Quantum sensor multi-detector array fault diagnosis method combined with spatial topology

Abstract

The embodiment of the invention provides a quantum sensor multi-detector array fault diagnosis method in combination with spatial topology. The method comprises the steps that original data of a marked fault type of a quantum sensor multi-detector array are acquired, and the original data comprise time sequence signal data, space topology data and environment data; preprocessing the original data, and constructing a fault diagnosis data set; training the CatBoost model by using the fault diagnosis data set, and performing hyper-parameter optimization on the CatBoost model by using an improved sparrow search algorithm fused with a genetic algorithm, a particle swarm optimization algorithm and a differential evolution algorithm in the training process; and performing fault diagnosis on the quantum sensor multi-detector array by using the trained CatBoost model. In this way, high-precision and real-time fault diagnosis can be achieved, and the fault diagnosis effect is improved.

Description

Technical Field

[0001] This disclosure relates to the field of quantum technology, and in particular to a fault diagnosis method for a quantum sensor multi-detector array that incorporates spatial topology. Background Technology

[0002] With the rapid development of quantum technology, quantum sensors, due to their high sensitivity, high precision, and broad-spectrum response characteristics, have been widely used in quantum communication, quantum computing, and precision measurement. Especially in multi-detector array systems, quantum sensors, through collaborative work, can achieve high-resolution signal acquisition and processing in complex environments. However, current multi-detector arrays of quantum sensors face the challenge of difficult fault diagnosis in practical operation, making it difficult to effectively diagnose whether a fault has occurred in the array. Summary of the Invention

[0003] In a first aspect, embodiments of this disclosure provide a fault diagnosis method for a quantum sensor multi-detector array incorporating spatial topology, the method comprising:

[0004] Acquire raw data of a quantum sensor multi-detector array with labeled fault types. The raw data includes time-series signal data, spatial topology data, and environmental data.

[0005] The raw data is preprocessed to construct a fault diagnosis dataset;

[0006] The CatBoost model was trained using a fault diagnosis dataset, and an improved sparrow search algorithm that integrates genetic algorithm, particle swarm optimization algorithm and differential evolution algorithm was used to optimize the hyperparameters of the CatBoost model during the training process.

[0007] Fault diagnosis of quantum sensor multi-detector arrays was performed using a trained CatBoost model.

[0008] In some possible implementations of the first aspect, raw data labeled with fault types for a quantum sensor multi-detector array is obtained, including:

[0009] Time-series signal data is obtained by acquiring signals from each detector unit in the quantum sensor multi-detector array at a fixed sampling rate using a high-frequency sampling device.

[0010] Based on the physical layout and signal interaction relationship of the quantum sensor multi-detector array, an adjacency matrix is ​​constructed to obtain spatial topology data;

[0011] Collect environmental data at the location of the quantum sensor multi-detector array;

[0012] By summarizing and analyzing time-series signal data, spatial topology data, and environmental data, we obtained raw data of quantum sensor multi-detector array with labeled fault types.

[0013] In some possible implementations of the first aspect, the raw data is preprocessed to construct a fault diagnosis dataset, including:

[0014] Denoising and feature extraction are performed on time-series signal data to obtain signal features;

[0015] A dynamic topology graph is constructed based on spatial topology data, and a graph neural network (GNN) model is used to extract features from the dynamic topology graph to obtain spatiotemporal features.

[0016] Environmental features are obtained by extracting features from environmental data;

[0017] By splicing together signal features, spatiotemporal features, and environmental features, a comprehensive feature is obtained.

[0018] Using comprehensive features as sample features and their corresponding fault types as sample labels, samples are constructed, and an initial dataset is built based on this.

[0019] The initial dataset is augmented with a small number of fault type samples to obtain a fault diagnosis dataset.

[0020] In some possible implementations of the first aspect, the CatBoost model is trained using a fault diagnosis dataset, and during training, an improved sparrow search algorithm that integrates genetic algorithms, particle swarm optimization algorithms, and differential evolution algorithms is used to optimize the hyperparameters of the CatBoost model, including:

[0021] The fault diagnosis dataset is divided into a training set and a validation set according to a preset ratio. The CatBoost model is trained using the training set, and the performance of the CatBoost model is evaluated using K-fold cross-validation. The improved sparrow search algorithm, which integrates genetic algorithm, particle swarm optimization algorithm, and differential evolution algorithm, is used to optimize the hyperparameters of the CatBoost model.

