61 results about "Phasor measurement unit" patented technology

The invention provides a power system voltage control method and device based on dynamic steady state collaborative correction, and the method comprises the steps: building an electromagnetic transient model for a preset voltage node in electromagnetic transient simulation software PSCAD, simplifying a power grid with the electrical distance to a target region exceeding a preset distance into a dynamic Thevenin equivalent model, initializing load flow calculation in the BPA to obtain a dynamic-steady state hybrid model; updating dynamic Thevenin equivalent parameters of the power system in real time according to the data of the phasor measurement unit of the power system, and synchronizing the dynamic Thevenin equivalent parameters to the BPA; based on the dynamic-steady state hybrid model, the transient overvoltage of the power system is suppressed within the time shorter than the first preset duration, and the steady state voltage of the power system is optimized within the time shorter than the second preset duration in combination with wide-area coordination control, so that the parameter bidirectional correction of the dynamic and steady state models is realized, and the head end voltage prediction error is reduced. And the voltage control robustness is improved.

The invention relates to the technical field of safety and stability control of a power distribution network, in particular to a regional power distribution network collection point voltage reactive coupling characteristic evaluation method. The method comprises the following steps: deploying a synchronous phasor measurement unit for a plurality of transformer area convergent points, collecting electrical parameters and generating original state data of the convergent points; performing quality verification, feature extraction and time sequence reconstruction on the original state data to obtain a convergence point key feature parameter set; based on the key characteristic parameter set, an inverter group initial model is constructed through inverter operation characteristic clustering and control link characteristic analysis; performing electromagnetic transient-steady state hybrid simulation on the inverter group based on the model to generate a transient-steady state hybrid simulation scheme; on the basis of the scheme, through multi-working-condition interactive simulation and interactive gain calculation, a key coupling path of a collection point is mined. According to the method, the dynamic interaction influence between the inverter group and the power grid can be accurately captured, and the resonance path and the reactive coupling path can be effectively identified.

The invention discloses an intelligent peak regulation method with optimized combination of energy storage and conventional peak regulation means, and the method comprises the following steps: building a holographic situation awareness layer: capturing the voltage, current phase angle and frequency deviation of a power grid, and monitoring the pressure fluctuation in a diesel engine cylinder; data transmission and preprocessing: encrypting data of the miniature phasor measurement unit and the quantum sensor by using a quantum key distribution technology; the self-adaptive decision engine is used for performing Nash equilibrium calculation by utilizing a federal reinforcement learning architecture, through a local Agent based on a power generation / energy storage unit state autonomous optimization control strategy and in combination with a digital twin platform, solving multi-target conflicts of economy, reliability and carbon emission, and generating a global optimal scheduling instruction; the execution control layer is used for tracking resonance points and dispersing harmonic energy spectrums; dynamic topology reconstruction: a harmonic source is positioned by using a digital twin platform, and the topology of the energy storage system is automatically switched according to the harmonic frequency and position; continuously optimizing the digital twin platform; and full-process closed-loop feedback and self-learning are carried out.

The invention discloses an identification method and device for sub-hyper-synchronous oscillation parameters of a power system, a terminal and a medium, and the method comprises the steps: firstly collecting the electric quantity sampling data of the power system through a broadband synchronous phasor measurement unit, and constructing a superposition model comprising a fundamental component and a plurality of oscillation modes; then, calculating a residual error between an actually measured signal and a model predicted value, constructing a nonlinear least square objective function, and generating a parameter updating vector in combination with a dynamic damping factor; in the iterative optimization process, the size of a damping factor is dynamically adjusted by comparing the change of an objective function value so as to adapt to the damping factor numerical value requirement of the current iterative stage, and finally, an optimized oscillation parameter is output as a final identification result. According to the scheme, through a dynamic damping factor adjustment mechanism, a large damping factor is adopted in the initial stage of iteration to inhibit matrix ill-conditioned influence, the damping factor is reduced when convergence is close to improve the convergence speed, and the stability and efficiency of the optimization process are effectively balanced.

