Fault diagnosis method and fault diagnosis system for dilution refrigerator

Abstract

The present invention provides a fault diagnosis method and a fault diagnosis system for a dilution refrigerator, which relate to the technical fields of cryogenic engineering, quantum information technology, and signal processing technology. The method includes: collecting vibration signals of multiple components of the dilution refrigerator by using a qubit sensor, the qubit sensor being arranged on the cold plate of the dilution refrigerator, and the vibration signals of each of the multiple components being transmitted to the cold plate through a connecting pipe; performing spectrum analysis on the vibration signals to determine the frequency distribution characteristics of each of the multiple components; determining a target component with abnormal vibration based on the frequency distribution characteristics of each of the multiple components; and determining a fault diagnosis result of the dilution refrigerator based on the frequency distribution characteristics of the target component and a preset fault frequency characteristic table of the target component, the fault frequency characteristic table including multiple fault types of the target component and the reference frequency characteristics respectively corresponding to the multiple fault types. Thereby, the accuracy and intelligence of the fault diagnosis of the dilution refrigerator are improved.

Description

Technical Field

[0001] The present invention relates to the technical fields of cryogenic engineering, quantum information technology, and signal processing technology, and particularly relates to a fault diagnosis method and a fault diagnosis system for a dilution refrigerator. Background Art

[0002] A dilution refrigerator is a device that uses the characteristics of a mixture of helium-3 and helium-4 to achieve extremely low-temperature refrigeration, and its lowest temperature can reach the order of mK (millikelvin). When the helium-3 and helium-4 mixture is at 0.86K, it will separate into two phases. The upper layer is called the concentrated phase, mainly composed of helium-3, and the lower layer is called the dilute phase, mainly composed of a mixture of helium-3 and helium-4. When helium-3 atoms are removed from the dilute phase, in order to maintain the balance of the two phases, the helium-3 atoms in the concentrated phase will enter the dilute phase through the phase interface to supplement the removed helium-3 atoms. This process is endothermic, and a dilution refrigerator can be made using this endothermic phenomenon.

[0003] Due to long-term operation at low temperatures, various faults may occur in the dilution refrigerator, resulting in the interruption of equipment operation. However, due to the complex components of the dilution refrigerator and the suddenness of faults, it is difficult to distinguish them with the naked eye, and there has been a lack of effective early warning measures for a long time. In addition, as a key component of a quantum computing system, for a long time, unplanned shutdown events caused by sudden faults in the dilution refrigerator during the measurement and control experiment of a quantum processor have been a pain point for researchers in related fields and need to be solved urgently. Summary of the Invention

[0004] In view of this, the present invention provides a fault diagnosis method and a fault diagnosis system for a dilution refrigerator.

[0005] In one aspect of the present invention, a fault diagnosis method for a dilution refrigerator is provided. The method includes: using a qubit sensor to collect vibration signals of multiple components of the dilution refrigerator, where the qubit sensor is arranged on the cold plate of the dilution refrigerator, and the vibration signals of each of the multiple components are transmitted to the cold plate through a connecting pipeline; performing spectral analysis on the vibration signals to determine the frequency distribution characteristics of each of the multiple components; based on the frequency distribution characteristics of each of the multiple components, determining a target component with abnormal vibration; based on the frequency distribution characteristics of the target component and a preset fault frequency characteristic table for the target component, determining a fault diagnosis result of the dilution refrigerator, where the fault frequency characteristic table includes multiple fault types of the target component and the reference frequency characteristics corresponding to each of the multiple fault types.

[0006] According to an embodiment of the present invention, the vibration signals of the multiple components include the vibration signals of a pulse tube cold head, a compressor assembly, a pump assembly, a valve assembly, and a pipeline, and each of the multiple components transmits its vibration signal to the cold plate in a different vibration direction.

[0007] According to an embodiment of the present invention, the qubit sensor includes a qubit chip integrated with resonators in three orthogonal directions of the X-axis, Y-axis, and Z-axis to collect vibration signals in multiple different directions from a pulse tube cryocooler, a compressor assembly, a pump assembly, a valve assembly, and a pipeline.

[0008] According to an embodiment of the present invention, spectrum analysis is performed on the vibration signals to determine the frequency distribution characteristics of each of the multiple components, including: performing Fourier transform or cepstrum analysis on the vibration signals to determine the initial frequency distribution characteristics of the vibration signals; obtaining the reference frequency characteristics of each of the multiple components, where the reference frequency characteristics include the reference frequency of the target component and the amplitude corresponding to the reference frequency; and using the reference frequency characteristics of each of the multiple components to perform normalization processing on the initial frequency distribution characteristics respectively to obtain the frequency distribution characteristics of each of the multiple components.