[0022] Among the possible implementations of the first aspect, an improved sparrow search algorithm that integrates genetic algorithms, particle swarm optimization algorithms, and differential evolution algorithms includes:

[0023] For explorers, the crossover operation of the genetic algorithm is combined to enhance the explorer's global search capability in the hyperparameter space;

[0024] For followers, a speed update mechanism of particle swarm optimization algorithm is introduced to accelerate the followers toward a better solution;

[0025] For the vigilant, the random walk characteristic of the vigilant is preserved, and the mutation operation of the differential evolution algorithm is used to avoid getting trapped in local optima;

[0026] The optimization objective is set to maximize the fault diagnosis accuracy of the CatBoost model on the validation set.

[0027] Among the possible implementations of the first aspect, genetic algorithms include:

[0028] Selection: The roulette wheel selection method is used to select high-quality individuals based on their fitness. Here, an individual is a combination of hyperparameters of the CatBoost model, and the fitness of an individual is the fault diagnosis accuracy of the CatBoost model on the validation set when the individual is selected.

[0029] Crossover: A single-point crossover occurs with a crossover probability of 0.8, generating a new individual;

[0030] Mutation: Individuals are randomly perturbed with a mutation probability of 0.1 to increase population diversity.

[0031] Among some possible implementations of the first aspect, particle swarm optimization algorithms include:

[0032] The position and velocity of an individual are updated based on its historical best position and global best position, where the individual is a combination of hyperparameters of the CatBoost model.

[0033] Among the possible implementations of the first aspect, differential evolution algorithms include:

[0034] The individual's position is updated based on the positions of multiple randomly selected individuals.

[0035] Among some possible implementations of the first aspect, the method also includes:

[0036] The fault diagnosis results of the quantum sensor multi-detector array output by the trained CatBoost model are compared with the actual operating status of the quantum sensor multi-detector array. False alarms and / or missed alarms are recorded, and the parameters of the CatBoost model are updated accordingly.

[0037] Among some possible implementations of the first aspect, the method also includes:

[0038] The weights of the GNN model and the hyperparameters of the CatBoost model are updated online based on real-time monitored environmental parameters.

[0039] Secondly, embodiments of this disclosure provide a fault diagnosis device for a quantum sensor multi-detector array incorporating spatial topology, the device comprising:

[0040] The acquisition module is used to acquire raw data of the quantum sensor multi-detector array with labeled fault types. The raw data includes time-series signal data, spatial topology data, and environmental data.

[0041] The building module is used to preprocess the raw data and build a fault diagnosis dataset;

[0042] The training module is used to train the CatBoost model using a fault diagnosis dataset, and during the training process, it uses an improved sparrow search algorithm that integrates genetic algorithm, particle swarm optimization algorithm, and differential evolution algorithm to optimize the hyperparameters of the CatBoost model.

[0043] The diagnostic module is used to perform fault diagnosis on the quantum sensor multi-detector array using a trained CatBoost model.

[0044] Thirdly, embodiments of this disclosure provide an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; the memory storing instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the method described above.

[0045] Fourthly, embodiments of this disclosure provide a non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the methods described above.

[0046] Compared with the prior art, this disclosure has at least the following technical effects:

[0047] (1) By integrating the GNN model and the CatBoost model, this disclosure can capture the dynamic spatiotemporal dependence of quantum sensor multi-detector arrays and realize the early prediction of potential faults.

[0048] (2) An adaptive learning mechanism is introduced, which dynamically adapts to complex and non-stationary environments by updating the weights of the GNN model and the hyperparameters of the CatBoost model online, so as to ensure the robustness and adaptability of diagnostic performance in dynamic environments.

[0049] (3) Effectively integrate multimodal data (time series signal data, spatial topology data, and environmental data), extract the topological relationship between sensors through the GNN model, and process high-dimensional features through the CatBoost model to achieve comprehensive fault mode analysis and improve the accuracy of fault prediction.

