An Internet of Things-based power equipment condition monitoring system

Abstract

The present invention discloses an Internet of Things-based power equipment status monitoring system, which relates to the technical field of wireless communication networks. It includes: a spectrum sensing module that detects and identifies the noise characteristics in electromagnetic signals through a quantum spectrum sensing array to obtain corresponding spectrum data and classification results; a dynamic evaluation module that filters out a list of available channels through dynamic channel survivability, and determines corresponding channels for signal transmission from the list of available channels according to the classification results of the SVW classifier; a frequency hopping optimization module that determines a globally optimal frequency hopping sequence through the list of available channels and the total energy between different channels. The present invention realizes the capture of microsecond-level transient noise through a superconducting resonator and a quantum compressive sensing algorithm, and at the same time, real-time filters out anti-interference channels through dynamic channel survivability and obtains an optimal channel combination, enabling it to complete channel switching within 10 ms.

Description

Technical Field

[0001] The present invention relates to the technical field of wireless communication networks, and specifically to a power equipment status monitoring system based on the Internet of Things. Background Art

[0002] With the growth of global energy demand and the continuous expansion of the power grid scale, the complexity of the power system has been increasing day by day. Traditional power equipment monitoring methods have been difficult to meet the requirements of modern power grids for reliability and efficiency. Due to the lack of real-time monitoring means, power equipment failures can often only be discovered after they occur, which will not only cause huge economic losses but also may have a serious impact on society.

[0003] In recent years, the Internet of Things (IoT) technology has developed rapidly, bringing revolutionary changes to all walks of life. By connecting physical devices to the Internet and using sensors, communication technologies, and data analysis algorithms, real-time monitoring and control of various environmental parameters can be achieved. For the power industry, the application of the Internet of Things has opened a new era of smart grids, making the status monitoring of power equipment more intelligent and efficient.

[0004] The Chinese invention patent with the publication number CN109548032A discloses a distributed cooperative spectrum sensing method for full-band detection in a dense network, and the process is as follows: S1: Group all the frequency points to be detected, and each node is respectively assigned a group of frequency points to be detected, and ensure that all frequency points are completely assigned within the one-hop range of each node; S2: Each node respectively performs energy detection on the assigned frequency points; S3: Obtain the interaction order of all nodes in the whole network; S4: According to the interaction order, interact and share the detected energy information; S5: The node that receives the shared information iterates it with the result of its own previous iteration until all nodes have completed one sharing and finally reach a consistent convergence result, otherwise return to S3; S6: Each node compares the iteration result with the decision threshold to obtain the final frequency usage decision for the global frequency points. This method can effectively meet the requirements of full-band detection in a dense network scenario, and effectively reduce the sensing delay while ensuring high detection accuracy.

[0005] When the above-mentioned and similar detection methods are used to monitor the status of power equipment, most traditional wireless monitoring solutions rely on static frequency band division (such as reserving 8 sub-channels in the 868MHz frequency band). However, in areas with dense power equipment (such as 500kV substations), there is a serious mismatch between the static spectrum allocation mechanism and dynamic electromagnetic interference. At the same time, the transient arc caused by the operation of the disconnector in the substation can instantaneously excite broadband electromagnetic noise in the range of 0.5MHz to 5GHz. Its equivalent isotropic radiated power (EIRP) jumps from -40dBm to +20dBm within 2μs, which will cause the target channel to be submerged in the noise floor. At the same time, due to the lack of dynamic spectrum hole sensing and fast frequency hopping (<10ms switching) capabilities, the signal-to-noise ratio (SNR) of the channel will deteriorate from 20dB to below 3dB, and the node is forced to maintain communication in the blocked channel, resulting in the bit error rate (BER) soaring from 1e⁻ 6 soaring to 1e⁻², and the number of data packet retransmissions exceeding the maximum threshold of the protocol stack (such as the LoRa ADR mechanism is limited to 3 times), ultimately triggering a link interruption. Summary of the Invention

[0006] The purpose of the present invention is to provide an Internet of Things-based power equipment status monitoring system to solve the problems proposed in the above background technology.

[0007] To achieve the above purpose, the present invention provides the following technical solution: An Internet of Things-based power equipment status monitoring system, including:

[0008] A spectrum sensing module, which detects and identifies the noise characteristics in the electromagnetic signal through a quantum spectrum sensing array, and obtains the corresponding spectrum data and classification results, including:

[0009] S1: Through the constructed superconducting resonator, obtain the power amplification factor of the signal after passing through the resonator, and amplify the transient arc noise power obtained, specifically:

[0010]

[0011] Where: is the amplified transient arc noise power, is the original transient arc noise power, is the power amplification factor of the signal after passing through the resonator;

[0012] S2: Take the amplified transient arc noise power as the input of the parameterized quantum circuit, obtain the measurement matrix and the observation vector, construct an optimization problem, and obtain the reconstructed signal according to the constraint conditions of the reconstructed signal;

[0013] S3: Extract the quantum fingerprint features of the reconstructed signal, combine the quantum fingerprint features with the traditional features to obtain a three-dimensional feature vector, and at the same time, train and test the SVW classifier through the three-dimensional feature vector to obtain a classification result;

[0014] The dynamic evaluation module filters out the available channel list through the dynamic channel survivability, and determines the corresponding channels for signal transmission from the available channel list according to the classification result of the SVW classifier;

[0015] The frequency hopping optimization module determines the globally optimal frequency hopping sequence through the available channel list and the total energy between different channels.