[0009] According to an embodiment of the present invention, using the reference frequency characteristics of each of the multiple components to perform normalization processing on the initial frequency distribution characteristics respectively to obtain the frequency distribution characteristics of each of the multiple components includes: using the reference frequency characteristics of each of the multiple components to perform normalization processing on the initial frequency distribution characteristics respectively to obtain the normalized frequency distribution characteristics of each of the multiple components; and determining the frequency distribution characteristics of each of the multiple components based on the reference frequencies of each of the multiple components and the frequencies that are integer multiples of the reference frequencies in the normalized frequency distribution characteristics.

[0010] According to an embodiment of the present invention, based on the frequency distribution characteristics of each of the multiple components, determining the target component with abnormal vibration includes: obtaining the working frequency characteristics of the multiple components in the normal state, where the working frequency characteristics are obtained by collecting the vibration signals of the multiple components in the normal state using the qubit sensor respectively; and determining the target component with abnormal vibration based on the frequency distribution characteristics of each of the multiple components and the working frequency characteristics of each of the multiple components.

[0011] According to an embodiment of the present invention, based on the frequency distribution characteristics of the target component and the fault frequency characteristic table of the target component, determining the fault diagnosis result of the dilution refrigerator includes: calculating the average value of the amplitudes in the frequency distribution characteristics; screening the frequency distribution characteristics of the target component based on the average value to obtain the target frequency characteristics of the target component; and determining the fault diagnosis result of the dilution refrigerator based on the target frequency characteristics of the target component and the fault frequency characteristic table of the target component.

[0012] According to an embodiment of the present invention, determining a fault diagnosis result of a dilution refrigerator based on a target frequency characteristic of a target component and a fault frequency characteristic table of the target component includes: sorting out the target frequency characteristic and the fault frequency characteristic table to obtain a target characteristic vector of the target component and reference characteristic vectors of multiple fault types respectively; determining a fault diagnosis result of the target component according to the similarity between the target characteristic vector and the reference characteristic vectors of multiple fault types respectively.

[0013] According to an embodiment of the present invention, the method further includes: determining an acquisition frequency of a vibration signal according to a reference frequency of each of multiple components; acquiring vibration signals of multiple components of the dilution refrigerator according to the acquisition frequency.

[0014] On the other hand, the present invention also provides a fault diagnosis system for a dilution refrigerator, including: a dilution refrigerator; a qubit sensor disposed on a cold plate of the dilution refrigerator;

[0015] an electronic device configured to receive vibration signals collected by the qubit sensor and execute the above method.

[0016] According to an embodiment of the present invention, based on the sensitivity of the qubit sensor to vibration, vibration signals of multiple components of the dilution refrigerator can be acquired with high sensitivity. By performing spectrum analysis to determine the frequency distribution characteristics of each of the multiple components, it is possible to preliminarily determine whether the multiple components are in a normal vibration state and identify a target component with abnormal vibration. Then, by comparing with a preset fault frequency characteristic table of the target component, it is possible to accurately and quickly determine the fault type of the target component. Thus, the accuracy and intelligence of fault diagnosis of the dilution refrigerator are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 Shows a flowchart of a fault diagnosis method according to an embodiment of the present invention;

[0018] Figure 2a Shows a spectrogram obtained by performing spectrum analysis on a vibration signal according to an embodiment of the present invention;

[0019] Figure 2b Shows a spectrogram obtained after normalizing an initial frequency distribution characteristic according to an embodiment of the present invention;

[0020] Figure 3 Shows a structural block diagram of a dilution refrigerator according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] To make the objectives, technical solutions, and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to specific embodiments and the accompanying drawings.

[0022] The terms used herein are merely for describing specific embodiments and are not intended to limit the present invention. The terms "comprising", "including" and the like used herein indicate the presence of features, steps, operations and / or components, but do not preclude the presence or addition of one or more other features, steps, operations or components.

[0023] All terms used herein, including technical and scientific terms, have the meanings commonly understood by those of ordinary skill in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.