[0050] (4) The improved sparrow search algorithm, which integrates genetic algorithm, particle swarm optimization algorithm and differential evolution algorithm, is used for hyperparameter optimization, which can shorten the hyperparameter search time and improve the optimization accuracy.

[0051] (5) By using model pruning, quantization techniques and efficient optimization strategies, the response speed and operating efficiency of the model have been improved, which can meet the high efficiency requirements of real-time status assessment and fault diagnosis of quantum sensor multi-detector arrays.

[0052] (6) Through dynamic prediction, online learning and multi-source data fusion, the model can identify potential faults in a timely manner and provide early warnings, significantly improving the operational reliability of quantum sensor arrays.

[0053] It should be understood that the description in the Summary of the Invention is not intended to limit the key or essential features of the embodiments of this disclosure, nor is it intended to restrict the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0054] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. The drawings are provided for a better understanding of the invention and are not intended to limit the scope of this disclosure. In the drawings, the same or similar reference numerals denote the same or similar elements, wherein:

[0055] Figure 1 A flowchart illustrating a fault diagnosis method for a quantum sensor multi-detector array combining spatial topology, provided by an embodiment of this disclosure, is shown.

[0056] Figure 2 A structural diagram of a fault diagnosis device for a quantum sensor multi-detector array combining spatial topology, provided by an embodiment of this disclosure, is shown.

[0057] Figure 3 A structural diagram of an exemplary electronic device capable of implementing embodiments of the present disclosure is shown. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0059] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0060] To address the problems in the background art, embodiments of this disclosure provide a method, apparatus, device, and storage medium for fault diagnosis of a quantum sensor multi-detector array incorporating spatial topology. Specifically, the method involves acquiring raw data of the quantum sensor multi-detector array labeled with fault types, including time-series signal data, spatial topology data, and environmental data; preprocessing the raw data to construct a fault diagnosis dataset; training a CatBoost model using the fault diagnosis dataset, and optimizing the hyperparameters of the CatBoost model during training using an improved sparrow search algorithm that integrates genetic algorithms, particle swarm optimization algorithms, and differential evolution algorithms; and using the trained CatBoost model to perform fault diagnosis on the quantum sensor multi-detector array. This approach enables high-precision, real-time fault diagnosis, improving the effectiveness of fault diagnosis.

[0061] The following detailed description, with reference to the accompanying drawings and specific embodiments, illustrates a method, apparatus, device, and storage medium for fault diagnosis of a quantum sensor multi-detector array incorporating spatial topology, provided by the present disclosure.

[0062] Figure 1 A flowchart illustrating a fault diagnosis method for a quantum sensor multi-detector array combining spatial topology, as provided in an embodiment of this disclosure, is shown. Figure 1 As shown, method 100 may include the following steps:

[0063] S110: Acquire raw data of the quantum sensor multi-detector array with labeled fault types. The raw data includes time-series signal data, spatial topology data, and environmental data.

[0064] In some embodiments, signals (such as voltage, current, and phase) can be acquired from each detection unit in the quantum sensor multi-detector array at a fixed sampling rate (e.g., 10 kHz) using a high-frequency sampling device to obtain time-series signal data. Based on the physical layout and signal interaction relationships of the quantum sensor multi-detector array, an adjacency matrix is ​​constructed to represent the connection strength between sensor nodes, thereby obtaining spatial topology data. Environmental data (such as temperature, electromagnetic field strength, and humidity) at the location of the quantum sensor multi-detector array are collected. The time-series signal data, spatial topology data, and environmental data are summarized and analyzed to obtain raw data of the quantum sensor multi-detector array labeled with fault types (such as normal, sensor failure, and signal drift).

[0065] As an example, the various data mentioned above can be acquired in real time through embedded systems or high-performance computing platforms (such as NVIDIA Jetson), with an update frequency of 1 second.

[0066] S120: Preprocess the raw data to construct a fault diagnosis dataset.