[0016] Furthermore, obtaining the reconstructed signal includes:

[0017] S2.1: Obtain the observation vector: Compress the original signal through a superconducting resonator, construct a measurement matrix, and at the same time, compress the amplified transient arc noise power through the measurement matrix to obtain the observation vector;

[0018] S2.2: Obtain the preliminary reconstructed signal: According to the measurement matrix and the observation vector, construct an optimization problem and encode the optimization problem as a quantum Hamiltonian to obtain the preliminary reconstructed signal. Specifically:

[0019]

[0020] Where: is the quantum Hamiltonian, is the i-th preliminary reconstructed signal, is the index of the preliminary reconstructed signal, is the low-dimensional observation signal, is the measurement matrix, is the regularization parameter, is the amplified transient arc noise power, is the preliminary reconstructed signal;

[0021] S2.3: Obtain the final reconstructed signal: According to the qubit resources, block the preliminary reconstructed signal and encode it as a quantum state, and at the same time, minimize the loss function through the quantum state parameters in the quantum state to obtain the final reconstructed signal.

[0022] Furthermore, obtaining the observation vector includes:

[0023] S2.1.1: Obtain the compressed signal dimension: Compress the original signal through the superconducting resonator to obtain the bandwidth compression ratio of the signal, and determine the dimension of the signal compressed by the superconducting resonator according to the bandwidth compression ratio. Specifically:

[0024]

[0025] wherein: is the dimension of the signal compressed by the superconducting resonator, is the dimension size of the original signal, is the bandwidth compression ratio;

[0026] Step SA2.1.2: Construct a measurement matrix: According to the dimension of the signal compressed by the superconducting resonator and the dimension size of the original signal, construct a measurement matrix, specifically:

[0027]

[0028] wherein: is the measurement matrix, is the complex number field, is the dimension of the signal compressed by the superconducting resonator, is the dimension size of the original signal;

[0029] S2.1.3: Determine the observation vector: Compress the amplified transient arc noise power through the measurement matrix to obtain an observation vector, specifically:

[0030]

[0031] wherein: is the low-dimensional observation signal, is the measurement matrix, is the amplified transient arc noise power, is the observation noise.

[0032] Furthermore, obtaining the final reconstructed signal includes:

[0033] S2.3.1: Obtain quantum state parameters: Through the quantum bit resource, block the preliminary reconstructed signal, and map each sub-signal to the quantum state probability amplitude through amplitude encoding, specifically:

[0034]

[0035] wherein: is the parameterized quantum state, is the quantum state parameter, is the j-th component of the preliminary reconstructed signal, is the index of the component of the preliminary reconstructed signal, is the sum of the absolute values of all signal components;

[0036] S2.3.2: Determine the final reconstructed signal: Through the gradient descent method, update the quantum state parameters, and according to the updated quantum state parameters, obtain the minimized loss function. The reconstructed signal corresponding to the minimized loss function is the final reconstructed signal.

[0037] Furthermore, obtaining the classification result includes:

[0038] S3.1: Extract quantum fingerprint features: According to the frequency-domain signal and the coupling signal in the final reconstructed signal, construct a Hamiltonian, and obtain the expectation value of the quantum state under the Hamiltonian. At the same time, determine the magnitude of the adjustable parameter according to the ground state energy;

[0039] S3.2: Classification decision: Combine the ground state energy with the traditional features to obtain a three-dimensional feature vector, and divide the three-dimensional feature vector into a training set and a test set. Train and test the SVW classifier, and at the same time, through the SVW classifier after training and testing, identify the arc operating state.

[0040] Furthermore, determining the magnitude of the adjustable parameter includes:

[0041] S3.1.1: Construct the Hamiltonian: Encode the frequency-domain signal and the coupling signal in the final reconstructed signal into the Hamiltonian of the quantum system, specifically:

[0042]

[0043] Where: is the quantum Hamiltonian, is the total number of frequency components in the final reconstructed signal, is the energy of the g-th frequency component in the final reconstructed signal, is the creation operator, is the annihilation operator, is the interaction strength between the g-th frequency component and the o-th frequency component in the final reconstructed signal, , are the indices of the frequency components in the final reconstructed signal;

[0044] S3.1.2: Prepare the variational quantum state: According to the Hamiltonian, obtain the magnitude of the expectation value of the quantum state under the Hamiltonian, specifically:

[0045]

[0046] Where: is the energy expectation value, is the parameterized quantum state, is the quantum Hamiltonian, is the adjustable parameter;

[0047] S3.1.3: Conduct parameter optimization: Adjust and update the adjustable parameters through the gradient descent method to obtain the ground state energy. The update formula for the adjustable parameters is specifically as follows:

[0048]

[0049] Where: is the adjustable parameter at the (k + 1)-th iteration, is the adjustable parameter at the k-th iteration, is the learning rate, is the expected energy value at the k-th iteration with respect to the adjustable parameter at the k-th iteration gradient.

[0050] Furthermore, determine the corresponding channel for signal transmission, including:

[0051] SB1: Obtain available channels: Fuse the quantum entanglement entropy, electromagnetic field strength, and vibration acceleration to obtain the dynamic channel viability, and compare the dynamic channel viability with a preset viability. Determine the dynamic channel viability not less than the preset viability from them. The dynamic channel viability not less than the preset viability is the available channel. The formula for obtaining the dynamic channel viability is specifically as follows:

[0052]

[0053] Where: is the channel viability, is the normalization constant, is the quantum entanglement entropy, is the threshold entropy, is the adjustment factor;

[0054] SB2: Select channels: According to the available channels and the classification results of the SVW classifier, set the channels of normal arcs as regular channels and the channels of faulty arcs as the available channels.

[0055] Furthermore, optimize the dynamic channel viability, including:

[0056] SB1.1: Obtain the electromagnetic-vibration coupling factor: Use the historical interference power density, the time-series data of the electromagnetic field strength, and the time-series data of the vibration acceleration as the input of the LSTM model, and output to obtain the predicted value of the interference power density at a preset time to obtain the electromagnetic-vibration coupling factor. Specifically:

[0057]

[0058] Where: is the electromagnetic-vibration coupling factor, is the instantaneous amplitude of the electromagnetic field, is the root mean square of the mechanical vibration acceleration, is the integration time window of the interference power density;

[0059] SB1.2: Adjust the dynamic weight: Compare the electromagnetic-vibration coupling factor with a preset coupling factor, and adjust the weight of the quantum entanglement entropy according to the comparison result. Specifically:

[0060] When the electromagnetic-vibration coupling factor is greater than the preset coupling factor, increase the weight of the quantum entanglement entropy through the weight constraint formula; otherwise, keep the weight of the quantum entanglement entropy unchanged;

[0061] SB1.3: Determine the quantum entanglement entropy: Determine the final size of the quantum entanglement entropy according to the adjusted weight of the quantum entanglement entropy, and determine the optimized dynamic channel survival degree according to the final size of the quantum entanglement entropy. The acquisition formula of the final quantum entanglement entropy is specifically:

[0062]

[0063] Where: is the final quantum entanglement entropy, is the weight of the quantum entanglement entropy, is the quantum entanglement entropy.

