Machine room equipment intelligent remote operation and maintenance method and platform

Abstract

The application discloses a kind of machine room equipment intelligent remote operation and maintenance method and platform, it is related to intelligent operation and maintenance technical field, including, construct bayesian probability model, and combine optical signal characteristics and current harmonic characteristics, predict the failure probability distribution of machine room equipment;According to the failure probability distribution of machine room equipment, convert fault coordinates into equipment logical identifier, and detect the quantum teleportation link of machine room equipment, generate machine room equipment switching instruction according to detection result;According to machine room equipment switching instruction, send Bell state measurement command to fault equipment, and carry out quantum state reconstruction and machine room equipment switching.The application constructs multilayer bayesian probability model based on Gaussian distribution, Granger causality test and variational inference, realizes the nonlinear correlation analysis of optical signal wavelength shift and current harmonic characteristics, can accurately capture the weak signs of potential equipment failure, to significantly improve the sensitivity and reliability of failure prediction.

Description

Technical Field

[0001] This invention relates to the field of intelligent operation and maintenance technology, and in particular to an intelligent remote operation and maintenance method and platform for data center equipment. Background Technology

[0002] As data centers continue to expand, intelligent remote operation and maintenance (O&M) technologies for data center equipment have become crucial for ensuring the reliability of IT infrastructure. Traditional O&M methods primarily rely on threshold alarms and manual inspections. In recent years, real-time monitoring based on the Internet of Things (IoT) and machine learning prediction technologies have been gradually introduced. However, these methods still have limitations in terms of real-time performance, multi-source data fusion, and proactive intervention capabilities. In particular, their accuracy in capturing physical layer anomalies such as optical signal wavelength shifts and current harmonics is insufficient, leading to a high false alarm rate for fault warnings.

[0003] In existing technologies, data center equipment operation and maintenance systems typically lack intelligent management capabilities for quantum communication links. When equipment malfunctions, traditional methods can only perform simple service switching, failing to guarantee the continuity of quantum state transmission. Furthermore, existing solutions often rely on fixed thresholds to determine equipment risk levels, making them ill-suited for complex and ever-changing operating environments and resulting in delayed responses to sudden failures. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides an intelligent remote operation and maintenance method for data center equipment to solve the technical problems of existing data center operation and maintenance systems being unable to guarantee the continuity of quantum state transmission and the lag in response to sudden faults.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides an intelligent remote operation and maintenance method for computer room equipment, which includes monitoring optical signal wavelength offset data and current harmonic data, and performing preprocessing.

[0008] A Bayesian probability model is constructed, and combined with optical signal characteristics and current harmonic characteristics, the failure probability distribution of equipment in the computer room is predicted.

[0009] Based on the failure probability distribution of the equipment in the data center, the fault coordinates are converted into logical identifiers of the equipment, and the quantum teleportation links of the equipment in the data center are detected. Based on the detection results, the switching instructions for the equipment in the data center are generated.

[0010] Based on the equipment switching instructions in the computer room, a Bell state measurement command is sent to the faulty equipment, and quantum state reconstruction and equipment switching in the computer room are performed.

[0011] As a preferred embodiment of the intelligent remote operation and maintenance method for computer room equipment described in this invention, the optical signal wavelength offset data includes the deviation between the center wavelength value and the reference wavelength, the offset duration, and the rate of change.

[0012] The current harmonic data includes the amplitude of the harmonic components, the total harmonic distortion rate, the phase angle, and the spectral energy distribution.

[0013] The preprocessing includes data normalization, noise filtering, outlier extraction, and time series alignment.

[0014] As a preferred embodiment of the intelligent remote operation and maintenance method for data center equipment described in this invention, the steps for constructing the Bayesian probability model are as follows:

[0015] A Gaussian distribution is used to fit the optical signal wavelength offset data and current harmonic data into physical layer nodes;

[0016] Based on conditional probability tables and historical fault data, latent variable nodes are constructed by defining the dependency relationship between optical signal anomalies and latent variable nodes.

[0017] Application layer nodes are constructed based on latent variable nodes using a logistic regression algorithm;

[0018] Granger causality test is used to analyze the causal relationship between physical layer nodes and latent variable nodes, and a hierarchical directed acyclic graph is constructed using BayesPy;

[0019] Based on physical layer nodes, latent variable nodes, application layer nodes, and a hierarchical directed acyclic graph, a Bayesian probabilistic model is constructed using the EM algorithm and variational inference.

[0020] As a preferred embodiment of the intelligent remote operation and maintenance method for data center equipment described in this invention, the steps for predicting the fault probability distribution of data center equipment by combining optical signal characteristics and current harmonic characteristics are as follows:

[0021] Based on the preprocessed optical signal wavelength offset data, optical signal features are extracted through time-frequency analysis.