[0067] In some embodiments, signal denoising and feature extraction can be performed on the time-series signal data in the original data to obtain signal features; a dynamic topology map can be constructed based on the spatial topology data in the original data, and a GNN model can be used to extract features from the dynamic topology map to obtain spatiotemporal features; features can be extracted from the environmental data in the original data to obtain environmental features; the extracted signal features, spatiotemporal features, and environmental features can be concatenated to obtain comprehensive features; samples can be constructed using the comprehensive features as sample features and their corresponding fault types as sample labels, and an initial dataset can be constructed based on this; to ensure the class balance of the dataset, a few fault type samples can be augmented in the initial dataset to obtain a fault diagnosis dataset.

[0068] For time-series signal data, signal denoising can be achieved using wavelet transform algorithms, specifically the Daubechies wavelet basis (db4) with a 5-layer decomposition. High-frequency noise is filtered out using a soft thresholding method, preserving the main signal features. Feature extraction mainly includes extracting time-domain features (such as mean, variance, and peak value), frequency-domain features (such as dominant frequency and spectral energy), and statistical features (such as skewness and kurtosis). To reduce data dimensionality, Principal Component Analysis (PCA) is used to map high-dimensional features to a low-dimensional space, retaining 95% of the variance contribution rate, resulting in signal features suitable for the CatBoost model.

[0069] The dynamic topology graph constructed here based on spatial topology data is represented by G(V,E,V), where:

[0070] V: Node set, representing sensor units (such as 16 superconducting quantum interference devices).

[0071] E: Edge set, representing the signal interaction strength between sensors (calculated based on Pearson correlation coefficient).

[0072] A: Adjacency matrix, dynamically updated (edge ​​weights are recalculated every second based on the latest signal).

[0073] The change of G(V,E,A) over time reflects the dynamic spatiotemporal dependence of the quantum sensor multi-detector array.

[0074] The GNN model described above is pre-built based on the GNN algorithm. Its input is a dynamic topological graph G(V,E,A). It extracts the spatiotemporal features between sensor nodes through graph convolution operations, which can be expressed by the following formula:

[0075]

[0076] Among them, H (l) This represents the feature matrix of the nodes in the l-th layer. W represents the normalized adjacency matrix (including self-loops); (l) σ represents the weight matrix of the l-th layer; σ represents the activation function (such as ReLU).

[0077] For example, the number of GNN layers can be set to 2, and the output dimension can be set to 64 dimensions, which means the spatiotemporal feature dimension is 64 dimensions. If the signal feature dimension is 20 dimensions and the environmental feature dimension is 5 dimensions, then the comprehensive feature obtained by concatenating the signal feature, spatiotemporal feature, and environmental feature will have a dimension of 89 dimensions.

[0078] It's important to understand that the aforementioned sample augmentation primarily addresses the scarcity of fault samples in multi-detector arrays of quantum sensors. It employs the Synthetic Minority Over-sampling Technique (SMOTE) algorithm to generate synthetic fault samples. Specifically, based on the K-nearest neighbor algorithm (K=5), new samples are generated by interpolation from a few fault types (such as sensor failure or signal drift), ensuring class balance in the dataset and thus improving the model's training performance in scenarios with limited sample sizes.

[0079] S130 uses a fault diagnosis dataset to train the CatBoost model, and during the training process, an improved sparrow search algorithm that integrates genetic algorithm, particle swarm optimization algorithm and differential evolution algorithm is used to optimize the hyperparameters of the CatBoost model.

[0080] In some embodiments, the fault diagnosis dataset can be divided into a training set and a validation set according to a preset ratio (e.g., 8:2). The CatBoost model is trained using the training set, and the performance of the CatBoost model is evaluated using K-fold (e.g., 5-fold) cross-validation. An improved sparrow search algorithm that integrates genetic algorithm, particle swarm optimization algorithm, and differential evolution algorithm is used to optimize the hyperparameters of the CatBoost model. In each iteration, the optimized hyperparameter combination is used to update the model parameters and minimize the model loss.

[0081] It is worth noting that the CatBoost model here is pre-built based on the CatBoost algorithm, which is a machine learning algorithm based on Gradient Boosting Decision Tree (GBDT). It enhances the model's ability to classify complex features through automatic encoding of class features and ordered boosting, and is particularly suitable for handling high-dimensional nonlinear features.