[0024] In cases where expressions similar to "at least one of A, B, and C, etc." are used, generally, it should be interpreted according to the meaning commonly understood by those of ordinary skill in the art. For example, "a system having at least one of A, B, and C" should include, but not be limited to, a system having only A, only B, only C, having A and B, having A and C, having B and C, and / or having A, B, and C, etc. In cases where expressions similar to "at least one of A, B, or C, etc." are used, generally, it should be interpreted according to the meaning commonly understood by those of ordinary skill in the art. For example, "a system having at least one of A, B, or C" should include, but not be limited to, a system having only A, only B, only C, having A and B, having A and C, having B and C, and / or having A, B, and C, etc.

[0025] It should also be noted that the directional terms mentioned in the embodiments, such as "up", "down", "front", "back", "left", "right", etc., are only references to the directions in the drawings and are not used to limit the protection scope of the present invention. Throughout the drawings, the same elements are denoted by the same or similar reference numerals. When it may cause confusion in the understanding of the present invention, conventional structures or configurations will be omitted.

[0026] Figure 1 The flowchart of a fault diagnosis method according to an embodiment of the present invention is shown.

[0027] As Figure 1 shown, the method includes: operations S110 to S140.

[0028] In operation S110, a quantum bit sensor is used to collect vibration signals of multiple components of a dilution refrigerator. The quantum bit sensor is disposed on the cold plate of the dilution refrigerator, and the multiple components are respectively connected to the cold plate.

[0029] In operation S120, frequency spectrum analysis is performed on the vibration signals to determine the frequency distribution characteristics of each of the multiple components.

[0030] In operation S130, based on the frequency distribution characteristics of each of the multiple components, determine the target component with abnormal vibration.

[0031] In operation S140, based on the frequency distribution characteristics of the target component and the pre-set fault frequency characteristic table for the target component, determine the fault diagnosis result of the dilution refrigerator. The fault frequency characteristic table includes multiple fault types of the target component and the reference frequency characteristics corresponding to each of the multiple fault types.

[0032] A qubit sensor is a quantum sensor that uses qubits as core components. Its principle is to utilize the sensitivity of quantum states to external mechanical perturbations, and by monitoring changes in quantum states (such as energy levels, phases, or entangled states), inversely deduce vibration parameters (frequency, amplitude, direction). Since qubits are extremely sensitive to external environmental perturbations (such as mechanical vibrations) when in a superposition state, a tiny vibration will cause the energy levels of the qubits to shift or the phase to change. Therefore, using qubit sensors can achieve high-sensitivity detection of multiple components, thereby improving the diagnostic accuracy of the components. Moreover, qubits can still detect sub-nanometer displacements or micro-strains at millikelvin temperatures, avoiding the sensitivity degradation or failure of traditional sensors caused by extremely low temperatures, and having strong cryogenic compatibility.

[0033] Performing spectral analysis on the vibration signal can include preprocessing the vibration signal. The preprocessing can include operations such as denoising and filtering the vibration signal, which can eliminate interference and noise in the signal and improve the quality of the signal.

[0034] After preprocessing the vibration signal, through spectral analysis, the vibration signal can be transformed from a time-domain signal into a frequency-domain signal, and then based on the frequency-domain signal, determine the frequency distribution characteristics of each of the multiple components.

[0035] The frequency distribution characteristics include multiple frequency distributions and the amplitudes of each frequency. The frequency distributions in the frequency distribution characteristics can reflect the structural characteristics and working states of the components. Components with abnormal vibration may exhibit harmonics (including integer multiples of the reference frequency). The magnitude of the amplitude can reflect the vibration intensity of the component. An abnormal increase in the harmonic amplitude may mean that the component has a fault or is more severely worn. Based on the frequency distribution and the corresponding amplitude of the component, it can be preliminarily determined whether the component is abnormal.

[0036] The pre-set fault frequency characteristic table for the target component can be determined based on historical data and experimental data, and contains multiple fault types and their corresponding reference frequency characteristics. By comparing the frequency distribution characteristics of the target component with the reference frequency characteristics corresponding to each of the multiple fault types one by one, the fault diagnosis result of the target component can be determined.

[0037] According to an embodiment of the present invention, based on the sensitivity of qubit sensors to vibrations, vibration signals of multiple components of a dilution refrigerator can be collected with high sensitivity. By performing spectral analysis, the respective frequency distribution characteristics of the multiple components can be determined, and it can be preliminarily judged whether the multiple components are in a normal vibration state and the target component with abnormal vibration can be determined. Then, by comparing with a preset fault frequency characteristic table of the target component, the fault type of the target component can be accurately and quickly judged. Thereby, the accuracy and automation of fault diagnosis of the dilution refrigerator are improved.