[0022] Based on the preprocessed current harmonic data, current harmonic features are extracted through fast Fourier transform and harmonic distortion analysis.

[0023] By inputting optical signal characteristics and current harmonic characteristics into a Bayesian probability model, and using a posterior probability algorithm based on dynamic feature fusion and time decay, the real-time failure probability P(D) of the data center equipment is predicted. k );

[0024] A spatial interpolation kernel density estimation algorithm is used to map the failure probability of all equipment in the computer room according to their spatial location, thereby generating a failure probability distribution of the equipment in the computer room.

[0025] As a preferred embodiment of the intelligent remote operation and maintenance method for data center equipment described in this invention, the step of converting fault coordinates into equipment logical identifiers based on the fault probability distribution of the data center equipment includes the following steps:

[0026] Based on statistical analysis of historical fault data, a risk warning threshold F1 is defined.

[0027] When P(D) k If F1 ≥ F1, then the current equipment in the computer room is considered to be high-risk equipment;

[0028] Based on the failure probability distribution of equipment in the computer room, the location coordinates of high-risk equipment are identified, and the logical identifiers of high-risk equipment are identified from the BIM database.

[0029] As a preferred embodiment of the intelligent remote operation and maintenance method for data center equipment described in this invention, the steps for detecting the quantum teleportation link of the data center equipment and generating a data center equipment switching command based on the detection results are as follows:

[0030] Based on the logical identifiers of high-risk devices, the liveness status of entangled pairs between devices is detected by QKD, and the liveness channel is identified.

[0031] The entanglement degree and quantum bit error rate of the live channel are measured in real time, and the quantum channel quality score Q is obtained by compressed sensing quantum tomography.

[0032] Based on statistical analysis of historical quantum communication data, a channel quality threshold Q1 is defined.

[0033] When Q < Q1, the quantum channel quality is considered substandard, and a device switching command is generated via QNMP.

[0034] As a preferred embodiment of the intelligent remote operation and maintenance method for data center equipment described in this invention, the steps of sending a Bell state measurement command to the faulty equipment according to the data center equipment switching instruction, and performing quantum state reconstruction and data center equipment switching are as follows:

[0035] Based on the device switching command, a Bell state measurement command is sent to the high-risk device through the quantum control interface. A polarized beam splitter is used to realize Bell state analysis. The entangled state particles in the quantum memory of the high-risk device are subjected to joint Bell basis measurement of qubits, and a 2-qubit classical measurement result is generated.

[0036] Based on the 2-qubit classical measurement results, quantum state reconstruction is performed, and the fidelity Y of the reconstructed state is verified by compressed sensing quantum tomography.

[0037] Based on the reconfiguration state fidelity, the service traffic is migrated to the backup data center equipment through the load balancer, and the quantum resources of high-risk equipment are released, generating a data center equipment switchover report.

[0038] Secondly, the present invention provides an intelligent remote operation and maintenance platform for data center equipment, including a data acquisition module, a fault probability prediction module, an instruction generation module, and an equipment switching module.

[0039] The data acquisition module is used to monitor optical signal wavelength offset data and current harmonic data, and to perform preprocessing.

[0040] The fault probability prediction module is used to build a Bayesian probability model and combine optical signal characteristics and current harmonic characteristics to predict the fault probability distribution of equipment in the computer room.

[0041] The instruction generation module is used to convert fault coordinates into device logical identifiers based on the fault probability distribution of the equipment in the data center, detect the quantum teleportation link of the equipment in the data center, and generate switching instructions for the equipment in the data center based on the detection results.

[0042] The equipment switching module is used to send Bell state measurement commands to the faulty equipment according to the equipment switching instructions in the computer room, and to perform quantum state reconstruction and equipment switching in the computer room.

[0043] Thirdly, the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the intelligent remote operation and maintenance method for data center equipment as described in the first aspect of the present invention.

[0044] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the intelligent remote operation and maintenance method for data center equipment as described in the first aspect of the present invention.

[0045] The beneficial effects of this invention are as follows: By constructing a multi-layer Bayesian probability model based on Gaussian distribution, Granger causality test, and variational inference, nonlinear correlation analysis of optical signal wavelength shift and current harmonic characteristics is achieved, which can accurately capture weak signs of potential equipment faults, thereby significantly improving the sensitivity and reliability of fault prediction. Through dynamic quality assessment of quantum teleportation links and compressed sensing quantum tomography verification, intelligent switching and recycling of quantum resources are realized while ensuring the fidelity of reconstructed states. This ensures both the continuity of quantum state transmission and optimizes the utilization rate of quantum communication resources. Attached Figure Description

[0046] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0047] Figure 1 This is a flowchart of the intelligent remote operation and maintenance method for computer room equipment in Example 1.