[0064] Furthermore, determine the global optimal frequency hopping sequence, including:

[0065] SC1: Obtain the coupling strength: According to the channel survival degree corresponding to each available channel, obtain the coupling strength between different channels. Specifically:

[0066]

[0067] Where: is the coupling strength between the h-th channel and the u-th channel, is the scaling factor, is the channel survival degree of the h-th channel, is the channel survival degree of the u-th channel, 、 are the indices of the channels;

[0068] SC2: Obtain the optimal channel combination: According to the coupling strength between different channels, obtain the total energy between channels, and determine the minimum total energy between channels from it. The channel combination corresponding to the minimum total energy between channels is the optimal channel combination. The acquisition formula of the total energy between channels is specifically:

[0069]

[0070] Wherein: is the total energy between the h-th channel and the u-th channel, is the coupling strength between the h-th channel and the u-th channel, is the Pauli Z matrix acting on the h-th channel, is the Pauli Z matrix acting on the u-th channel.

[0071] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0072] First: Through the superconducting resonator and the quantum compressive sensing algorithm, the present invention realizes the capture of microsecond-level transient noise, improves the spectrum hole detection sensitivity by 40 dB, and at the same time, through the dynamic channel survivability, anti-interference channels are screened out in real time, and the optimal channel combination is obtained, enabling the channel switching to be completed within 10 ms;

[0073] Second: Through the superconducting resonator, the present invention realizes non-linear power amplification, through the quantum compressive sensing algorithm, realizes the reconstruction of broadband signals, and through the quantum fingerprint feature extraction technology, reduces the feature dimension from the traditional 256 dimensions to 3 dimensions, thereby increasing the classification accuracy by 12.7%;

[0074] Third: By combining quantum entanglement entropy, electromagnetic field strength and vibration acceleration, the present invention obtains the corresponding channel survivability, and through the dynamic weight adjustment mechanism, makes the channel evaluation misjudgment rate lower than 0.5%. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Figure 1 is a schematic diagram of the monitoring process of the power equipment status monitoring system of the present invention.

[0076] Figure 2 is a comparison diagram of transient noise in the present invention.

[0077] Figure 3 is a performance diagram of the original signal in the present invention.

[0078] Figure 4 is a performance diagram of the reconstructed signal in the present invention.

[0079] Figure 5 is a graph of the change of quantum entanglement entropy with time in the present invention.

[0080] Figure 6 is a graph of the change of channel survivability with time in the present invention.

[0081] Figure 7 is a graph of the survivability threshold baseline in the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0082] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0083] When the existing detection methods monitor the state of power equipment, most traditional wireless monitoring solutions rely on static frequency band division (such as reserving 8 sub-channels in the 868 MHz frequency band). However, in areas with dense power equipment (such as 500 kV substations), there is a serious mismatch between the static spectrum allocation mechanism and dynamic electromagnetic interference. At the same time, the transient arc caused by the operation of the disconnector in the substation can instantaneously generate broadband electromagnetic noise in the range of 0.5 MHz to 5 GHz. Its equivalent isotropic radiated power (EIRP) jumps from -40 dBm to +20 dBm within 2 μs, which will cause the target channel to be submerged in the noise floor. At the same time, due to the lack of dynamic spectrum hole sensing and fast frequency hopping (<10 ms switching) capabilities, the signal-to-noise ratio (SNR) of the channel will deteriorate from 20 dB to below 3 dB, and the node is forced to maintain communication in the blocked channel, resulting in the bit error rate (BER) soaring from 1e⁻ 6 to 1e⁻², and the number of data packet retransmissions exceeds the maximum threshold of the protocol stack (such as the LoRa ADR mechanism is limited to 3 times), ultimately triggering a link interruption. The technical solution of this application amplifies the transient arc noise through a superconducting resonator, reconstructs the signal through quantum compressive sensing, and jointly trains the SVW classifier with quantum fingerprint features and traditional features for state classification. At the same time, according to the classification results, it dynamically selects dedicated channels for normal / fault arcs, and based on the channel survivability, obtains the coupling strength between channels, and determines the globally optimal frequency hopping sequence from it to effectively solve the broadband electromagnetic interference problem caused by transient arcs.

[0084] Embodiment 1

[0085] Refer to Figures 1 - 7 , this embodiment provides an Internet of Things-based power equipment status monitoring system. The power equipment status monitoring system includes a spectrum sensing module, a dynamic evaluation module, a frequency hopping optimization module, and a damage-resistant transmission module. Among them, the spectrum sensing module detects and identifies the noise characteristics in the electromagnetic signal through a quantum spectrum sensing array to obtain the corresponding spectrum data and corresponding classification results. The dynamic evaluation module establishes a calculation formula for dynamic channel survivability based on quantum entanglement entropy, electromagnetic field strength, and vibration acceleration, and uses the spectrum data and classification results as input data for dynamic channel survivability to screen out a list of available channels. The frequency hopping optimization module obtains the quantum state of the channel according to the list of available channels and the channel switching cost parameter, and obtains the globally optimal frequency hopping sequence through the quantum state of the channel.