The invention discloses a low-cost precise time keeping method, system and device based on a gating circulation unit, and relates to the technical field of power grid measurement. The method is executed when the synchronous phasor measurement device cannot receive external signals, and comprises the following steps: monitoring operation parameters which influence the frequency stability of an internal crystal oscillator of the phasor measurement unit in real time; calculating a current frequency error of the crystal oscillator in real time by using a crystal oscillator frequency drift prediction model based on a gated cycle unit network based on the operation parameters; generating a time correction amount according to the current frequency error; and applying the time correction to a local clock of the phasor measurement unit to compensate the frequency drift of the crystal oscillator so as to realize accurate time keeping. According to the invention, through accurate prediction of crystal oscillator errors and lightweight software compensation, the synchronous phasor measurement device adopting a low-cost crystal oscillator can also realize long-time high-precision time keeping, the hardware cost is obviously reduced, and the reliability of power grid monitoring during the Beidou / GPS lock losing period is improved.

A device and method for fault management for an electric power distribution network incorporating intermittent generation sources. The method involves configuring multiple hyperparameters for a series of Convolutional Neural Networks (CNNs). A first CNN is trained using current signal imagery from phasor measuring units (PMUs) during fault conditions to classify faults. A second CNN is trained with signal images from PMUs captured during pre-fault and fault cycles for identifying fault sections. Similarly, a third CNN is trained using these images to determine the exact fault location. Once trained, the CNNs are employed sequentially. The first CNN classifies the fault, the second detects the fault section, and the third ascertains the fault location. Subsequently, a comprehensive fault management strategy is deployed.

The invention discloses a distribution network overhead line fault self-healing method, system and device and a medium, and belongs to the technical field of power system fault self-healing, and the method comprises the steps: laying a synchronous phasor measurement unit, and generating a unified phasor reference through cloud edge cooperative calculation; the edge coordination engine calculates the load flow change after loop closing by means of a unified phasor reference and an instantaneous topological state to form an action sequence; the local controller obtains a coordination coefficient and controls the pre-compensation of the flexible interconnection device and the closing time of the section switch; pushing the temperature rise prediction value to a wire thermal model; and the edge controller circularly and adaptively corrects the compensation parameter of the flexible interconnection device, and writes the event into the self-healing knowledge base. According to the invention, automation and intelligentization of fault self-healing are realized, the efficiency of fault self-healing is significantly improved, and long-term stable operation of the power distribution network after fault recovery is ensured; and through intelligent and adaptive design, a brand new technical path is provided for long-term reliable operation of the distribution network.

The invention relates to the technical field of power system recovery control, in particular to a power grid black-start system and method based on cascade hydroelectric characteristics, and the method comprises the steps: constructing a software and hardware integrated system integrating physical equipment and intelligent control; situation awareness is carried out through the software and hardware integrated system, a power failure range after large-area power failure of a power grid is automatically identified, available cascade hydropower stations are positioned, and a black start plan is automatically triggered; the central decision-making unit simultaneously instructs a plurality of upstream hydropower stations to start units according to a preset rule so as to establish a plurality of independent power islands; after the frequency and voltage of each power island are stabilized, power is supplied to a next-stage transformer substation or power station in parallel, and partition collaborative recovery is carried out; dynamically distributing a load to be recovered, and performing synchronous parallel operation by using real-time phase angle difference data provided by a synchronous phasor measurement unit (PMU); and carrying out synchronous parallel operation on a plurality of subsystems which are stably operated, and gradually expanding a power grid architecture.

A phasor measurement unit (PMU) of the present disclosure measures phasor, i.e., magnitude and phase angle of voltage and current, and related data from a specific location on the electrical gird synchronized to a common time source. The time-synchronized phasor is called a synchrophasor. In a system of the present disclosure, a plurality of PMUs transmit the synchrophasors and related data to a phasor data concentrator (PDC), which aggregates and time-aligns the data for real time and post analysis. The PMU of the present disclosure further functions as a power quality meter determining at least one of symmetrical components' phasor, frequency, rate of change of frequency, high-speed digital inputs, analog fundamental power and / or displacement power factor.

The invention provides a power system synchronous phasor measurement unit configuration scheme generation method and related equipment. The method comprises the steps of determining a voltage distribution coefficient of each node and a current phase of each branch of a new energy grid-connected system; identifying a main influence area of the dominant oscillation mode of the new energy grid-connected system by using the voltage distribution coefficient of each node; identifying a propagation path of a dominant oscillation mode of the new energy grid-connected system by using the current phase of each branch; and generating a synchronous phasor measurement unit configuration scheme of the new energy grid-connected system by combining the main influence area and the propagation path based on the obtained measurement demand. Therefore, the key monitoring position related to the oscillation mechanism can be accurately determined, and the generated PMU configuration scheme can highlight the real-time monitoring capability on the weak oscillation node while ensuring the measurement coverage efficiency.