[0048] Figure 2 This is a schematic diagram of the intelligent remote operation and maintenance platform for computer room equipment in Example 1.

[0049] Figure 3 This is a flowchart of the fault probability prediction in Example 1.

[0050] Figure 4 This is a flowchart of the Bayesian probability model construction in Example 1. Detailed Implementation

[0051] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0052] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0053] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0054] Example 1, referring to Figures 1-4 This embodiment provides an intelligent remote operation and maintenance method for data center equipment, including the following steps:

[0055] S1. Monitor the wavelength offset data of the optical signal and the harmonic data of the current, and perform preprocessing.

[0056] The optical signal wavelength offset data includes the deviation between the center wavelength value and the reference wavelength, the offset duration, and the rate of change;

[0057] Current harmonic data includes the amplitude of harmonic components, total harmonic distortion, phase angle, and spectral energy distribution;

[0058] It should be noted that the signal wavelength offset data is obtained in real time through a tunable laser source and a wavelength meter. The deviation between the center wavelength value and the reference wavelength is recorded using a high-precision spectral analyzer at a fixed sampling interval (e.g., every 10ms). The offset duration is calculated using the timestamp difference, and the rate of change is obtained based on the difference in wavelength change at consecutive sampling points. The current harmonic data is acquired by a broadband current sensor. The amplitude, total harmonic distortion rate, and phase angle of the harmonic components are extracted synchronously through a lock-in amplifier (e.g., sampling rate 50kHz). The spectral energy distribution is calculated by performing a fast Fourier transform on the original waveform and then calculating the energy integral of each frequency band. All data are synchronously filtered digitally to eliminate power frequency interference before being stored in a time series database.

[0059] Preprocessing includes data normalization, noise filtering, outlier extraction, and time series alignment.

[0060] Furthermore, the data normalization adopts the maximum and minimum value scaling method, which linearly transforms the original values ​​of optical signal wavelength offset data and current harmonic data to the [0,1] interval. The maximum and minimum values ​​are dynamically calculated based on historical data within a sliding time window (e.g., window length of 60 seconds) to ensure the comparability of parameters with different dimensions.

[0061] Noise filtering is achieved using a digital Butterworth low-pass filter. A cutoff frequency (e.g., 5Hz) is set for the wavelength offset data of the optical signal to suppress high-frequency jitter. Adaptive threshold wavelet denoising is used for the current harmonic data to retain the fundamental and main harmonic components while eliminating random interference.

[0062] Outlier extraction is based on a dynamic threshold outlier detection algorithm. The moving average and three-times-standard-deviation range are calculated for optical signal wavelength offset data. For current harmonic data, the box plot method is used to identify data points exceeding 1.5 times the interquartile range. All outliers are marked and replaced by linear interpolation.

[0063] The time series alignment uses a dynamic time warping algorithm to compensate for the time difference between the acquisition times of optical signal wavelength offset data and current harmonic data. By finding the minimum curved path, the timestamps of the two types of data are forced to match (example: alignment accuracy reaches 1ms), forming a strictly synchronized feature sequence.

[0064] S2. Construct a Bayesian probability model and combine optical signal characteristics and current harmonic characteristics to predict the failure probability distribution of equipment in the computer room;

[0065] A Gaussian distribution is used to fit the optical signal wavelength offset data and current harmonic data into physical layer nodes;

[0066] Furthermore, statistical analysis is first performed on the wavelength offset data of the optical signal to calculate its distribution characteristics and fit it to a Gaussian distribution describing the wavelength offset characteristics, forming the optical signal physical layer node. Simultaneously, statistical analysis is performed on the current harmonic data to calculate its distribution characteristics and fit it to a Gaussian distribution describing the harmonic characteristics, forming the current harmonic physical layer node. These two physical layer nodes output the probability distribution information of their respective characteristics: the optical signal physical layer node outputs the wavelength offset probability distribution, and the current harmonic physical layer node outputs the harmonic characteristic probability distribution. All distribution parameters are calculated using historical operational data and are continuously updated and maintained.

[0067] Based on conditional probability tables and historical fault data, latent variable nodes are constructed by defining the dependency relationship between optical signal anomalies and latent variable nodes.

[0068] Application layer nodes are constructed based on latent variable nodes using a logistic regression algorithm;

[0069] Furthermore, we first analyze the co-occurrence frequency of optical signal wavelength shift and equipment failure in historical data to establish a conditional probability distribution P(failure state | optical signal anomaly). Then, we define the dependency relationship between the latent variable node and the optical signal physical layer node, and calculate the posterior probability distribution of the latent variable node using Bayes' theorem. Finally, we construct a latent variable node that reflects the potential failure risk of the equipment. The output of this node is a failure risk probability value that comprehensively considers the characteristics of optical signal anomalies.