[0086] In this embodiment, the spectrum sensing module performs corresponding amplification processing on the input signal through the provided superconducting resonator. At the same time, the amplified signal is mapped to the quantum state amplitude and verified through the reconstruction error. Further, the verified signal is classified through the SVW classifier to determine whether the input signal is a fault arc. Specifically as follows:

[0087] Step SA1: Perform quantum state amplification. That is, through the constructed superconducting resonator, obtain the quality factor of the resonator, and determine the power amplification factor of the signal after passing through the resonator through the quality factor of the resonator. Specifically:

[0088]

[0089] Where: is the power amplification factor of the signal after passing through the resonator, is the quality factor of the resonator, is the free space impedance, is the external coupling impedance, is the natural oscillation frequency of the resonator, is the frequency width at half of the maximum power of the resonance peak.

[0090] In this embodiment, a cylindrical cavity (radius 50 mm, height 100 mm) is constructed with niobium superconducting material, and the surface of the cavity is polished to a surface roughness of not more than 0.1 μm. At the same time, the temperature of the cavity is stabilized at 4K through a liquid helium closed-loop refrigeration system to ensure the stability of the superconducting state, and a magnetic shielding layer is wound with niobium-titanium alloy superconducting wire to suppress external magnetic field interference.

[0091] Further, the noise characteristics (transient arc noise) in the electromagnetic signal are amplified through the constructed superconducting resonator. In this embodiment, the power of the obtained transient arc noise is amplified through the power amplification factor of the signal after passing through the resonator. Specifically:

[0092]

[0093] Where: is the amplified power of the transient arc noise, is the original power of the transient arc noise, is the power amplification factor of the signal after passing through the resonator.

[0094] Step SA2: Perform sparse signal reconstruction. That is, use the amplified transient arc noise power obtained in Step SA1 as the input of the parameterized quantum circuit to obtain the measurement matrix and the observation vector, thereby constructing the corresponding optimization problem, and at the same time combining the constraint conditions of the reconstructed signal to obtain the reconstructed signal. Specifically as follows:

[0095] Step SA2.1: Obtain the observation vector. That is, compress the input original signal through the superconducting resonator constructed in Step SA1, and construct the measurement matrix according to the compressed signal. At the same time, compress the amplified transient arc noise power obtained in Step SA1 through the measurement matrix to obtain the observation vector. Specifically as follows:

[0096] Step SA2.1.1: Obtain the dimension of the compressed signal. That is, compress the input original signal through the superconducting resonator constructed in Step SA1 to obtain the signal bandwidth and bandwidth compression ratio of the compressed signal. Specifically:

[0097]

[0098] Where: is the bandwidth compression ratio, is the bandwidth of the original signal, is the signal bandwidth after compression by the superconducting resonator.

[0099] Furthermore, determine the dimension of the signal compressed by the superconducting resonator according to the dimension size of the original signal and the obtained bandwidth compression ratio. Specifically:

[0100]

[0101] Where: is the dimension of the signal compressed by the superconducting resonator, is the dimension size of the original signal, is the bandwidth compression ratio.

[0102] Step SA2.1.2: Construct the measurement matrix. That is, construct the measurement matrix according to the dimension of the signal compressed by the superconducting resonator determined in Step SA2.1.1 and the dimension size of the original signal. Specifically:

[0103]

[0104] Where: is the measurement matrix, is the complex number field, is the dimension of the signal compressed by the superconducting resonator, is the dimension size of the original signal.

[0105] Step SA2.1.3: Determine the observation vector. That is, through the measurement matrix constructed in step SA2.1.2, compress the amplified transient arc noise power obtained in step SA1, that is, compress it into a low-dimensional measurement value to obtain the observation vector, specifically:

[0106]

[0107] Where: is the low-dimensional observation signal, is the measurement matrix, is the amplified transient arc noise power, is the observation noise.

[0108] Step SA2.2: Obtain the preliminary reconstruction signal. That is, according to the measurement matrix constructed in step SA2.1.2 and the observation vector obtained in step SA2.1.3, construct an optimization problem, specifically:

[0109]

[0110] Where: is the preliminary reconstruction signal, is the low-dimensional observation signal, is the measurement matrix.

[0111] Furthermore, through the quantum approximate optimization algorithm, encode the optimization problem into a quantum Hamiltonian, specifically:

[0112]

[0113] Where: is the quantum Hamiltonian, is the i-th preliminary reconstruction signal, is the index of the preliminary reconstruction signal, is the low-dimensional observation signal, is the measurement matrix, is the regularization parameter.

[0114] It should be noted that the preliminary reconstruction signal obtained through the quantum Hamiltonian satisfies the following constraint conditions, specifically:

[0115]

[0116] Where: is the amplified transient arc noise power, is the preliminary reconstruction signal.

[0117] Step SA2.3: Obtain the final reconstructed signal. That is, according to the qubit resources, the preliminary reconstructed signal obtained in step SA2.2 is divided into blocks and encoded into quantum states. At the same time, the loss function is minimized through the quantum state parameters in the quantum states corresponding to the preliminary reconstructed signal to obtain the final reconstructed signal. Specifically as follows:

[0118] Step SA2.3.1: Obtain the quantum state parameters. That is, through the qubit resource size (block-encoded signal), the preliminary reconstructed signal obtained in step SA2.2 is divided into blocks, that is, divided into multiple sub-signals. At the same time, each sub-signal is mapped to the quantum state probability amplitude through amplitude encoding, specifically as follows:

[0119]

[0120] Where: is the parameterized quantum state, is the quantum state parameter, is the j-th component of the preliminary reconstructed signal, is the index of the components of the preliminary reconstructed signal, is the sum of the absolute values of all signal components.

[0121] Step SA2.3.2: Determine the final reconstructed signal. That is, through the gradient descent method, the quantum state parameters obtained in step SA2.3.1 are updated, and according to the updated quantum state parameters, the minimized loss function is obtained. The reconstructed signal corresponding to this minimized loss function is the final reconstructed signal.

[0122] Furthermore, the loss function in this embodiment is specifically:

[0123]

[0124] Where: is the loss function, is the quantum state parameter, is the regularization parameter, is the parameterized preliminary reconstructed signal, is the low-dimensional observation signal, is the measurement matrix, is the data fitting term, is the sum of the absolute values of all signal components.