[0038] Moreover, based on the vibration information measured by the qubit sensor, whose measurement accuracy can be close to the Heisenberg limit, the effect of capturing abnormal signals in advance compared with conventional sensors can be achieved, which helps to realize early warning of abnormal components. And this method can be used for diagnosis when the dilution refrigerator is in a non-stop state, and can be applied to the early warning and avoidance of the risk of unplanned shutdown of a superconducting quantum computer.

[0039] According to an embodiment of the present invention, the vibration signals of multiple components include the vibration signals of a pulse tube cold head, a compressor assembly, a pump assembly, a valve assembly, and a pipeline. The multiple components transmit their respective vibration signals to the cold plate in different vibration directions.

[0040] The qubit chip can be installed on the mixing chamber cold plate (temperature < 30 mK) of the dilution refrigerator. Since the cold plate is a key conduction path for the mechanical vibration of the dilution refrigerator, the vibrations of multiple components can be transmitted to the qubit through the cold plate. The wide-frequency response of the qubit can be used to capture the vibration information of multiple components simultaneously.

[0041] Specifically, the pulse tube refrigerator generates cooling capacity through the reciprocating compression-expansion of helium. The vibration signal of the pulse tube cold head is mainly caused by the periodic movement of the internal piston, resulting in mechanical vibration, and the axial vibration can be directly transmitted through the rigid connection between the cold head and the cold plate. The compressor assembly mainly generates high-frequency mechanical vibration through the reciprocating movement of the piston or the rotation of the scroll plate, and is transmitted to the cold plate through the support structure. The pump assembly is mainly caused by the rotation of the impeller, resulting in fluid pressure pulsation and mechanical unbalance vibration, and is transmitted to the cold plate through the pipeline connection. The valve assembly mainly generates transient impact vibration through rapid switching (such as solenoid valves or expansion valves), and is transmitted to the cold plate through the connection between the valve body and the pipeline, which may be transient multi-directional vibration. The vibration signal of the pipeline mainly comes from fluid turbulence and is transmitted to the cold plate through pipeline supports or flange connections, which may be vibrations in multiple directions.

[0042] According to an embodiment of the present invention, the qubit sensor includes a qubit chip integrated with resonators in three orthogonal directions of the X-axis, Y-axis, and Z-axis to collect multiple vibration signals in different directions from the pulse tube cold head, the compressor assembly, the pump assembly, the valve assembly, and the pipeline.

[0043] According to an embodiment of the present invention, by deploying qubit chips in three orthogonal directions of the X-axis, Y-axis, and Z-axis at a position, full-degree-of-freedom coverage can be achieved, and lateral, longitudinal, and vertical vibrations can be captured to accurately and comprehensively collect vibration signals.

[0044] In some embodiments, the qubit sensor can also adopt a planar waveguide type superconducting qubit sensor. The installation method of the superconducting qubit sensor on the cryocooler determines its sensitivity to the direction of spatial vibration signals. For example, setting the plane of the superconducting qubit sensor perpendicular to the cold plate and parallel to the working direction of the working medium gas in the pulse tube cold head, perpendicular to the cold plate and perpendicular to the working direction of the working medium gas in the pulse tube cold head, and parallel to the cold plate can respectively sense vibrations in the X-axis (lateral), Y-axis (longitudinal), and Z-axis (vertical) directions. In some other embodiments, the superconducting qubit chip can also be installed obliquely, and the tilt angle can be optimized according to the actual spatial vibration mode distribution to obtain the best signal detection effect.

[0045] According to an embodiment of the present invention, before using the sensor to collect vibration signals of multiple components of the dilution refrigerator, it may further include: determining the acquisition frequency of the vibration signals according to the respective reference frequencies of the multiple components; and collecting the vibration signals of the multiple components of the dilution refrigerator according to the acquisition frequency.

[0046] The reference frequency characterizes the natural vibration frequency that the component has in the normal working state. The acquisition frequency characterizes the sampling rate used by the sensor when collecting vibration signals, that is, the number of vibration data points collected per second. Since the selection of the acquisition frequency can directly affect the resolution and accuracy of the vibration signals, the acquisition frequency should be high enough to ensure that all frequency components in the vibration signals can be accurately captured.

[0047] When using qubits as sensors to collect vibration signals, the sampling frequency can be determined according to the maximum reference frequency among the multiple components.