[0082] When constructing the CatBoost model, its input is set as comprehensive features, and its output is set as the fault category and its corresponding probability. The hyperparameters of the CatBoost model include the learning rate, tree depth, number of iterations, and L2 regularization coefficient (l2_leaf_reg). Here, the initial hyperparameters are set as follows: learning rate = 0.1, depth = 6, iterations = 1000, l2_leaf_reg = 3.0, and subsequently adjusted through optimization strategies. Optionally, the multiclass cross-entropy loss function is used as the objective function for optimizing the CatBoost model parameters, and its formula is shown below:

[0083]

[0084] Where L represents the loss value, N represents the number of samples, M represents the number of fault categories, and y ij Indicates a label, p ij This represents the predicted probability of the fault category.

[0085] For example, the above-mentioned use of the improved sparrow search algorithm, which integrates genetic algorithm, particle swarm optimization algorithm, and differential evolution algorithm, to optimize the hyperparameters of the CatBoost model may include:

[0086] Multiple sets of CatBoost model hyperparameter combinations are randomly generated. Each set of CatBoost model hyperparameter combinations is used as an individual to construct the initial population. The population size is set to 50, and the hyperparameter range is set to:

[0087] learning_rate∈[0.01,0.5];

[0088] depth∈[4,10];

[0089] iterations∈[500,2000];

[0090] l2_leaf_reg∈[1,10].

[0091] Genetic algorithms are used to enhance the global search capability of the population. Specific operations include:

[0092] Selection: The roulette wheel selection method is used to select high-quality individuals based on their fitness (i.e., the fault diagnosis accuracy of the CatBoost model on the validation set when the individual is adopted);

[0093] Crossover: A single-point crossover occurs with a crossover probability of 0.8, generating a new individual;

[0094] Mutation: Randomly perturb individuals with a mutation probability of 0.1 to increase population diversity.

[0095] Particle Swarm Optimization (PSO) is introduced to accelerate the population's iteration towards a high-quality solution through its fast convergence characteristic. Specifically, the individual's position and velocity are updated based on its historical best position (pbest) and global best position (gbest). This process can be represented by the following formula:

[0096] v i (t+1)=w·v i (t)+c1·r1·(pbest i -x i (t))+c2·r2·(gbest-x i (t))

[0097] Among them, v i (t) represents the velocity vector of the i-th individual, x i (t) represents the current position vector of the i-th individual, w = 0.5 is the inertia weight, c1 = c2 = 1.5 is the learning factor, and r1, r2 ∈ [0, 1] are random numbers.

[0098] Differential evolution algorithm is introduced to enhance the local search capability of hyperparameters. Specifically, the position of an individual is updated based on the positions of multiple randomly selected individuals using the following formula:

[0099] x i (t+1)=x r1 (t)+F·(x r2 (t)-x r3 (t))

[0100] Where, x i (t+1) represents the next-generation position of the i-th individual, x r1 (t),x r2 (t),x r3 (t) represents the randomly selected individual position, and F = 0.5 is the scaling factor.

[0101] The explorer, follower, and watchdog mechanism of the sparrow search algorithm is integrated with the global search capability of the genetic algorithm, the fast iterative characteristics of the particle swarm optimization algorithm, and the local search capability of the differential evolution algorithm to form a hybrid optimization strategy. This strategy optimizes the efficiency and accuracy of hyperparameter optimization in the CatBoost model, as detailed below:

[0102] For explorers, the crossover operation of the genetic algorithm is combined to enhance the explorer's global search capability in the hyperparameter space;

[0103] For followers, a speed update mechanism of particle swarm optimization algorithm is introduced to accelerate the followers toward a better solution;

[0104] For the vigilant, the random walk characteristic of the vigilant is preserved, and the mutation operation of the differential evolution algorithm is used to avoid getting trapped in local optima;

[0105] The optimization objective is set to maximize the fault diagnosis accuracy of the CatBoost model on the validation set. The termination condition for hyperparameter optimization is set to the number of iterations reaching 100 or the accuracy reaching the convergence condition (accuracy reaching 95% or accuracy change less than 0.001).

[0106] As an example, the optimal combination of hyperparameters chosen here is learning_rate = 0.08, depth = 7, iterations = 1200, l2_leaf_reg = 2.5.