[0125] Specifically, when the loss function in this embodiment is not greater than 0.1, the reconstructed signal corresponding to this loss function not greater than 0.1 is the final reconstructed signal.

[0126] Step SA3: Perform quantum fingerprint recognition. That is, extract the quantum fingerprint features from the final reconstructed signal, and combine the quantum fingerprint features with the traditional features (pulse width and peak amplitude) to obtain a three-dimensional feature vector for training and testing the SVW classifier. That is to say, through the SVW classifier after training and testing, obtain the corresponding classification results (normal arc and faulty arc). Specifically as follows:

[0127] Step SA3.1: Extract quantum fingerprint features. That is, according to the frequency-domain signal and coupling signal in the final reconstructed signal determined in step SA2.3.2, construct a Hamiltonian, and obtain the expected value of the quantum state under the Hamiltonian. At the same time, according to the obtained minimum expected value, determine the corresponding adjustable parameter size. Specifically as follows:

[0128] Step SA3.1.1: Construct a Hamiltonian. That is, extract the frequency-domain energy and coupling strength of the final reconstructed signal from step SA2.3.2, and encode them as the Hamiltonian of the quantum system. Specifically:

[0129]

[0130] Where: is the quantum Hamiltonian, is the total number of frequency components in the final reconstructed signal, is the energy of the g-th frequency component of the final reconstructed signal, is the creation operator, is the annihilation operator, is the interaction strength between the g-th frequency component and the o-th frequency component of the final reconstructed signal, 、 are the indices of the frequency components in the final reconstructed signal.

[0131] Step SA3.1.2: Prepare a variational quantum state. That is, according to the Hamiltonian of the quantum system obtained in step SA3.1, combine it with 3 qubits to obtain the expected value of the quantum state under the Hamiltonian. Specifically:

[0132]

[0133] Where: is the energy expectation value, is the parameterized quantum state, is the quantum Hamiltonian, is the adjustable parameter.

[0134] Step SA3.1.3: Perform parameter optimization. That is, through the gradient descent method, continuously adjust and update the adjustable parameters to obtain the minimum energy expectation value, i.e., the ground state energy. Further, the update formula for the adjustable parameters is specifically:

[0135]

[0136] Where: is the adjustable parameter at the (k + 1)-th iteration, is the adjustable parameter at the k-th iteration, is the learning rate, is the energy expectation value at the k-th iteration with respect to the adjustable parameter at the k-th iteration gradient.

[0137] Step SA3.2: Classification decision. That is, merge the ground state energy obtained in step SA3.1.3 with traditional features to obtain a three-dimensional feature vector, specifically:

[0138]

[0139] Where: is the multi-dimensional feature vector, is the ground state energy, is the pulse width in the time-domain signal of the final reconstructed signal, is the peak amplitude in the time-domain signal of the final reconstructed signal.

[0140] Further, train the SVW classifier with the obtained multiple three-dimensional feature vectors. Specifically, divide the multiple three-dimensional feature vectors, where 80% of the data is used as the training set and 20% of the data is used as the test set to perform relevant training and testing on the SVW classifier. That is, according to the trained and tested SVW classifier, the corresponding classification result can be obtained, that is, effectively distinguish normal arcs and fault arcs.

[0141] In this embodiment, the dynamic evaluation module determines the dynamic channel survivability according to the quantum entanglement entropy, electromagnetic field strength, and vibration acceleration to filter out the available channel list. At the same time, take the spectrum data and the arc classification result output by the SVW classifier in step SA3.2 as the input data of the dynamic channel survivability, and select the corresponding channel from the filtered available channel list for signal transmission. Specifically as follows:

[0142] Step SB1: Obtain available channels. That is, fuse the quantum entanglement entropy, electromagnetic field strength, and vibration acceleration to obtain the dynamic channel survival degree, compare the dynamic channel survival degree with the preset survival degree, and determine the dynamic channel survival degree not less than the preset survival degree from them. The channel corresponding to the determined dynamic channel survival degree is the available channel.

[0143] In this embodiment, the formula for obtaining the dynamic channel survival degree is specifically:

[0144]

[0145] Where: is the channel survival degree, is the normalization constant, is the quantum entanglement entropy, is the threshold entropy, is the adjustment factor.

[0146] In the process of specific implementation, the threshold entropy is set to 1.0, the normalization constant is set to 1.0, the adjustment factor is set to 5.0, and the quantum entanglement entropy corresponding to the normal channel is 0.8, that is, its corresponding channel survival degree is 0.73. It should be noted that the preset survival degree set in this embodiment is 0.8. That is to say, the channel survival degree (0.73) obtained in this embodiment is less than the preset survival degree (0.8). That is, the channel corresponding to the channel survival degree in this embodiment is not retained, that is, it is not an available channel.

[0147] Step SB2: Select channels. That is, according to the arc classification result output by the SVW classifier in step SA3.2 and the available channels obtained in step SB1, set the channel of the normal arc as the regular channel, and the regular channel is the channel corresponding to the channel survival degree not less than the regular survival degree. At the same time, set the channel of the faulty arc as the available channel obtained in step SB1, that is, the channel corresponding to the channel survival degree not less than the preset survival degree.

[0148] In the process of specific implementation, when the arc classification result output by the SVW classifier includes both normal arcs and faulty arcs. Specifically, during the transmission of a normal arc, it is transmitted through the regular channel. It should be noted that the regular survival degree is set to 0.6 in this embodiment. That is to say, any channel corresponding to a channel survival degree not less than 0.6 can be used to transmit normal arcs. Further, during the transmission of a faulty arc, it is transmitted through the available channels obtained in step SB1. That is to say, any channel corresponding to a channel survival degree not less than 0.8 can be used to transmit faulty arcs.