The invention relates to the technical field of power system automation and electric energy quality control, in particular to an electric energy quality optimization control method and system based on distributed collaborative decision. The method comprises the following steps: deploying synchronous phasor measurement units and corresponding edge intelligent decision nodes at a plurality of nodes of a power grid, and constructing a distributed network; each node calculates an electric energy quality characteristic index based on local data; dynamic role negotiation is carried out through local communication, and a leading controller is determined; the dominant controller fuses strategies and confidence information of local and adjacent nodes to generate a cooperative control strategy; and finally, equipment such as a static var generator and an active power filter is controlled to perform coordinated execution. Through a decentralized collaborative decision-making mechanism, the defects that centralized control is slow in response and low in reliability are overcome, rapid self-adaptive treatment on voltage sag, harmonic waves and other problems is achieved, and the response speed, the treatment success rate and the system robustness are remarkably improved.

The application discloses a power distribution feeder voltage transformer online monitoring method and system based on a phasor measurement, and the method comprises the following steps: collecting phasor data of each key node in real time through a phasor measurement unit arranged at a key node of a power distribution network; based on topological structure information and a switch state of the power distribution network, taking the key node which has an electrical correlation with a node to be monitored as a neighboring measurement node; based on the phasor data, identifying line parameters between the node to be monitored and each neighboring measurement node, and obtaining the line parameters; calculating a virtual reference voltage of the node to be monitored based on the line parameters; comparing a measured voltage of the node to be monitored with the virtual reference voltage, and obtaining a ratio error and a phase error; and based on error criteria corresponding to the ratio error and the phase error, monitoring an operation state of a voltage transformer.

The present invention discloses a broadband oscillation monitoring method based on waveform data, which relates to the field of power systems and includes: performing real-time sampling on the power grid system through a synchronized phasor measurement unit of a wide-area measurement system to obtain electrical signal waveform sequences for each time period; performing parameter identification on the electrical signal waveform sequences for each time period; comparing the amplitudes of harmonics and interharmonics in each time period, and if the amplitude of the harmonic or interharmonic continuously increases, it indicates that the power grid system has oscillated; if the amplitudes of both the harmonic and interharmonic do not continuously increase, it indicates that the power grid system has not oscillated. The present invention does not require the installation of additional equipment and can monitor oscillations in a broadband range only by using the data of the synchronized phasor measurement device, reducing the monitoring cost, solving the problem of online monitoring of oscillation signals in a broadband range, and can adaptively determine the window length of the required data according to the noise intensity to reduce the algorithm calculation amount, which has important significance in engineering.

The present application relates to a kind of active power distribution network multi-type voltage regulating equipment real-time coordination control method, comprising: step 1: by phasor measurement unit (PMU) or monitoring and data acquisition unit (SCADA) obtains data, based on the data obtained, the selection of window period is carried out, and according to line voltage sensitivity, voltage is estimated, according to the charge-discharge equation of battery energy storage, the battery state of charge (SOC) is predicted;Step 2: based on the bandwidth limit of on-load tap changer (OLTC) and the power deviation of distributed new energy (DG), voltage is smoothed and corrected, and the constraint condition based on distributed new energy and battery energy storage system (BESS) is constructed;Step 3: to avoid early depletion of battery energy storage system, according to the difference between battery state of charge and battery reference state of charge, the active power of battery energy storage system is adjusted.The method can solve the problem of mutual interference when mechanical voltage regulating equipment and energy storage equipment operate.

The invention discloses an electric power system robust dynamic state estimation method and system based on complex field measurement, complex measured values of a synchronous phasor measurement unit (PMU) are introduced into a dynamic state estimation process, abnormal data interference in PMU measurement is overcome through a complex field robust generalized maximum likelihood unscented Kalman filtering algorithm, and the dynamic state estimation process is optimized. And robust state estimation of the power system is realized. The method provided by the invention can solve the problem that the estimation precision is influenced by the abnormal value on the premise of keeping the PMU full complex number field measurement information, improves the accuracy and reliability of state estimation, and better describes the operation state of the power system.