[0107] In other embodiments, to adapt to non-stationary environments, the weights of the GNN model and the hyperparameters of the CatBoost model can be updated online based on real-time monitored environmental parameters (such as temperature, electromagnetic field strength, and humidity). The update frequency can be: adjusting the model every minute based on the latest environmental parameters.

[0108] S140 uses a trained CatBoost model to perform fault diagnosis on a quantum sensor multi-detector array.

[0109] In some embodiments, the trained CatBoost model can be deployed in a real-time monitoring system for a quantum sensor multi-detector array (such as NVIDIA Jetson or a GPU server). The system continuously collects relevant data from the quantum sensor multi-detector array, which is then preprocessed and input into the trained CatBoost model. The model outputs the fault type, probability (such as normal, sensor failure, signal interference), and potential fault time.

[0110] It is worth noting that, in addition to S110-S140, method 100 may also include: pruning and quantizing the trained CatBoost model (such as INT8 quantization) to reduce the inference time of the CatBoost model and ensure that the time for a single diagnosis is less than 80 milliseconds, thus meeting the real-time requirements.

[0111] Meanwhile, method 100 may also include: displaying the fault diagnosis results in a visual form, including a fault probability distribution map (showing the confidence level of various faults), a condition assessment report (such as "Sensor 1 will fail within 10 minutes, confidence level 0.95"), and time series outlier annotations. Furthermore, the fault diagnosis results can also be transmitted to the monitoring system via a human-machine interface or an application programming interface (API).

[0112] Furthermore, method 100 may also include: comparing the fault diagnosis results of the quantum sensor multi-detector array output by the trained CatBoost model with the actual operating state of the quantum sensor multi-detector array, recording false alarms and / or missed alarms, and updating the parameters of the CatBoost model accordingly to continuously improve the accuracy of fault diagnosis.

[0113] In summary, this disclosure achieves at least the following technical effects:

[0114] (1) By integrating the GNN model and the CatBoost model, this disclosure can capture the dynamic spatiotemporal dependence of quantum sensor multi-detector arrays and realize the early prediction of potential faults.

[0115] (2) An adaptive learning mechanism is introduced, which dynamically adapts to complex and non-stationary environments by updating the weights of the GNN model and the hyperparameters of the CatBoost model online, so as to ensure the robustness and adaptability of diagnostic performance in dynamic environments.

[0116] (3) Effectively integrate multimodal data (time series signal data, spatial topology data, and environmental data), extract the topological relationship between sensors through the GNN model, and process high-dimensional features through the CatBoost model to achieve comprehensive fault mode analysis and improve the accuracy of fault prediction.

[0117] (4) The improved sparrow search algorithm, which integrates genetic algorithm, particle swarm optimization algorithm and differential evolution algorithm, is used for hyperparameter optimization, which can shorten the hyperparameter search time and improve the optimization accuracy.

[0118] (5) By using model pruning, quantization techniques and efficient optimization strategies, the response speed and operating efficiency of the model have been improved, which can meet the high efficiency requirements of real-time status assessment and fault diagnosis of quantum sensor multi-detector arrays.

[0119] (6) Through dynamic prediction, online learning and multi-source data fusion, the model can identify potential faults in a timely manner and provide early warnings, significantly improving the operational reliability of quantum sensor arrays.

[0120] The method 100 will be described in detail below with reference to a specific embodiment, as shown in the following figure:

[0121] This method was applied to a multi-detector array of quantum sensors containing 16 detection units (superconducting quantum interference devices)100. The signal sampling rate was set to 10 kHz, and the fault types included normal operation, sensor failure, and signal drift. A fault diagnosis dataset consisting of 5000 normal samples and 500 fault samples was generated by preprocessing the acquired raw data labeled with fault types using a GNN model. Each sample had 89 feature dimensions. The CatBoost model was trained using the fault diagnosis dataset, and an improved sparrow search algorithm incorporating genetic algorithms, particle swarm optimization, and differential evolution was used to optimize the hyperparameters of the CatBoost model during training. Furthermore, online learning was performed based on real-time monitored environmental parameters. The final optimal hyperparameters were: learning_rate = 0.08, depth = 7, iterations = 1200, l2_leaf_reg = 2.5. The trained CatBoost model achieved a classification accuracy of 97.2% on the validation set, a prediction lead time of 10 minutes, and a false alarm rate of less than 2%, meeting the system requirements. The trained CatBoost model was then used to diagnose faults in the quantum sensor multi-detector array.