[0149] In this embodiment, the frequency hopping optimization module determines the global optimal frequency hopping sequence according to the available channel list determined in step SB2 and obtains the total energy between different channels. Specifically as follows:

[0150] Step SC1: Obtain the coupling strength. That is, according to the channel survivability corresponding to each available channel determined in step SB2, obtain the coupling strength between different channels. Specifically:

[0151]

[0152] Where: is the coupling strength between the h-th channel and the u-th channel, is the scaling factor, is the channel survivability of the h-th channel, is the channel survivability of the u-th channel, 、 are the indices of the channels.

[0153] In the process of specific implementation, the channel survivability of the first channel is 0.92, the channel survivability of the second channel is 0.88, and the scaling factor is set to 10. Then the coupling strength between the first channel and the second channel is 8.1.

[0154] Step SC2: Obtain the optimal channel combination. That is, determine the total energy between channels through the coupling strength between different channels obtained in step SC1. Specifically:

[0155]

[0156] Where: is the total energy between the h-th channel and the u-th channel, is the coupling strength between the h-th channel and the u-th channel, is the Pauli Z matrix acting on the h-th channel, is the Pauli Z matrix acting on the u-th channel.

[0157] In the process of specific implementation, the channel survivability of the first channel is 0.92, the channel survivability of the third channel is 0.90, and the channel survivability of the fifth channel is 0.88. Then the coupling strength between the first channel and the third channel is 8.28, the coupling strength between the first channel and the third channel is 8.1, and the coupling strength between the third channel and the fifth channel is 7. Therefore, the total energy corresponding to these three coupling strengths is -24.3.

[0158] Furthermore, according to the total energy determined between channels, the lowest energy state is determined from the total energy between all channels, and the channel combination corresponding to this lowest energy state is the optimal channel combination.

[0159] Reference Figure 2 , Figure 2 is the comparison graph of the transient noise in this embodiment. It can be seen from Figure 2 that the peak power of the input signal is +20 dBm. After being amplified by the superconducting resonator cavity, its peak power is increased to +60 dBm. At the same time, the Gaussian pulse shape of the input signal (centered at 0.5 μs and the pulse width is controlled by the standard deviation of 0.1 μs) is completely retained after amplification. Therefore, during the amplification process, no waveform distortion is introduced. Furthermore, the rise time of the input signal (from -40 dBm to +20 dBm) is about 0.2 μs, and the rise time of the output signal is the same as it.

[0160] Reference Figure 3 and Figure 4 , Figure 3 is the performance graph of the original signal in this embodiment, Figure 4 is the performance graph of the reconstructed signal in this embodiment. It can be seen from Figure 3 and Figure 4 that the original signal has the sparse characteristics of a typical Laplace distribution: a small number of large-amplitude spikes (sparse components) and a large number of small-amplitude noises close to zero. The reconstructed signal accurately captures all significant spikes, but there is a smoothing effect on the tiny components. The maximum local error appears at the signal mutation point (such as the spike falling edge near index 500), and the error amplitude is about 0.3, corresponding to a relative error of 15%.

[0161] Reference Figures 5 - 7 , Figure 5 is the graph of the quantum entanglement entropy varying with time in this embodiment, Figure 6 is the graph of the channel survivability varying with time in this embodiment, Figure 7 is the graph of the survivability threshold baseline in this embodiment. It can be seen from Figures 5 - 7 that the random uniform distribution of the quantum entanglement entropy simulates the dynamic fluctuation characteristics of the quantum channel in the actual environment, covering the typical working condition range (0.6 - 0.9). The channel survivability shows the characteristics of a Sigmoid curve near the threshold of 7 for the quantum entanglement entropy, meeting the probabilistic evaluation requirements of the channel state. At the same time, the normal working condition accounts for 82%, and the abnormal working condition accounts for 18%, which is consistent with the statistics of the failure rate of power equipment. The dotted line threshold, as the early warning threshold, can identify potential channel deterioration 5 - 10 ms in advance (such as when the quantum entanglement entropy is close to 0.7 at 30 ms and the channel survivability begins to decline).

[0162] Embodiment 2

[0163] This embodiment provides a power equipment status monitoring system based on the Internet of Things. The specific implementation method is the same as that of Embodiment 1, except that the dynamic channel survivability is compared with a preset survivability, and the dynamic channel survivability not less than the preset survivability is determined. The channel corresponding to the determined dynamic channel survivability is the available channel. The present invention will be illustrated by way of the specific implementation manner of this embodiment below.

[0164] In this embodiment, the interference power density within a preset time is obtained through the LSTM model, and the weight of the quantum entanglement entropy is dynamically adjusted based on the interference power density within the preset time to optimize the initially obtained available channels. Specifically as follows:

[0165] Step SB1.1: Obtain the electromagnetic-vibration coupling factor. That is, the historical interference power density (for example, the time window is set to 100 ms and the sampling interval is set to 1 ms), the time-series data of the electromagnetic field intensity, and the time-series data of the vibration acceleration are used as the inputs of the LSTM model, and the predicted value of the interference power density for the preset time is output.

[0166] In the process of specific implementation, the time-series data of the electromagnetic field intensity is set to [48, 49, 50] V / m, and the time-series data of the vibration acceleration is set to [0.18, 0.19, 0.20] m / s 2 , and the interference power density obtained through the LSTM model for the preset time is 120 μW / Hz.

[0167] Furthermore, the electromagnetic-vibration coupling factor is obtained through the predicted value of the interference power density for the preset time, specifically:

[0168]

[0169] Where: is the electromagnetic-vibration coupling factor, is the instantaneous amplitude of the electromagnetic field, is the root mean square of the mechanical vibration acceleration, is the integration time window of the interference power density.

[0170] In the process of specific implementation, when in the normal working condition, the instantaneous amplitude of the electromagnetic field is 50 V / m, the root mean square of the mechanical vibration acceleration is 0.2 m / s 2 , and the integration time window of the interference power density is 0.1 s. Therefore, the electromagnetic-vibration coupling factor is 31.6. Furthermore, when in the strong interference working condition, the instantaneous amplitude of the electromagnetic field is 200 V / m, the root mean square of the mechanical vibration acceleration is 0.5 m / s 2 , and the integration time window of the interference power density is 0.1 s. Therefore, the electromagnetic-vibration coupling factor is 316.