[0070] It should be noted that the conditional probability table is a lookup table built by analyzing the correspondence between optical signal anomalies and equipment failure states in historical operational data. Specifically, the construction process involves: statistically analyzing the proportion of equipment failures when optical signal wavelength deviations exceed a specific range, and simultaneously statistically analyzing the proportion of failures within the normal range. These proportions are then compiled into a failure probability lookup table corresponding to different degrees of optical signal anomalies. This table directly reflects the probabilistic correlation between optical signal anomaly characteristics and equipment failure states.

[0071] Granger causality test is used to analyze the causal relationship between physical layer nodes and latent variable nodes, and a hierarchical directed acyclic graph is constructed using BayesPy;

[0072] Furthermore, the Granger causality test is used to verify the predictive ability of physical layer node time series data for latent variable node time series data. By comparing the prediction error variances of latent variable nodes with and without historical information from physical layer nodes, it determines whether physical layer nodes constitute a Granger cause of latent variable nodes. For causal relationships that pass the test, a hierarchical directed acyclic graph is constructed using BayesPy based on the verification results. In the graph, physical layer nodes act as parent nodes pointing to their corresponding latent variable nodes, forming a probabilistic graph structure with a clear causal direction. During the test, the lag order is automatically determined based on the Bayesian information criterion; for example, the maximum lag order is set to 5.

[0073] Based on physical layer nodes, latent variable nodes, application layer nodes, and a hierarchical directed acyclic graph, a Bayesian probabilistic model is constructed using the EM algorithm and variational inference.

[0074] Furthermore, physical layer nodes, latent variable nodes, and application layer nodes are probabilistically modeled based on the structural relationships of a hierarchical directed acyclic graph. Parameter optimization is achieved using the EM algorithm: in the expectation step, the expected posterior probability distribution of each node is calculated using the current conditional probability parameters; in the maximization step, the conditional probability parameters are updated based on the expected statistic, gradually increasing the model's log-likelihood function value. Simultaneously, variational inference techniques are employed to approximate complex posterior distributions into tractable variational distributions, achieving efficient probabilistic inference by optimizing the variational lower bound. The causal relationships determined by the Granger causality test guide the connection methods between nodes, ensuring that physical layer nodes maintain an effective influence on latent variable nodes. The final constructed Bayesian probabilistic model fully preserves the topological structure of the hierarchical directed acyclic graph, and the conditional dependencies between nodes are accurately expressed through the optimized parameter matrix.

[0075] Based on the preprocessed optical signal wavelength offset data, optical signal features are extracted through time-frequency analysis.

[0076] It should be noted that after the optical signal wavelength offset data is preprocessed, short-time Fourier transform is used for time-frequency analysis to calculate the energy distribution of the signal in the time-frequency domain and extract parameters such as the main frequency band energy ratio, frequency band energy entropy, and instantaneous frequency fluctuation rate, which together constitute the optical signal characteristics characterizing the abnormal state of the optical signal.

[0077] Based on the preprocessed current harmonic data, current harmonic features are extracted through fast Fourier transform and harmonic distortion analysis.

[0078] It should be noted that after the current harmonic data is preprocessed, a fast Fourier transform is first performed to obtain the spectral distribution. Then, the ratio of the amplitude of each harmonic component to the amplitude of the fundamental wave is calculated. At the same time, the total harmonic distortion rate and the phase relationship of each harmonic are analyzed. Finally, the energy proportion of the main frequency band is calculated by spectral integration. These analysis results together constitute the current harmonic characteristics that reflect power quality anomalies.

[0079] By inputting optical signal characteristics and current harmonic characteristics into a Bayesian probability model, and using a posterior probability algorithm based on dynamic feature fusion and time decay, the real-time failure probability P(D) of the data center equipment is predicted. k The expression is:

[0080]

[0081] Among them, P(D) k ) represents the real-time failure probability of the k-th computer room device. The observed value of the m-th feature at time t, w m It is the weight coefficient of the m-th feature. Indicates the fault state D of the kth computer room equipment. k Fault characteristics x m The conditional probability, P'(D) k ) represents the basic failure probability of equipment k in the computer room, γ is the confidence adjustment coefficient for the basic failure probability of the equipment in the computer room, and D k The fault status of the kth equipment in the computer room is represented by M, which is the total number of optical signal features and current harmonic features, m is the index variable of all features (including optical signal features and current harmonic features), λ represents the exponential decay coefficient of historical data on the current fault probability prediction, Δt represents the time difference between time t and the historical data acquisition time, and Z represents the normalization factor.