[0122] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this disclosure is not limited to the described order of actions, because according to this disclosure, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to this disclosure.

[0123] The above is an introduction to the method embodiments. The following describes the solution described in this disclosure further through device embodiments.

[0124] Figure 2 A structural diagram of a fault diagnosis device for a quantum sensor multi-detector array combining spatial topology, as provided in an embodiment of this disclosure, is shown. Figure 2 As shown, the device 200 may include:

[0125] The acquisition module 210 is used to acquire raw data of the quantum sensor multi-detector array with labeled fault types. The raw data includes time-series signal data, spatial topology data, and environmental data.

[0126] Module 220 is used to preprocess the raw data and build a fault diagnosis dataset.

[0127] Training module 230 is used to train the CatBoost model using a fault diagnosis dataset, and during the training process, it uses an improved sparrow search algorithm that integrates genetic algorithm, particle swarm optimization algorithm and differential evolution algorithm to optimize the hyperparameters of the CatBoost model.

[0128] The diagnostic module 240 is used to perform fault diagnosis on the quantum sensor multi-detector array using a trained CatBoost model.

[0129] Understandable, Figure 2 Each module / unit in the illustrated device 200 has the ability to implement Figure 1 The functions of each step in method 100 shown, and their corresponding technical effects, will not be elaborated here for the sake of brevity.

[0130] Figure 3 A structural diagram of an exemplary electronic device capable of implementing embodiments of the present disclosure is shown. Electronic device 300 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. Electronic device 300 may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0131] like Figure 3 As shown, the electronic device 300 may include a computing unit 301, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 302 or a computer program loaded from a storage unit 308 into a random access memory (RAM) 303. The RAM 303 may also store various programs and data required for the operation of the electronic device 300. The computing unit 301, ROM 302, and RAM 303 are interconnected via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0132] Multiple components in electronic device 300 are connected to I / O interface 305, including: input unit 306, such as keyboard, mouse, etc.; output unit 307, such as various types of displays, speakers, etc.; storage unit 308, such as disk, optical disk, etc.; and communication unit 309, such as network card, modem, wireless transceiver, etc. Communication unit 309 allows electronic device 300 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0133] The computing unit 301 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 301 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 301 performs the various methods and processes described above, such as method 100. For example, in some embodiments, method 100 may be implemented as a computer program product, including a computer program tangibly contained in a computer-readable medium, such as storage unit 308. In some embodiments, part or all of the computer program may be loaded and / or installed on device 300 via ROM 302 and / or communication unit 309. When the computer program is loaded into RAM 303 and executed by the computing unit 301, one or more steps of method 100 described above may be performed. Alternatively, in other embodiments, the computing unit 301 may be configured to perform method 100 by any other suitable means (e.g., by means of firmware).

[0134] The various embodiments described above 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), system-on-a-chip (SoCs), payload programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0135] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0136] In the context of this disclosure, a computer-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of computer-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0137] It should be noted that this disclosure also provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to execute method 100 and achieve the corresponding technical effects achieved by the embodiments of this disclosure in executing the method. For the sake of brevity, these will not be elaborated here.

[0138] In addition, this disclosure also provides a computer program product including a computer program that implements method 100 when executed by a processor.

[0139] To provide interaction with a user, the embodiments described above can be implemented on a computer 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 pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0140] The embodiments described above can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with the implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication (e.g., a communication network) of any form or medium. Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0141] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0142] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0143] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A fault diagnosis method for a quantum sensor multi-detector array combining spatial topology, characterized in that, The method includes: Acquire raw data of a quantum sensor multi-detector array with labeled fault types. The raw data includes time-series signal data, spatial topology data, and environmental data. The raw data is preprocessed to construct a fault diagnosis dataset; The CatBoost model was trained using a fault diagnosis dataset, and an improved sparrow search algorithm that integrates genetic algorithm, particle swarm optimization algorithm and differential evolution algorithm was used to optimize the hyperparameters of the CatBoost model during the training process. Fault diagnosis of quantum sensor multi-detector arrays was performed using a trained CatBoost model.