[0171] Step SB1.2: Adjust the dynamic weight. That is, compare the electromagnetic-vibration coupling factor obtained in step SB1.1 with the preset coupling factor, and adjust the weight of the quantum entanglement entropy according to the comparison result. Specifically:

[0172] When the obtained electromagnetic-vibration coupling factor is greater than the preset coupling factor, increase the weight of the quantum entanglement entropy; otherwise, keep the weight of the quantum entanglement entropy unchanged.

[0173] Furthermore, during the process of adjusting the weight of the quantum entanglement entropy, the magnitude of the corresponding weight can be adjusted according to actual needs. It should be noted that during the process of weight adjustment, the weight satisfies the following constraint formula, which is specifically:

[0174]

[0175] Where: is the weight of the quantum entanglement entropy, is the weight of the instantaneous amplitude of the electromagnetic field, is the weight of the mechanical vibration acceleration.

[0176] Step SB1.3: Determine the quantum entanglement entropy. That is, determine the final magnitude of the quantum entanglement entropy according to the weight of the quantum entanglement entropy determined in step SB1.2. Specifically:

[0177]

[0178] Where: is the final quantum entanglement entropy, is the weight of the quantum entanglement entropy, is the quantum entanglement entropy.

[0179] During the specific implementation process, the threshold entropy is set to 1.0, the normalization constant is set to 1.0, the adjustment factor is set to 5.0, and the quantum entanglement entropy corresponding to the normal channel is 0.8, that is, the corresponding channel survival degree is 0.73. It should be noted that the preset survival degree in this embodiment is 0.8. That is to say, the obtained channel survival degree (0.73) in this embodiment is less than the preset survival degree (0.8), that is, the channel corresponding to the channel survival degree in this embodiment is not retained, that is, it is not an available channel.

[0180] Furthermore, in this embodiment, the weight of the quantum entanglement entropy is set to 0.75, that is, the final quantum entanglement entropy is 0.6, so the corresponding channel survival degree is 0.88. That is to say, the obtained channel survival degree (0.88) in this embodiment is not less than the preset survival degree (0.8), that is, the channel corresponding to the channel survival degree in this embodiment will be retained, that is, it is an available channel.

[0181] Although embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended embodiments and their equivalents.

Claims

1. An Internet of Things-based power equipment status monitoring system, characterized in that, It includes: A spectrum sensing module that detects and identifies the noise characteristics in electromagnetic signals through a quantum spectrum sensing array, and obtains corresponding spectrum data and classification results, including: S1: Through the constructed superconducting resonator, obtain the power amplification factor of the signal after passing through the resonator, and amplify the obtained transient arc noise power signal. Specifically: ； Wherein: is the amplified transient arc noise power, is the original transient arc noise power, is the power amplification factor after the signal passes through the resonant cavity; S2: Use the amplified transient arc noise power as the input of a parameterized quantum circuit, obtain a measurement matrix and an observation vector, construct an optimization problem, and obtain a reconstructed signal according to the constraint conditions of the reconstructed signal. S3: Extract the quantum fingerprint features of the reconstructed signal, combine the quantum fingerprint features with traditional features to obtain a three-dimensional feature vector, and use the three-dimensional feature vector to train and test an SVW classifier to obtain a classification result. A dynamic evaluation module that screens out a list of available channels through dynamic channel viability, and determines corresponding channels for signal transmission from the list of available channels according to the classification result of the SVW classifier. A frequency hopping optimization module that determines a globally optimal frequency hopping sequence through the list of available channels and the total energy between different channels.

2. The power equipment status monitoring system based on the Internet of Things according to claim 1, characterized in that Obtaining the reconstructed signal includes: S2.1: Obtain the observation vector: Compress the original signal through a superconducting resonator, construct a measurement matrix, and compress the amplified transient arc noise power through the measurement matrix to obtain the observation vector. S2.2: Obtain a preliminary reconstructed signal: According to the measurement matrix and the observation vector, construct an optimization problem, encode the optimization problem as a quantum Hamiltonian, and obtain a preliminary reconstructed signal. Specifically: ； Wherein: is the quantum Hamiltonian,[[]] is the i-th preliminary reconstruction signal,[[]] is the index of the preliminary reconstruction signal,[[]] is the low-dimensional observation signal,[[]] is the measurement matrix,[[]] is the regularization parameter,[[]] is the amplified transient arc noise power,[[]] is the preliminary reconstruction signal;[[]] S2.3: Obtain the final reconstructed signal: According to the qubit resources, divide the preliminary reconstructed signal into blocks and encode it into a quantum state. At the same time, minimize the loss function through the quantum state parameters in the quantum state to obtain the final reconstructed signal.

3. The power equipment status monitoring system based on the Internet of Things according to claim 2, characterized in that, Obtaining the observation vector includes: S2.1.1: Obtain the dimension of the compressed signal: Compress the original signal through the superconducting resonator, obtain the bandwidth compression ratio of the signal, and determine the dimension of the signal compressed by the superconducting resonator according to the bandwidth compression ratio. Specifically: ； Wherein: is the dimension of the signal compressed by the superconducting resonator cavity, is the dimension size of the original signal, is the bandwidth compression ratio; Step SA2.1.2: Construct a measurement matrix: Construct a measurement matrix according to the dimension of the signal compressed by the superconducting resonator and the dimension of the original signal. Specifically: ； Wherein: is a measurement matrix, is the complex number field, is the signal dimension after being compressed by a superconducting resonator, is the dimension size of the original signal; S2.1.3: Determine the observation vector: Compress the amplified transient arc noise power through the measurement matrix to obtain the observation vector. Specifically: ； Wherein: is a low-dimensional observation signal, is a measurement matrix, is the amplified transient arc noise power, is the observation noise.