[0082] Furthermore, after inputting the optical signal characteristics and current harmonic characteristics into the Bayesian probability model, the model first queries the conditional probability table for each characteristic. In the current device state D k conditional probability For example, the probability of the main lobe energy concentration of the optical signal under fault conditions is 0.85. Then, the weighted sum of the logarithmic probabilities of each feature is calculated, with the weighting coefficient w... m The weights are dynamically adjusted based on feature importance; in this example, the weight of the optical signal feature is set to 0.6, and the weight of the current harmonic feature is set to 0.4. The probability of basic equipment failure, P'(D), is also considered. kThe logarithmic term of the equation is used, with a confidence adjustment coefficient γ of 0.3 to balance prior influence. The time decay term λΔt exponentially decays the contribution of historical data, with a decay coefficient λ set to 0.01 / minute. Finally, all terms are summed and the exponent is taken, and a normalization factor Z is used to ensure that the probability value is between 0 and 1. During the calculation, the feature index variable m traverses all M optical signal features and current harmonic features, and the device status D... k A value of 0 or 1 represents normal and fault states, respectively. The entire calculation process achieves real-time updates of the fault probability through dynamic feature fusion and a time decay mechanism.

[0083] A spatial interpolation kernel density estimation algorithm is used to map the failure probability of all equipment in the computer room according to their spatial location, thereby generating a failure probability distribution of the equipment in the computer room.

[0084] It should be noted that the spatial interpolation kernel density estimation algorithm first obtains the three-dimensional coordinate positions of all data center equipment and the real-time failure probability P(D) of the corresponding k-th data center equipment. k The algorithm employs a Gaussian kernel function as the spatial weighting function, with the kernel bandwidth automatically determined using the Silverman rule. For any point in the space, the algorithm calculates the weighted average of the failure probabilities of all equipment within a radius surrounding that point. The weights are determined by the distance of the equipment from the target point, with closer equipment receiving greater weights. For example, when calculating a certain point, the algorithm considers 12 pieces of equipment within a 5-meter radius, with the contribution weight of each piece of equipment decreasing with distance according to the Gaussian kernel function. By traversing the grid points in the data center space, a three-dimensional failure probability distribution map covering the entire data center area is finally generated. The value of each grid point in the map represents the failure risk probability density at that location. During the calculation process, the base failure probability P'(D) is maintained. k The numerical value remains unchanged, and the probability density is expressed continuously only through spatial interpolation.

[0085] Dynamic feature fusion specifically refers to the ability to identify equipment faults by dynamically quantizing optical signal features and current harmonic features using KL divergence.

[0086] It is based on historical data (such as abnormal records of optical signal wavelength offset and current harmonic distortion events, which are statistically analyzed according to equipment failure status to generate a conditional probability distribution) and fitted using the kernel density estimation method;

[0087] D k ∈{0,1}, where 0 represents normal and 1 represents fault.

[0088] S3. Based on the failure probability distribution of the equipment in the computer room, convert the fault coordinates into equipment logical identifiers, detect the quantum teleportation links of the equipment in the computer room, and generate switching instructions for the equipment in the computer room based on the detection results.

[0089] Based on statistical analysis of historical fault data, a risk warning threshold F1 is defined.

[0090] Furthermore, the risk warning threshold F1 is defined based on the probabilistic characteristics exhibited by the equipment before failure in historical fault data. The specific definition process is as follows: First, a large amount of historical monitoring data is collected during normal equipment operation and before failure, and the real-time failure probability P(D) of the k-th equipment in the computer room before the failure is analyzed. k The typical change pattern of the fault is determined; then, a statistical distribution analysis method is used to determine the probability critical value that can cover the vast majority of real fault cases; finally, this critical value is used as the risk warning threshold F1 to identify high-risk equipment in real-time monitoring.

[0091] When P(D) k If F1 ≥ F1, then the current equipment in the computer room is considered to be high-risk equipment;

[0092] Based on the failure probability distribution of equipment in the computer room, the location coordinates of high-risk equipment are identified, and the logical identifiers of high-risk equipment are identified from the BIM database.

[0093] Furthermore, the failure probability distribution data of the equipment in the computer room is spatially matched with the BIM database. First, the location of high-risk equipment with a failure probability exceeding the threshold F1 is located in the three-dimensional coordinate system, and the X / Y / Z coordinate values ​​of the corresponding location are extracted. Then, the spatial indexing function of the BIM database is used to retrieve matching equipment objects with coordinate values ​​as query conditions, and the unique logical identifier of the equipment in the BIM model is obtained, including metadata information such as equipment code, type number and system to which it belongs.

[0094] Based on the logical identifiers of high-risk devices, the liveness status of entangled pairs between devices is detected by QKD, and the liveness channel is identified.