The application provides a method for participating in primary frequency modulation control of a power system based on a modular large-scale hydrogen production power supply, comprising: obtaining original data streams from a high-precision synchronous phasor measurement unit, an electrolytic cell and a sub-module, performing multi-scale decomposition and noise reduction processing after time synchronization and abnormality detection repair; using unscented Kalman filtering for state estimation, combining empirical mode decomposition and Hilbert transform to extract frequency disturbance characteristics and estimate grid inertia; using an established electrolytic cell dynamic model to identify and optimize parameters through recursive least squares; designing a control strategy based on a multi-time scale control framework, optimizing virtual synchronous generator parameters through online Gaussian process regression, combining scenario tree stochastic model predictive control and minimax regret learning to realize frequency modulation control; finally, coordinating control through a distributed consistency algorithm and performing multi-objective performance evaluation optimization. The system's frequency modulation performance and operating efficiency are improved.

The application discloses an artificial intelligence-based power system transient stability evaluation method considering sample imbalance, and comprises the following steps: step 1, generating an original data set containing system operation information based on a time domain simulation technology; step 2, proposing a synthetic minority over-sampling technique (SMOTE) method with adaptive eligible synthesizer selection to process the original data set, and establishing a sample balanced and rich data set; step 3, constructing a feature extraction model based on a parallel convolution algorithm; step 4, training and establishing a transient stability evaluation model based on a rich random forest based on an offline data set; step 5, quickly establishing an updated model; and step 6, completing online transient stability evaluation based on data collected by a phasor measurement unit. SM The application has the advantages that an ASN-SMOTE method is proposed, the sample imbalance problem is effectively solved, a feature extraction framework is constructed based on a parallel convolution algorithm, highly relevant key operation variables T SM are extracted, and the calculation efficiency of subsequent model evaluation is greatly improved.

The invention provides a flexible regulation load remote control terminal and method for an integrated photovoltaic distribution network, and relates to the technical field of photovoltaic distribution networks, and the terminal comprises a photovoltaic measurement and control unit, a synchronous phasor measurement unit, a reconfigurable power electronic control unit, a load control unit, an edge calculation unit, a communication and safety module, and a local autonomous unit. According to the method, spectral decomposition is carried out on photovoltaic output and load power, high-frequency disturbance is rapidly adjusted by power electronics, and low-frequency disturbance is coordinately processed by an energy packet and energy storage; strategy prediction and verification execution are realized by adopting double-layer optimization and digital twinning; and in combination with federal learning and an intelligent contract mechanism, load flexible regulation and control, photovoltaic friendly consumption and distributed energy transaction are realized. According to the method, the photovoltaic consumption level can be effectively improved, the voltage fluctuation and the light abandoning rate of the power distribution network are reduced, the terminal autonomy capability and the privacy protection performance are enhanced, the power grid stability, the user comfort and the economical efficiency are considered, and the method has relatively high popularization and application values.

Disclosed is a resilient route generation system for reliable communication of a phasor measurement system of a power grid. A named data network is a new network architecture to improve the communication reliability between a phasor measurement unit and a phasor measurement concentrator in power transmission and distribution networks. The lost data packets in a current router can be directly recovered from an upstream router with resilient route, and the optimal RR selection will maximize the success rate of retransmission of lost data packets, thus maximizing the network reliability. The mesh network and ring network structure of the power grid are fully utilized, and a resilient route generation system is provided, wherein the resilient route of each communication pair includes a corresponding primary path and a plurality of redundant sub paths, so that the success rate of retransmission of lost data packets and the network reliability are maximized.

The application relates to the technical field of medical big data processing, and relates to a medical data classification and labeling method and system applied to big model mining, which comprises a heterogeneous data phasor measurement unit, which samples multi-modal medical data and calculates a semantic deviation angle and a semantic angular velocity; a transient energy evaluation engine, which is configured with a Lyapunov function calculation core and is used for real-time calculation of semantic kinetic energy, semantic potential energy and instantaneous total energy of the sample; a stability discrimination and shunt controller, which calculates an energy margin and shunts the data to different channels; and an excitation feedback adjustment module, which executes an adaptive excitation control strategy and dynamically adjusts big model parameters in the medical field. The application quantifies data semantic reliability by using transient energy stability theory, accurately identifies critical high-value samples by calculating an energy margin, effectively suppresses model hallucinations by adaptive parameter adjustment while greatly reducing manual labeling costs, and significantly improves the clinical safety and robustness of the medical big model.