4. An Internet of Things-based power equipment status monitoring system according to claim 2, characterized in that Obtaining the final reconstructed signal includes: S2.3.1: Obtain quantum state parameters: Divide the preliminary reconstructed signal into blocks according to qubit resources, and map each sub-signal to a quantum state probability amplitude through amplitude encoding. Specifically: ； Wherein: is a parameterized quantum state, is a quantum state parameter, is the j-th component of the preliminary reconstruction signal, is the index of the component of the preliminary reconstruction signal, is the sum of the absolute values of all signal components; S2.3.2: Determine the final reconstructed signal: Update the quantum state parameters through the gradient descent method, and obtain the minimized loss function according to the updated quantum state parameters. The reconstructed signal corresponding to the minimized loss function is the final reconstructed signal.

5. An Internet of Things-based power equipment status monitoring system according to claim 1, characterized in that, Obtain classification results, including: S3.1: Extract quantum fingerprint features: Based on the frequency-domain signal and coupling signal in the final reconstructed signal, construct a Hamiltonian, and obtain the expectation value of the quantum state under the Hamiltonian. At the same time, determine the magnitude of the adjustable parameter according to the ground-state energy; S3.2: Classification decision: Combine the ground-state energy with traditional features to obtain a three-dimensional feature vector. Divide the three-dimensional feature vector into a training set and a test set, train and test the SVW classifier, and identify the arc operating state through the trained and tested SVW classifier.

6. The power equipment status monitoring system based on the Internet of Things according to claim 5, characterized in that, Determine the magnitude of the adjustable parameter, including: S3.1.1: Construct the Hamiltonian: Encode the frequency-domain signal and coupling signal in the final reconstructed signal into the Hamiltonian of the quantum system, specifically: ； Wherein: is the quantum Hamiltonian, is the total number of frequency components in the final reconstructed signal, is the energy of the g-th frequency component in the final reconstructed signal, is the creation operator, is the annihilation operator, is the interaction strength between the g-th and o-th frequency components in the final reconstructed signal, and are the indices of the frequency components in the final reconstructed signal; S3.1.2: Prepare the variational quantum state: According to the Hamiltonian, obtain the magnitude of the expectation value of the quantum state under the Hamiltonian, specifically: ； Wherein: is the energy expectation value, is the parameterized quantum state, is the quantum Hamiltonian, is the adjustable parameter; S3.1.3: Perform parameter optimization: Adjust and update the adjustable parameter by the gradient descent method to obtain the ground-state energy. The update formula of the adjustable parameter is specifically: ； Wherein: is the adjustable parameter at the (k + 1)-th iteration, is the adjustable parameter at the k-th iteration, is the learning rate, is the expected value of the energy at the k-th iteration with respect to the adjustable parameter at the k-th iteration gradient.

7. The power equipment status monitoring system based on the Internet of Things according to claim 1, characterized in that Determine the corresponding channel for signal transmission, including: SB1: Obtain available channels: Fuse the quantum entanglement entropy, electromagnetic field strength, and vibration acceleration to obtain the dynamic channel survivability. Compare the dynamic channel survivability with the preset survivability, and determine the dynamic channel survivability not less than the preset survivability from them. The dynamic channel survivability not less than the preset survivability is the available channel. The acquisition formula of the dynamic channel survivability is specifically: ； Wherein: is the channel survival degree, is the normalization constant, is the quantum entanglement entropy, is the threshold entropy, is the adjustment factor; SB2: Select channels: According to the available channels and the classification results of the SVW classifier, set the channels of normal arcs as conventional channels and the channels of faulty arcs as the available channels.

8. An Internet of Things-based power equipment status monitoring system according to claim 7, characterized in that, Optimize the dynamic channel survivability, including: SB1.1: Obtain the electromagnetic-vibration coupling factor: Use the historical interference power density, the time-series data of the electromagnetic field strength, and the time-series data of the vibration acceleration as the input of the LSTM model, and output the predicted value of the interference power density at a preset time to obtain the electromagnetic-vibration coupling factor, specifically: ； Wherein: is the electromagnetic-vibration coupling factor, is the instantaneous amplitude of the electromagnetic field, is the root mean square of the mechanical vibration acceleration, is the integration time window of the interference power density; SB1.2: Adjust the dynamic weight: Compare the electromagnetic-vibration coupling factor with the preset coupling factor, and adjust the weight of the quantum entanglement entropy according to the comparison result, specifically: When the electromagnetic-vibration coupling factor is greater than the preset coupling factor, increase the weight of the quantum entanglement entropy through the weight constraint formula. Otherwise, keep the weight of the quantum entanglement entropy unchanged; SB1.3: Determine the quantum entanglement entropy: Determine the final magnitude of the quantum entanglement entropy according to the adjusted weight of the quantum entanglement entropy, and determine the optimized dynamic channel survivability according to the final magnitude of the quantum entanglement entropy. The acquisition formula of the final quantum entanglement entropy is specifically: ； Wherein: is the final quantum entanglement entropy, is the weight of the quantum entanglement entropy, is the quantum entanglement entropy.

9. The power equipment status monitoring system based on the Internet of Things according to claim 1, characterized in that Determine the global optimal frequency-hopping sequence, including: SC1: Obtain the coupling strength: According to the channel survivability corresponding to each available channel, obtain the coupling strength between different channels, specifically: ； Wherein: is the coupling strength between the h-th channel and the u-th channel, is the scaling factor, is the channel viability of the h-th channel, is the channel viability of the u-th channel, and are the indices of the channels; SC2: Obtain the optimal channel combination: According to the coupling strength between different channels, obtain the total energy between channels, and determine the minimum total energy between channels therefrom. The channel combination corresponding to the minimum total energy between channels is the optimal channel combination. The specific formula for obtaining the total energy between channels is as follows: ； Wherein: is the total energy between the h-th channel and the u-th channel, is the coupling strength between the h-th channel and the u-th channel, is the Pauli Z matrix acting on the h-th channel, is the Pauli Z matrix acting on the u-th channel.

Images