[0095] Furthermore, after the logical identifier of a high-risk device is input into the QKD detection process, the state of all entangled pairs that have established quantum connections with that device is first queried through the quantum key distribution protocol. Belli measurements are then used to verify the survival status of each entangled pair. For surviving entangled pairs, the corresponding quantum channel number is recorded, and channel parameters, including entanglement fidelity, quantum bit error rate, and transmission delay, are measured. For example, among the three quantum channels corresponding to the logical identifier MEP-AC-025, channel QCH-1024 has a measured entanglement fidelity of 0.95, a quantum bit error rate of 2%, and a delay of 120 μs, and is therefore marked as a surviving channel. All surviving channel information is stored in the quantum channel state table for subsequent quality assessment.

[0096] The entanglement degree and quantum bit error rate of the live channel are measured in real time, and the quantum channel quality score Q is obtained by compressed sensing quantum tomography, expressed as:

[0097]

[0098] Where S is the entanglement fidelity, R is the entanglement generation rate, QBER is the quantum bit error rate, τ is the channel delay, and τ0 is the reference delay.

[0099] It should be noted that the calculation process of the quantum channel quality score Q for the surviving channel is as follows: First, the entanglement fidelity S of the surviving channel is measured in real time using quantum state tomography. The entanglement generation rate R is calculated using the coincidence count rate. Simultaneously, the quantum bit error rate QBER is collected using a single-photon detector array, and the channel transmission delay τ is recorded. The reference delay τ0 is set to 100 μs as a reference. During the calculation, the product of the entanglement fidelity S and the entanglement generation rate R is used as the numerator to characterize the channel transmission quality; the quantum bit error rate QBER and the delay influence factor are calculated. The product of these terms serves as the denominator, reflecting the effects of channel noise and delay. Finally, the quantum channel quality score Q is obtained by dividing the numerator by the denominator.

[0100] Based on statistical analysis of historical quantum communication data, a channel quality threshold Q1 is defined.

[0101] It should be noted that the channel quality threshold Q1 is defined based on the correspondence between channel quality scores Q and communication reliability in historical quantum communication data. The specific definition process is as follows: Collect quantum channel quality scores Q from a large amount of historical operational data; analyze the relationship between these Q values ​​and actual communication quality; statistically determine the minimum Q value level that can guarantee reliable quantum state transmission; determine a critical value through probability distribution analysis, ensuring that the Q values ​​of the vast majority of reliable communication cases are higher than this critical value; finally, determine this critical value as the channel quality threshold Q1, used to judge whether the current quantum channel meets the communication quality requirements.

[0102] When Q < Q1, the quantum channel quality is considered substandard, and a device switching command is generated via QNMP.

[0103] S4. Based on the equipment switching instruction in the computer room, send a Bell state measurement command to the faulty equipment, and perform quantum state reconstruction and equipment switching in the computer room.

[0104] Based on the device switching command, a Bell state measurement command is sent to the high-risk device through the quantum control interface. A polarized beam splitter is used to realize Bell state analysis. The entangled state particles in the quantum memory of the high-risk device are subjected to joint Bell basis measurement of qubits, and a 2-qubit classical measurement result is generated.

[0105] Furthermore, upon receiving the device switching command, the quantum control interface immediately sends a Bell state measurement command to the target high-risk device. The entangled particle pairs stored in the high-risk device's quantum memory are processed by precision optical components: a polarization beam splitter separates the entangled photon pairs according to their polarization states, and the separated photons enter two independent measurement channels. Each measurement channel is equipped with tunable optical elements, enabling projection measurements of different Bell bases. After polarization analysis, the photons are detected by a high-sensitivity single-photon detector. The detectors in both channels work together; when photons are detected simultaneously, a coincidence counting event is generated, which is encoded as a 2-bit classical measurement result. For example, under a specific Bell base measurement, simultaneous detection of photons by both channels may correspond to a 00 result, indicating successful projection onto a specific Bell state.

[0106] It should be noted that the 2-bit classical measurement result represents two binary bits obtained when performing a Belli measurement on a quantum entangled state. This result is a combination of the independent measurement outputs of the two entangled particles; the first bit corresponds to the measurement value of the first particle, and the second bit corresponds to the measurement value of the second particle. The four possible combinations reflect different quantum state correlation characteristics: the 00 combination indicates that the measurement results of the two particles are perfectly correlated, the 11 combination indicates that the measurement results of the two particles are perfectly anticorrelated, and the 01 and 10 combinations show that there is a specific correlation pattern between the particles. These measurement results are obtained through a polarized beam splitter and a single-photon detector, directly reflecting the non-classical correlation characteristics of quantum entanglement, providing crucial information for verifying quantum state quality and executing quantum communication protocols. Each binary combination corresponds to a specific quantum state projection measurement result, fully recording the statistical properties of quantum entanglement.