2. The method according to claim 1, characterized in that, The acquisition of raw data with labeled fault types for the quantum sensor multi-detector array includes: Time-series signal data is obtained by acquiring signals from each detector unit in the quantum sensor multi-detector array at a fixed sampling rate using a high-frequency sampling device. Based on the physical layout and signal interaction relationship of the quantum sensor multi-detector array, an adjacency matrix is ​​constructed to obtain spatial topology data; Collect environmental data at the location of the quantum sensor multi-detector array; By summarizing and analyzing time-series signal data, spatial topology data, and environmental data, we obtained raw data of quantum sensor multi-detector array with labeled fault types.

3. The method according to claim 1, characterized in that, The preprocessing of the raw data to construct the fault diagnosis dataset includes: Denoising and feature extraction are performed on time-series signal data to obtain signal features; A dynamic topology graph is constructed based on spatial topology data, and a graph neural network (GNN) model is used to extract features from the dynamic topology graph to obtain spatiotemporal features. Environmental features are obtained by extracting features from environmental data; By splicing together signal features, spatiotemporal features, and environmental features, a comprehensive feature is obtained. Using comprehensive features as sample features and their corresponding fault types as sample labels, samples are constructed, and an initial dataset is built based on this. The initial dataset is augmented with a small number of fault type samples to obtain a fault diagnosis dataset.

4. The method according to claim 1, characterized in that, The method involves training the CatBoost model using a fault diagnosis dataset, and optimizing its hyperparameters during training using an improved sparrow search algorithm that integrates genetic algorithms, particle swarm optimization algorithms, and differential evolution algorithms. This includes: The fault diagnosis dataset is divided into a training set and a validation set according to a preset ratio. The CatBoost model is trained using the training set, and the performance of the CatBoost model is evaluated using K-fold cross-validation. The improved sparrow search algorithm, which integrates genetic algorithm, particle swarm optimization algorithm, and differential evolution algorithm, is used to optimize the hyperparameters of the CatBoost model.

5. The method according to claim 4, characterized in that, The improved sparrow search algorithm, which integrates genetic algorithm, particle swarm optimization algorithm, and differential evolution algorithm, includes: For explorers, the crossover operation of the genetic algorithm is combined to enhance the explorer's global search capability in the hyperparameter space; For followers, a speed update mechanism of particle swarm optimization algorithm is introduced to accelerate the followers toward a better solution; For the vigilant, the random walk characteristic of the vigilant is preserved, and the mutation operation of the differential evolution algorithm is used to avoid getting trapped in local optima; The optimization objective is set to maximize the fault diagnosis accuracy of the CatBoost model on the validation set.

6. The method according to claim 5, characterized in that, The genetic algorithm includes: Selection: The roulette wheel selection method is used to select high-quality individuals based on their fitness. Here, an individual is a combination of hyperparameters of the CatBoost model, and the fitness of an individual is the fault diagnosis accuracy of the CatBoost model on the validation set when the individual is selected. Crossover: A single-point crossover occurs with a crossover probability of 0.8, generating a new individual; Mutation: Randomly perturb individuals with a mutation probability of 0.1 to increase population diversity.

7. The method according to claim 5, characterized in that, The particle swarm optimization algorithm includes: The position and velocity of an individual are updated based on its historical best position and global best position, where the individual is a combination of hyperparameters of the CatBoost model.

8. The method according to claim 5, characterized in that, The differential evolution algorithm includes: The individual's position is updated based on the positions of multiple randomly selected individuals.

9. The method according to claim 1, characterized in that, The method further includes: The fault diagnosis results of the quantum sensor multi-detector array output by the trained CatBoost model are compared with the actual operating status of the quantum sensor multi-detector array. False alarms and / or missed alarms are recorded, and the parameters of the CatBoost model are updated accordingly.

10. The method according to claim 3, characterized in that, The method further includes: The weights of the GNN model and the hyperparameters of the CatBoost model are updated online based on real-time monitored environmental parameters.

Images