronic device 1100 includes at least one processor 1101, and a memory connected to the at least one processor 1101 in communication, such as a read-only memory (ROM) 1102, a random access memory (RAM) 1103, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 1101 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 1102 or the computer program loaded from the storage unit 1108 to the random access memory (RAM) 1103. In the RAM 1103, various programs and data required for the operation of the electronic device 1100 can also be stored. The processor 1101, the ROM 1102, and the RAM 1103 are connected to each other through a bus 1104. An input / output (I / O) interface 1105 is also connected to the bus 1104.

[0177] Multiple components in the electronic device 1100 are connected to the I / O interface 1105, including: an input unit 1106, such as a keyboard, a mouse, etc.; an output unit 1107, such as various types of displays, speakers, etc.; a storage unit 1108, such as a disk, an optical disk, etc.; and a communication unit 1109, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 1109 allows the electronic device 1100 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0178] The processor 1101 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 1101 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 1101 executes the various methods and processes described above, such as a distribution network line planning method.

[0179] In some embodiments, the distribution network line planning method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 1108. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 1100 via the ROM 1102 and / or the communication unit 1109. When the computer program is loaded into the RAM 1103 and executed by the processor 1101, one or more steps of the distribution network line planning method described above may be performed. Alternatively, in other embodiments, the processor 1101 may be configured to execute the distribution network line planning method in any other appropriate manner (e.g., by means of firmware).

[0180] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0181] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0182] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, device, or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0183] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0184] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0185] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS (Virtual Private Server) services.

[0186] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0187] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A fault detection method for a magnetic bearing system, characterized in that: include: Obtain real-time data of the control system of the magnetic bearing during the control process; Based on the fault diagnosis rule, determining a first fault detection result according to the real-time data; Processing the real-time data to obtain input data; Inputting the input data into a pre-trained fault detection model for processing, and outputting a second fault detection result; The first fault detection result and the second fault detection result are integrated to determine a system fault detection result of the control system of the magnetic bearing.

2. The method according to claim 1, characterized in that The magnetic bearing includes: an electromagnetic ring, a first end cover, a second end cover, a rotor, an X-axis electromagnetic coil, a Y-axis electromagnetic coil and a Z-axis electromagnetic coil; the first end cover and the second end cover are arranged on the electromagnetic ring, the first end cover and the second end cover are respectively arranged on both sides of the electromagnetic ring, and the first end cover and the second end cover are parallel; the rotor is placed inside the electromagnetic ring; the X-axis electromagnetic coil and the Y-axis electromagnetic coil are arranged on the inner side of the electromagnetic ring; the Z-axis electromagnetic coil is arranged on the inner side of the first end cover and the inner side of the second end cover.

3. The method according to claim 2, characterized in that The control system of the magnetic bearing comprises: the magnetic bearing, a controller, a power amplifier and a displacement sensor; The displacement sensor is used to obtain the actual position of the rotor; The controller obtains a position deviation between an actual position of the rotor and a desired position, and outputs a pulse width modulation signal according to the position deviation; The power amplifier is used to output a control current according to the pulse width modulation signal; The electromagnetic coil is used to convert the control current into electromagnetic thrust and adjust the position of the rotor.

4. The method according to claim 1, characterized in that: The determining the first fault detection result based on the real-time data based on the fault diagnosis rule includes: Detecting whether the electromagnetic coil current in the real-time data is within a reasonable current range; When the electromagnetic coil current is greater than a reasonable current range, determining a first fault detection result according to the pulse width modulation signal; When the electromagnetic coil current is less than the reasonable current range, determining a first fault detection result according to the DC voltage, the DC current, the coil current and the pulse width modulation signal; A first fault detection result is determined according to a signal of a displacement sensor in the control system and a change in coil current.

5. The method according to claim 1, characterized in that The processing of the real-time data to obtain input data includes: extracting rotor position voltage from the real-time data; Extracting a low-frequency component from a pulse width modulation signal in the real-time data to obtain a low-frequency modulation signal; extracting electromagnetic coil current from the real-time data; The rotor position voltage, the low frequency modulation signal and the electromagnetic coil current are determined as input data.

6. The method according to claim 5, characterized in that The fault detection model is trained as follows: Acquire signal data of the control system of the magnetic bearing in normal operation and signal data under various fault types, and generate a training sample set; The fault detection model is trained using the training sample set to obtain a trained fault detection model.

7. The method according to claim 1, characterized in that The step of fusing the first fault detection result and the second fault detection result to determine a system fault detection result of the control system of the magnetic bearing includes: When the first fault detection result and the second fault detection result are different, it is determined that the system fault detection result of the control system of the magnetic bearing is the first fault detection result.

8. A fault detection device for a magnetic bearing system, characterized in that: The device comprises: A bearing data acquisition module is used to acquire real-time data of the control system of the magnetic suspension bearing during the control process; A first fault detection module, configured to determine a first fault detection result according to the real-time data based on a fault diagnosis rule; An input data processing module, used for processing the real-time data to obtain input data; A second fault detection module, used for inputting the input data into a pre-trained fault detection model for processing, and outputting a second fault detection result; The fault result fusion module is used to fuse the first fault detection result and the second fault detection result to determine the system fault detection result of the control system of the magnetic bearing.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the fault detection method for a magnetic bearing system according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the fault detection method for a magnetic bearing system according to any one of claims 1 to 7 when executed.

Images