[0107] Based on the 2-qubit classical measurement results, quantum state reconstruction is performed, and the fidelity of the reconstructed state is verified by compressed sensing quantum tomography. The expression is:

[0108]

[0109] Where Y is the fidelity of the reconstructed state, d is the Hilbert space dimension, ρ is the density matrix of the reconstructed state, and ψ i is the i-th measurement basis selected in the compressed sensing quantum tomography method, N is the number of measurement basis vectors actually used in the compressed sensing quantum tomography method, and i is the index variable of the measurement basis vector;

[0110] Furthermore, the quantum state reconstruction process first establishes a measurement statistical distribution using 2-qubit classical measurement results, and then reconstructs the quantum state density matrix from the limited measurement data using compressed sensing quantum tomography. During reconstruction, N specific measurement basis vectors are selected in Hilbert space for projection measurements, and the expected measurement value under each basis vector is recorded. The fidelity calculation considers the influence of the Hilbert space dimension d, and the projection results under each measurement basis vector are weighted and summed to obtain a normalized fidelity value between 0 and 1. For example, in a two-qubit system, when using 9 measurement basis vectors and the measurement data exhibits good consistency, a reconstructed state fidelity of over 0.9 can be obtained. This method effectively reduces the measurement resources required by traditional quantum tomography while ensuring reconstruction accuracy.

[0111] Based on the reconfiguration state fidelity, the service traffic is migrated to the backup data center equipment through the load balancer, and the quantum resources of high-risk equipment are released, generating a data center equipment switchover report.

[0112] The quantum resources of a high-risk device include entangled particle pairs stored in the device's quantum memory, established quantum channel connections, and available quantum key distribution resources.

[0113] The equipment switchover report includes the identification information of the equipment to be switched, the comparison data of the quantum channel status before and after the switchover, the quantum measurement results during the switchover process, and the resource reallocation plan after the switchover is completed.

[0114] This embodiment also provides an intelligent remote operation and maintenance platform for data center equipment, including: a data acquisition module, a fault probability prediction module, an instruction generation module, and an equipment switching module;

[0115] The data acquisition module is used to monitor optical signal wavelength offset data and current harmonic data, and to perform preprocessing.

[0116] The fault probability prediction module is used to build a Bayesian probability model and combine optical signal characteristics and current harmonic characteristics to predict the fault probability distribution of equipment in the computer room.

[0117] The instruction generation module is used to convert fault coordinates into device logical identifiers based on the fault probability distribution of the equipment in the data center, detect the quantum teleportation link of the equipment in the data center, and generate switching instructions for the equipment in the data center based on the detection results.

[0118] The equipment switching module is used to send Bell state measurement commands to the faulty equipment according to the equipment switching instructions in the computer room, and to perform quantum state reconstruction and equipment switching in the computer room.

[0119] This embodiment also provides a computer device applicable to the intelligent remote operation and maintenance method for data center equipment, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the intelligent remote operation and maintenance method for data center equipment as proposed in the above embodiment.

[0120] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0121] This embodiment also provides a storage medium storing a computer program. When executed by a processor, the program implements the intelligent remote operation and maintenance method for data center equipment as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0122] In summary, this invention achieves nonlinear correlation analysis of optical signal wavelength shift and current harmonic characteristics by constructing a multilayer Bayesian probabilistic model based on Gaussian distribution, Granger causality test, and variational inference. This model accurately captures subtle signs of potential equipment faults, significantly improving the sensitivity and reliability of fault prediction. Furthermore, through dynamic quality assessment of the quantum teleportation link and compressed sensing quantum tomography verification, intelligent switching and recycling of quantum resources are achieved while ensuring the fidelity of the reconstructed state. This guarantees the continuity of quantum state transmission and optimizes the utilization rate of quantum communication resources.

[0123] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for intelligent remote operation and maintenance of data center equipment, characterized in that: include, Monitor optical signal wavelength offset data and current harmonic data, and perform preprocessing; A Bayesian probability model is constructed, and combined with optical signal characteristics and current harmonic characteristics, the failure probability distribution of equipment in the computer room is predicted. Based on the failure probability distribution of the equipment in the data center, the fault coordinates are converted into logical identifiers for the equipment. The quantum teleportation links of the equipment in the data center are then detected. Based on the detection results, a data center equipment switching command is generated. The steps are as follows. Based on the logical identifiers of high-risk devices, the liveness status of entangled pairs between devices is detected by QKD, and the liveness channel is identified. Real-time measurements of entanglement degree and quantum bit error rate were performed on the live channel, and a quantum channel quality score was obtained using compressed sensing quantum tomography. ; Based on statistical analysis of historical quantum communication data, a channel quality threshold Q1 is defined. when If the value is less than Q1, the quantum channel quality is considered substandard, and a device switching command is generated via QNMP. Based on the equipment switching instructions in the computer room, a Bell state measurement command is sent to the faulty equipment, and quantum state reconstruction and equipment switching in the computer room are performed.