[0048] Preferably, the acquisition frequency is not less than 20 times the reference frequency. For example, if the reference working frequency of the working medium gas reciprocatingly doing work in the pulse tube cold head of a typical pulse tube cryocooler is 1.4 Hz, the upper limit of the acquisition frequency of its vibration signals should not be less than 28 Hz. The present invention is not limited thereto, and the multiple relationship between the acquisition frequency and the reference frequency can also be adjusted according to specific situations and requirements.

[0049] According to an embodiment of the present invention, by determining the acquisition frequency according to the reference frequency, it can be ensured that the sensor can accurately capture the vibration signals of the components, improving the accuracy and reliability of the signals. The acquisition frequency not being less than 20 times the reference frequency can ensure that all frequency components can be collected to improve the integrity of the information.

[0050] According to an embodiment of the present invention, spectral analysis is performed on the vibration signal to determine the frequency distribution characteristics of each of multiple components, including: performing Fourier transform or cepstrum analysis on the vibration signal to determine the initial frequency distribution characteristics of the vibration signal; obtaining the reference frequency characteristics of each of the multiple components, where the reference frequency characteristics include the reference frequency of the target component and the amplitude corresponding to the reference frequency; and using the reference frequency characteristics of each of the multiple components to perform normalization processing on the initial frequency distribution characteristics respectively to obtain the frequency distribution characteristics of each of the multiple components.

[0051] In some embodiments, performing spectral analysis on the vibration signal in a preset time period to determine the initial frequency distribution characteristics of the vibration signal may include: performing discrete Fourier transform processing on the vibration signal to obtain the initial frequency distribution characteristics of the vibration signal. Alternatively, performing cepstrum processing on the vibration signal to obtain the initial frequency distribution characteristics of the vibration signal.

[0052] Specifically, when the preprocessing of the vibration signal is simple, the real-time requirement is high, or the characteristics in the vibration signal (such as resonance frequency, fault frequency, etc.) are obvious in the frequency domain and are easy to identify and extract, discrete Fourier transform can be selected for use. Through discrete Fourier transform, the vibration signal can be converted from the time domain to the frequency domain, so that the amplitude and phase information of different frequency components in the signal can be intuitively observed.

[0053] When it is necessary to analyze the periodic structure of the vibration signal, especially when the signal contains multiple harmonic components, or there are noise, interference or nonlinear distortion in the vibration signal, and these components are not easy to directly remove in the frequency domain, or the original signal is a non-stationary signal, and the spectrogram may contain complex periodic structures and is difficult to directly identify, techniques such as cepstrum, filtering, Hilbert-Huang transform, time-dependent Fourier transform, etc. can be used to process and analyze the vibration signal in the preset time period to extract frequency characteristics helpful for diagnosis.

[0054] In some embodiments, the combined characteristics of the reference frequency and the derived frequency under the normal working condition of the component can be obtained through theoretical calculation, experimental measurement or consulting relevant technical materials. It is also possible to determine the reference frequency characteristics by measuring the vibration signal when the component is in a normal operating state. Since the reference frequency of each mechanical component is determined by its inherent physical parameters (such as the number of gear teeth, the number of bearing balls, the rotational speed), the combined characteristics of the reference frequency and the derived frequency under the normal working condition of the component are unique.

[0055] The reference frequency and the amplitude corresponding to the reference frequency can be used as normalization factors. Applying the normalization factors to each frequency and amplitude in the initial frequency distribution characteristics, after normalizing the initial frequency distribution characteristics, the amplitude and frequency of the reference frequency are standardized to 1 (or a fixed reference value), and other derived frequency components are scaled proportionally. Thus, the quantization analysis scale of the vibration signals of different components is unified.

[0056] Since the reference frequency of each component under normal operating conditions is unique, after normalizing the initial frequency distribution characteristics using the reference frequency of the target component, the frequency distribution characteristics of each component occupy independent positions in their respective frequency domains. Thus, the frequency distribution characteristics of different components can be distinguished, and further, the frequency distribution characteristics of the target component can be distinguished from the initial frequency distribution characteristics.

[0057] According to an embodiment of the present invention, by normalizing the initial frequency distribution characteristics using the reference frequency of the target component, the uniqueness of the reference frequency can be utilized to avoid spectral confusion of different components, thereby improving the accuracy of fault diagnosis of the target component.