The invention relates to the technical field of power systems, in particular to a power distribution network fault positioning method and system based on load impedance, and the method comprises the following steps: classifying faults of a distributed power supply power distribution network; fault positioning is carried out in the completely observable power distribution network according to current and voltage data recorded by the miniature phasor measurement unit; in the partial observable power distribution network, a fast reverse or forward load flow calculation method is adopted to obtain initial load impedance of each node, and based on the initial load impedance of each node, a particle swarm optimization algorithm is adopted to optimize complex power of each node to obtain a load value of each node; in the partial observable power distribution network, the fault distance of each section is calculated by adopting a least square method for the fault position. Aiming at different characteristics of a completely observable power grid and a partially observable power grid, technical means such as a fast reverse or forward load flow calculation method, a particle swarm algorithm and a least square method are combined, and the limitation that a traditional method is difficult to apply to the partially observable power grid is solved.

The application relates to a photovoltaic power station outgoing line voltage diagnosis method and system, and belongs to the technical field of supervisory control and data acquisition systems. The system realizes accurate distinction of transient disturbance and permanent fault and high-precision fault positioning by real-time acquisition of electrical quantity data through a synchronous phasor measurement unit, data preprocessing, multi-time scale feature extraction, transient and steady state identification and fault determination module analysis, and outputs diagnosis instructions based on a dynamic self-adaptive threshold and a closed-loop decision mechanism, thereby improving the accuracy, response speed and operation and maintenance efficiency of photovoltaic outgoing line voltage state diagnosis.

The invention relates to an electric power system inertial parameter real-time evaluation method and system based on recursive system identification, and the method comprises the following steps: carrying out measurement by using a phasor measurement unit installed on a power generation bus, and collecting active power and bus frequency measurement values in real time; the collected data are preprocessed; calculating the rotor speed change of the synchronous generator or the internal frequency change of the converter interface generator through an estimation algorithm of the rotor speed; constructing an adaptive infinite input response filter model through a recursive least square equation error algorithm, and obtaining a z-domain transfer function through the filter model; and converting the z-domain transfer function into an s-domain transfer function, and extracting a mechanical time constant, a damping coefficient and a droop gain of the generator, namely the inertial parameters of the power system to be evaluated. Based on an advanced recursive algorithm and real-time measurement data, the inertia of the generator in the power system is estimated in real time, the method can adapt to dynamic changes of the system, and a more accurate evaluation result is provided.

The present invention relates to the technical field of safe and stable control of distribution networks, and in particular to a method for evaluating voltage-reactive coupling characteristics at a regional distribution network collection point. The method comprises the following steps: deploying synchronous phasor measurement units at multiple collection points in different areas, collecting electrical parameters and generating original state data of the collection point; performing quality verification and cleaning on the original state data to generate pre-processed state data of the collection point; performing feature extraction and time series reconstruction on the pre-processed state data of the collection point to obtain a set of key characteristic parameters of the collection point; based on the set of key characteristic parameters, constructing an initial model of the inverter group through clustering of inverter operation characteristics and analysis of control link characteristics; performing electromagnetic transient-steady-state hybrid simulation on the inverter group based on the model to generate a transient-steady-state hybrid simulation scheme. The present invention can accurately capture the dynamic interaction between the inverter group and the power grid, and effectively identify the resonant path and the reactive coupling path.

The invention belongs to a power system measurement technology, and discloses a synchronous phasor standard device and a frequency precision measurement method adopting the synchronous phasor standard device. According to the device, a 10MHz clock, a 1PPS pulse per second and an IRIG-B code are provided by using a Beidou time service module, and a high-precision synchronous sampling platform is formed by the Beidou time service module, a synchronous phasor signal source, a proportional conversion standard device and a dual-channel sampling system. The method comprises the following steps of: aiming at the frequency measurement error of a synchronous phasor measurement unit, firstly removing direct current offset from a sampling signal, segmenting the sampling signal, obtaining initial frequency estimation by utilizing fast Fourier transform, constructing a residual sum of squares optimization model in a set amplitude, frequency and phase boundary, obtaining optimal frequency estimation through iterative solution, and further calculating the frequency change rate.