2. The intelligent remote operation and maintenance method for data center equipment as described in claim 1, characterized in that: The optical signal wavelength offset data includes the deviation between the center wavelength value and the reference wavelength, the offset duration, and the rate of change. The current harmonic data includes the amplitude of the harmonic components, the total harmonic distortion rate, the phase angle, and the spectral energy distribution. The preprocessing includes data normalization, noise filtering, outlier extraction, and time series alignment.

3. The intelligent remote operation and maintenance method for data center equipment as described in claim 1, characterized in that: The steps for constructing the Bayesian probability model are as follows: A Gaussian distribution is used to fit the optical signal wavelength offset data and current harmonic data into physical layer nodes; Based on conditional probability tables and historical fault data, latent variable nodes are constructed by defining the dependency relationship between optical signal anomalies and latent variable nodes. Application layer nodes are constructed based on latent variable nodes using a logistic regression algorithm; Granger causality test is used to analyze the causal relationship between physical layer nodes and latent variable nodes, and a hierarchical directed acyclic graph is constructed using BayesPy; Based on physical layer nodes, latent variable nodes, application layer nodes, and a hierarchical directed acyclic graph, a Bayesian probabilistic model is constructed using the EM algorithm and variational inference.

4. The intelligent remote operation and maintenance method for data center equipment as described in claim 3, characterized in that: The steps for predicting the failure probability distribution of equipment in the computer room by combining optical signal characteristics and current harmonic characteristics are as follows: Based on the preprocessed optical signal wavelength offset data, optical signal features are extracted through time-frequency analysis. Based on the preprocessed current harmonic data, current harmonic features are extracted through fast Fourier transform and harmonic distortion analysis. By inputting optical signal characteristics and current harmonic characteristics into a Bayesian probability model, and using a posterior probability algorithm based on dynamic feature fusion and time decay, the real-time failure probability of data center equipment is predicted. ; A spatial interpolation kernel density estimation algorithm is used to map the failure probability of all equipment in the computer room according to their spatial location, thereby generating a failure probability distribution of the equipment in the computer room.

5. The intelligent remote operation and maintenance method for data center equipment as described in claim 1, characterized in that: The steps for converting fault coordinates into device logical identifiers based on the fault probability distribution of the equipment in the computer room are as follows: Based on statistical analysis of historical fault data, a risk warning threshold F1 is defined. when If the value is ≥F1, then the current equipment in the computer room is considered to be high-risk equipment; Based on the failure probability distribution of equipment in the computer room, the location coordinates of high-risk equipment are identified, and the logical identifiers of high-risk equipment are identified from the BIM database.

6. The intelligent remote operation and maintenance method for data center equipment as described in claim 5, characterized in that: The steps for sending a Bell state measurement command to the faulty device according to the data center equipment switching instruction, and performing quantum state reconstruction and data center equipment switching are as follows: Based on the device switching command, a Bell state measurement command is sent to the high-risk device through the quantum control interface. A polarized beam splitter is used to realize Bell state analysis. The entangled state particles in the quantum memory of the high-risk device are subjected to joint Bell basis measurement of qubits, and a 2-qubit classical measurement result is generated. Based on the 2-qubit classical measurement results, quantum state reconstruction was performed, and the fidelity of the reconstructed state was verified using compressed sensing quantum tomography. ; Based on the reconfiguration state fidelity, the service traffic is migrated to the backup data center equipment through the load balancer, and the quantum resources of high-risk equipment are released, generating a data center equipment switchover report.

7. A smart remote operation and maintenance platform for data center equipment, based on the smart remote operation and maintenance method for data center equipment according to any one of claims 1 to 5, characterized in that: It includes a data acquisition module, a fault probability prediction module, an instruction generation module, and a device switching module; The data acquisition module is used to monitor optical signal wavelength offset data and current harmonic data, and to perform preprocessing. The fault probability prediction module is used to build a Bayesian probability model and combine optical signal characteristics and current harmonic characteristics to predict the fault probability distribution of equipment in the computer room. The instruction generation module is used to convert fault coordinates into device logical identifiers based on the fault probability distribution of the equipment in the data center, detect the quantum teleportation link of the equipment in the data center, and generate switching instructions for the equipment in the data center based on the detection results. The equipment switching module is used to send Bell state measurement commands to the faulty equipment according to the equipment switching instructions in the computer room, and to perform quantum state reconstruction and equipment switching in the computer room.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the intelligent remote operation and maintenance method for computer room equipment as described in any one of claims 1 to 5.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the intelligent remote operation and maintenance method for computer room equipment as described in any one of claims 1 to 5.

Images