[0058] According to an embodiment of the present invention, using the respective reference frequency characteristics of multiple components to normalize the initial frequency distribution characteristics respectively to obtain the respective frequency distribution characteristics of the multiple components, including: using the respective reference frequency characteristics of multiple components to normalize the initial frequency distribution characteristics respectively to obtain the respective normalized frequency distribution characteristics of each component; based on the reference frequency and the frequencies that are integer multiples of the reference frequency in the normalized frequency distribution characteristics of each component, determining the respective frequency distribution characteristics of each component.

[0059] By normalizing the initial frequency distribution characteristics using the respective reference frequencies of multiple components, since the reference frequencies of different components are different, the obtained normalized frequency distribution characteristics of each component occupy independent positions in the spectrogram, and the frequency distribution characteristics of each component can be distinguished.

[0060] Since local damage of mechanical components will generate periodic impacts, which are manifested as harmonics of the reference frequency in the spectrogram. By using the frequency distribution characteristics composed of the reference frequency and the frequencies that are integer multiples of the reference frequency, the fault type of the target component can be accurately reflected.

[0061] Figure 2a The figure shows a spectrogram obtained by performing spectral analysis on a vibration signal according to an embodiment of the present invention.

[0062] As Figure 2a shown, the frequency distribution included in the vibration signal and the qubit flux noise spectral density (i.e., amplitude) of each frequency can be read from the spectrogram, and thus the initial frequency distribution characteristics of the vibration signal can be obtained. For example, there are multiple characteristic peaks in the range of 0 - 1000 Hz and one characteristic peak at 2000 Hz.

[0063] Figure 2b The figure shows a spectrogram obtained after normalizing the initial frequency distribution characteristics according to an embodiment of the present invention.

[0064] AsFigure 2b As shown, multiple frequencies in the initial frequency distribution characteristics are normalized using the reference frequency f of the target component, and the amplitudes corresponding to the multiple frequencies are used to normalize the amplitudes of the multiple frequencies respectively, obtaining a normalized spectrogram. Based on the spectrogram, the frequency distribution characteristics of the target component can be obtained.

[0065] Specifically, according to Figure 2b the frequencies that are integer multiples of the reference frequency of the target component and the corresponding amplitudes can be read out. For example, the amplitude of the reference frequency f is 1, the amplitude of 2f is 0.87, the amplitude of 3f is 0.58, the amplitude of 4f is 0.79, the amplitude of 5f is 0.35, the amplitude of 6f is 0.53, the amplitude of 7f is 0.35, the amplitude of 8f is 0.38, the amplitude of 9f is 0.58, the amplitude of 10f is 0.41, etc. Thus, the frequency distribution characteristics of the vibration signal can be obtained.

[0066] According to an embodiment of the present invention, based on the frequency distribution characteristics of multiple components respectively, the target component with abnormal vibration is determined, including: obtaining the working frequency characteristics of multiple components in the normal state, where the working frequency characteristics are obtained by respectively collecting the vibration signals of multiple components in the normal state using a quantum bit sensor; and determining the target component with abnormal vibration based on the frequency distribution characteristics of multiple components respectively and the working frequency characteristics of multiple components respectively.

[0067] First, the quantum bit sensor can be used to respectively collect the working frequency characteristics of multiple components in the normal state, and a working frequency characteristic library of multiple components can be constructed based on the working frequency characteristics of multiple components respectively. After determining the frequency distribution characteristics of multiple components respectively, by calling the working frequency characteristics of multiple components in the working frequency characteristic library and comparing them with the frequency distribution characteristics of multiple components, when there is a certain component whose working frequency characteristics and frequency distribution characteristics are inconsistent, that component is determined as the target component with abnormal vibration.

[0068] According to an embodiment of the present invention, based on the frequency distribution characteristics of the target component and the fault frequency characteristic table of the target component, the fault diagnosis result of the dilution refrigerator is determined, including: calculating the average value of the amplitudes in the frequency distribution characteristics; screening the frequency distribution characteristics of the target component based on the average value to obtain the target frequency characteristics of the target component; and determining the fault diagnosis result of the dilution refrigerator based on the target frequency characteristics of the target component and the fault frequency characteristic table of the target component.

[0069] In an embodiment of the present invention, since the signal contains multiple frequency components, but only a few of them play a dominant role in the characteristics or behavior of the vibration signal. By calculating the average value and screening based on this, it is easier to identify the dominant frequency components and use them as the target frequency characteristics.