The invention discloses a phasor measurement unit configuration strategy generation method and related equipment, and the method comprises the steps: determining a target power grid, and constructing a three-phase power grid model based on the target power grid; determining each power grid node contained in the three-phase power grid model, and screening each first power grid node which is not configured with the phasor measurement unit from each power grid node; constructing a configuration strategy generation model by taking minimization of the state estimation error and the configuration cost of the phasor measurement unit as targets; and performing optimization solution on the configuration strategy generation model by adopting NSGA-II in combination with the congestion distance and the elite strategy to generate an optimal configuration strategy, the optimal configuration strategy being used for representing whether each first power grid node needs to be configured with a phasor measurement unit or not. It can be seen that by constructing the three-phase power grid model and screening the first power grid node, the difficulty of constructing and solving the configuration strategy generation model is simplified, the strategy generation process is accelerated, and the optimization efficiency and the strategy rationality are improved.

The invention discloses a multi-modal cooperation-based stability control method for a network-building type power system. A power grid system is connected with a plurality of energy units for inputting electric energy to the power grid system; acquiring key state data of the power grid through the synchronous phasor measurement unit and the sensor; based on the key state data, a virtual inertia value is adjusted in real time through a dynamic virtual inertia nonlinear adaptive algorithm, and an improved Nash bargaining game model is adopted to carry out multi-energy power distribution optimization; when a power grid fault is detected, constructing a time-varying delay virtual impedance model, and dynamically adjusting impedance delay time; through the above models and parameters, the operation parameters of the energy unit and the operation parameters of the interface of the energy unit and the power grid system are adjusted, and the stability control of the power grid system is realized. The control method disclosed by the invention has the advantages of improving the inertial support capability and the frequency stability of the system, enhancing the fault ride-through capability, ensuring the frequency stability, the power balance and the voltage stability of the power grid system and the like.

Classical optimization techniques, when applied to the placement of measurement devices in very large power networks, become computationally intractable and are not guaranteed to converge. In contrast, disclosed embodiments guarantee an optimal placement strategy of measurement devices, such as phasor measurement units for measuring voltage and current, across a power network of any size, such that the voltage at each bus is observable, in dramatically less computational time than classical optimization techniques. In particular, embodiments traverse at least one spanning tree of a graph of the power network, and place measurement devices based on distance values of the nodes from the nearest measurement device. The distance values, and potentially the observability status, of all nodes are updated after each placement.

The invention discloses a false data injection method for a broadband synchronous phasor of a power system, and relates to the technical field of wide area measurement and control of the power system, and the method comprises the steps: collecting broadband synchronous phasor measurement unit (WPMU) data of a target area; when it is detected that broadband oscillation occurs in the power grid and the oscillation duration exceeds a preset threshold value, the attack starting opportunity is judged; randomly selecting the measurement value of one oscillation mode as a false data injection target, and obtaining the measurement value of the oscillation mode; constructing and solving a multi-target nonlinear programming model, generating tampered WPMU data, and injecting the tampered WPMU data into a scheduling center through man-in-the-middle attack MITM or protocol vulnerabilities; and monitoring whether the power system cuts off the target wind power plant or not, and if the target wind power plant is not successfully cut off, performing attack again. According to the method, the data tampering quantity can be reduced, the attack concealment is improved, the risk found by a detected system is reduced, and a powerful guarantee is provided for safe operation of a power system.

The invention discloses a multi-source power monitoring data alignment method based on adaptive multi-scale space-time coupling, and belongs to the field of intelligent sensing and data fusion. The method comprises the following steps: acquiring multi-source signals of a phasor measurement unit, a monitoring system, a sensor, an intelligent electric meter and the like; a unified input sequence is constructed through communication delay compensation, linear interpolation missing compensation, cubic spline resampling and standardization; multi-scale features are extracted by using expansion convolution, and space-time alignment is realized in combination with a learnable space-time coupling matrix and a multi-head attention mechanism; the reliability score of each channel is calculated on line through a specific coding branch network, a fusion weight is dynamically generated, weighted fusion is completed, and a collaborative monitoring sequence is output; and finally, combining the fusion error, the alignment error, the causal consistency and the scale regularization to form comprehensive loss and end-to-end optimization parameters. In the reasoning stage, a high-precision fusion result can be obtained on new data only through forward calculation, and the accuracy of power grid state monitoring and fault diagnosis is remarkably improved.