[0070] In some embodiments, the screening criteria can be set according to the average value. For example, frequencies with amplitudes several times higher than the average value can be selected, which can be 3 times, 5 times, 10 times higher than the average value, and so on. The present invention is not limited thereto, and frequencies within a certain ratio (such as the average value plus or minus a certain standard deviation) of the amplitude can also be selected. The screening criteria can be non-fixed and can be dynamically adjusted according to specific requirements or data characteristics, such as setting according to the distribution of data, the presence of outliers, or specific application scenarios.

[0071] According to an embodiment of the present invention, based on the average value of the amplitudes of multiple frequencies, the dominant frequency components can be screened out from multiple frequencies, thereby improving the accuracy and reliability of fault diagnosis.

[0072] According to an embodiment of the present invention, based on the target frequency characteristics of the target component and the fault frequency characteristics table of the target component, the fault diagnosis result of the dilution refrigerator is determined, including: sorting the target frequency characteristics and the fault frequency characteristics table to obtain the target feature vector of the target component and the reference feature vectors of each of the multiple fault types respectively; determining the fault diagnosis result of the target component according to the similarity between the target feature vector and the reference feature vectors of each of the multiple fault types.

[0073] Based on the differences in components, different fault frequency characteristics tables can be formulated to adapt to different devices and different working conditions. Table 1 and Table 2 respectively schematically show the fault frequency characteristics tables of two components.

[0074] Table 1:

[0075]

[0076] Table 2:

[0077]

[0078] Sorting the target frequency characteristics and the fault frequency characteristics table can be to sort multiple target frequencies according to the magnitude of the frequencies to obtain a two-dimensional matrix composed of multiple target frequencies and their respective amplitudes, thereby obtaining the target feature vector.

[0079] Similarly, in the same manner as the target frequency characteristics, the reference frequency characteristics are also sorted according to the magnitude of the frequencies to obtain the reference feature vectors of each of the multiple fault types.

[0080] According to the similarity between the target feature vector and the reference feature vectors of each of the multiple fault types, such as cosine similarity, Euclidean distance, correlation coefficient, etc., the fault diagnosis result of the target component is determined.

[0081] In some embodiments, a threshold value of similarity can be preset. If the similarity is higher than this threshold value, the faults corresponding to the diagnostic table are output in a preset format.

[0082] According to an embodiment of the present invention, based on the similarity between the target feature vector and the reference feature vectors of multiple fault types, the fault diagnosis result of the target component is determined, which can accurately and efficiently detect potential faults at an early stage, thereby avoiding equipment operation interruption and improving the reliability and stability of the equipment.

[0083] Figure 3 The structural block diagram of a dilution refrigerator according to an embodiment of the present invention is shown.

[0084] The present invention also provides a fault diagnosis system for a dilution refrigerator, including a dilution refrigerator; a qubit sensor disposed on the cold plate of the dilution refrigerator; and an electronic device for receiving the vibration signals collected by the qubit sensor and executing the above-mentioned fault diagnosis method.

[0085] In some embodiments, as Figure 3 shown, the dilution refrigerator includes a core unit 310, a pulse tube cooler 320, and a gas control system 330. The gas control system 330 includes multiple components such as a mechanical pump, a molecular pump, a compressor, and a pneumatic valve.

[0086] The core unit 310 of the dilution refrigerator includes the cold plate 311 of the dilution refrigerator and a qubit sensor 312. The qubit sensor 312 is disposed on the cold plate 311 of the mixing chamber of the dilution refrigerator. The qubit sensor 312 can collect the vibration signals of multiple components such as the pulse tube cooler 320, the mechanical pump, the molecular pump, the compressor, the pneumatic valve, and the pipeline. And transmit the vibration signals to the electronic device, and use the electronic device to process the vibration signals of multiple components and perform fault diagnosis.

[0087] In some embodiments, the electronic device may include one or more processors and a memory. The memory is used to store one or more computer programs, and one or more processors execute one or more computer programs to implement the above-mentioned fault diagnosis method.

[0088] The above describes the embodiments of the present invention. However, these embodiments are only for illustrative purposes and are not intended to limit the scope of the present invention. Although the embodiments are described separately above, this does not mean that the measures in each embodiment cannot be used advantageously in combination. The scope of the present invention is defined by the appended claims and their equivalents. Without departing from the scope of the present invention, those skilled in the art can make various substitutions and modifications, and these substitutions and modifications should all fall within the scope of the present invention.

Claims

1. A dilution refrigerator fault diagnosis method, characterized in that: The method comprises: Using a quantum bit sensor to collect vibration signals of multiple components of a dilution refrigerator, wherein the quantum bit sensor is disposed on a cold plate of the dilution refrigerator, and multiple components are connected to the cold plate respectively; Performing spectrum analysis on the vibration signal to determine frequency distribution characteristics of each of the plurality of components; determining a target component of abnormal vibration based on the frequency distribution characteristics of each of the plurality of components; The fault diagnosis result of the dilution refrigerator is determined based on the frequency distribution characteristics of the target component and a preset fault frequency characteristic table of the target component, wherein the fault frequency characteristic table includes multiple fault types of the target component and reference frequency characteristics corresponding to each of the multiple fault types.

2. The method according to claim 1, characterized in that The vibration signals of the plurality of components include vibration signals of a pulse tube cold head, a compressor assembly, a pump assembly, a valve assembly and a pipeline, and the plurality of components transmit their respective vibration signals to the cold plate in different vibration directions.

3. The method according to claim 2, characterized in that The quantum bit sensor includes a quantum bit chip that integrates resonant cavities in three orthogonal directions of X-axis, Y-axis and Z-axis to collect vibration signals in multiple different directions from the pulse tube cold head, compressor assembly, pump assembly, valve assembly and pipeline.

4. The method according to claim 1, characterized in that: The performing spectrum analysis on the vibration signal to determine the frequency distribution characteristics of each of the plurality of components comprises: Performing Fourier transform or cepstrum analysis on the vibration signal to determine initial frequency distribution characteristics of the vibration signal; Acquire reference frequency characteristics of each of the plurality of components, wherein the reference frequency characteristics include a reference frequency and an amplitude corresponding to the reference frequency; The initial frequency distribution characteristics are respectively normalized using the respective reference frequency characteristics of the plurality of components to obtain the respective frequency distribution characteristics of the plurality of components.

5. The method according to claim 4, characterized in that The method of using the respective reference frequency characteristics of the plurality of components to respectively normalize the initial frequency distribution characteristics to obtain the respective frequency distribution characteristics of the plurality of components includes: Using the respective reference frequency characteristics of the plurality of components, respectively normalizing the initial frequency distribution characteristics to obtain respective normalized frequency distribution characteristics of the plurality of components; Based on the reference frequencies of the plurality of components and the frequencies of integer multiples of the reference frequencies in the normalized frequency distribution characteristics, the frequency distribution characteristics of the plurality of components are determined.

6. The method according to claim 1, characterized in that The step of determining a target component with abnormal vibration based on the frequency distribution characteristics of each of the plurality of components comprises: Acquiring operating frequency characteristics of multiple components in a normal state, wherein the operating frequency characteristics are obtained by respectively collecting vibration signals of multiple components in a normal state using the quantum bit sensor; Based on the frequency distribution characteristics of each of the plurality of components and the operating frequency characteristics of each of the plurality of components, a target component of abnormal vibration is determined.

7. The method according to claim 1, characterized in that The method of determining the fault diagnosis result of the dilution refrigerator based on the frequency distribution characteristics of the target component and a preset fault frequency characteristic table of the target component includes: Calculating the average value of the amplitude in the frequency distribution feature; Based on the average value, the frequency distribution characteristics of the target component are screened to obtain the target frequency characteristics of the target component; and A fault diagnosis result of the dilution refrigerator is determined based on a target frequency characteristic of a target component and a fault frequency characteristic table of the target component.

8. The method according to claim 7, characterized in that The step of determining the fault diagnosis result of the dilution refrigerator based on the target frequency characteristic of the target component and the fault frequency characteristic table of the target component comprises: Arrange the target frequency characteristics and the fault frequency characteristics table to obtain the target characteristic vector of the target component and the reference characteristic vectors of each of the multiple fault types; The fault diagnosis result of the target component is determined according to the similarities between the target feature vector and the reference feature vectors of multiple fault types.

9. The method according to claim 1, characterized in that: The method further comprises: Determine the collection frequency of the vibration signal according to the respective reference frequencies of the multiple components; According to the acquisition frequency, vibration signals of multiple components of the dilution refrigerator are acquired.

10. A fault diagnosis system for a dilution refrigerator, characterized in that: include: Dilution refrigerator; A quantum bit sensor is arranged on a cold plate of the dilution refrigerator; An electronic device for receiving the vibration signal collected by the quantum bit sensor and executing the method according to any one of claims 1 to 9.

Images