120results about How to "Good engineering application value" patented technology

The invention relates to the field of gas leakage detection, and discloses a gridding directional laser methane detection system, which comprises the following steps of: acquiring area information which comprises a methane absorption signal, environmental data, background gas, an area image and historical leakage data; generating first two-dimensional structure data from the collected regional information, wherein the first two-dimensional structure data comprises a data matrix and a relation matrix; first one-dimensional structure data is obtained based on sorting of the first two-dimensional structure data, wherein the first one-dimensional structure data comprises n data items sorted according to time; and inputting the first two-dimensional structure data into the detection model, and outputting a methane leakage processing strategy. According to the invention, detection and treatment strategies of methane in fuel gas can be realized, and a distribution strategy of the methane laser detection device can be further given.

The invention relates to the technical field of drilling parameter optimization control, and provides a petroleum drilling parameter optimization method and system and a storage medium. Comprising the steps that torque fluctuation amplitude and bit pressure response delay data are collected in real time, an original data sequence is constructed, and parameter trend data are obtained; analyzing the parameter trend data to identify a stratum change mode, extracting trend feature information, and when the trend feature information meets a preset condition, determining a corresponding abnormal mode identifier; generating a parameter adjustment direction of drilling control according to the abnormal mode identifier, and optimizing the parameter adjustment direction in combination with parameter weight analysis and a real-time feedback mechanism to obtain final optimization guidance data for guiding drilling operation; and mapping the final optimization guidance data to a control interface, and generating a dynamic adjustment instruction for drilling operation. The problems that in the existing petroleum drilling process, parameter adjustment lags behind, and stratum changes are difficult to adapt are solved, and intelligent optimization and dynamic control over drilling parameters are achieved.

The invention discloses a wind turbine blade trailing edge beam and core material lap joint sandwich initial failure position prediction method, and relates to the technical field of wind turbine blade structure damage failure prediction methods. Damage of a sandwich structure panel and a core material in a shear failure mode is characterized as a lap joint sandwich structure formed by combining a laminated plate area, a conventional sandwich area and an inclined plane lap joint sandwich structure area; respectively calculating the cross section bending rigidity of the conventional sandwich structure, the neutral layer position of the inclined plane lap joint sandwich structure and the cross section bending rigidity distribution of the inclined plane lap joint sandwich structure, and calculating the cross section bending rigidity distribution of the inclined plane lap joint sandwich structure on the basis of the neutral layer distribution function inflection point in the length direction of the inclined plane lap joint sandwich structure and the descending slope of the cross section bending rigidity distribution function; the initial failure position of the lap joint sandwich structure of the blade trailing edge beam and the trailing edge core material is predicted. The method provides an important theoretical basis for in-service wind power blade design, health prediction and repair decision making, and has great method innovation and engineering application values.

The invention discloses a low-voltage transformer area topology modeling method and system based on hierarchical clustering, and the method comprises the steps: building a sensitivity relation between voltage and current, calculating a sensitivity coefficient of voltage to current, and calculating a sensitivity matrix according to the resistance value and reactance value of each line in transformer area topology; subtraction is carried out on the electrical quantity under the continuous time section, a sensitivity relational expression is converted, and an associated impedance matrix is obtained and comprises an associated resistance matrix and an associated reactance matrix; based on an improved hierarchical clustering algorithm, inputting an associated resistance matrix or an associated reactance matrix to generate a tree diagram; and identifying hidden nodes according to the tree diagram, determining a generation sequence of the hidden nodes and a subordinate connection relationship between the nodes, and constructing a transformer area topology. By implementing the method, not only can the correlated impedance between the end nodes be accurately calculated, but also all hidden nodes in the network can be identified based on the correlated impedance matrix and the network topology can be correctly deduced, high accuracy and robustness are shown, and the cost is low.

The invention discloses a high-power-supply-rejection-ratio band-gap reference circuit with base current correction and high-order temperature compensation, which comprises a pre-voltage-stabilizing circuit module used for generating and outputting internal power supply voltage according to external power supply voltage, the internal power supply voltage is supplied to the band-gap reference core circuit module so as to isolate transmission of power supply noise to reference output; the high-order temperature compensation circuit module is used for generating nonlinear compensation current according to the bias voltage; the band-gap reference core circuit module is used for generating current in direct proportion to absolute temperature, reference voltage and reference output voltage based on internal power supply voltage and feeding back the reference voltage to the pre-voltage-stabilizing circuit module to serve as bias reference; and the base current correction circuit module is used for sampling the base current of the bipolar transistor in the band-gap reference core circuit module. On the premise that the circuit structure and the power consumption are reasonable, the temperature drift of the reference voltage is reduced, and the power supply suppression capability is improved.

The invention discloses a fabricated steel-ultra-high-performance concrete combined bridge deck structure with shear grooves and a construction method of the fabricated steel-ultra-high-performance concrete combined bridge deck structure, and relates to the technical field of bridge deck constructions.The bridge deck structure comprises a lower-layer steel bridge deck, an upper-layer prefabricated ultra-high-performance concrete bridge deck and shear connectors; the prefabricated ultra-high performance concrete bridge deck is provided with a shear groove, an embedded connecting piece penetrates through a groove opening of the shear groove in the transverse bridge direction, the two ends of the embedded connecting piece are embedded in the prefabricated ultra-high performance concrete bridge deck, the top of the shear connecting piece is fixed to the embedded connecting piece in a lap joint mode, and the bottom of the shear connecting piece is fixed to the steel bridge deck. A post-pouring layer is arranged between the prefabricated ultra-high performance concrete bridge deck and the steel bridge deck. The bridge deck slab is good in structural integrity, excellent in crack resistance and reasonable in stress, the interface connection problem of the prefabricated ultra-high-performance concrete bridge deck slab and the steel bridge deck slab can be effectively solved, the bearing capacity of a combined bridge deck system is improved, and the service life of the combined bridge deck system is prolonged.

The invention provides a curvature-enhanced large-displacement image variational optical flow method for motion estimation, relates to the field of image processing, and aims to describe the local structure complexity of an image by introducing an image contour curvature. On the basis of the curvature, limited self-adaptive weighted adjustment is carried out on the brightness invariant constraint and the gradient invariant constraint on the data item level of the opto-rheological model, so that the interference of unreliable matching in a complex structure region on optical flow estimation is inhibited; the robustness of optical flow estimation in illumination variation, complex texture and large displacement scenes is improved; and the numerical stability and convergence of the model in the multi-scale calculation process are ensured.

The invention discloses a switched reluctance motor turn-to-turn short circuit fault diagnosis method. According to the method, a second harmonic characteristic value is extracted according to an acquired stator torque signal of the switched reluctance motor, and a fault diagnosis model based on OCSSA-VMD-CNN-BiLSTM is established. The method comprises the following implementation steps: 1) acquiring a signal; 2) feature value extraction; 3) preprocessing characteristic values; 4) extracting time-frequency domain features of the feature values, and labeling; (5) a VMD-CNN-BiLSTM model optimized by the OCSSA is constructed; (6) an OCSSA-VMD-CNN-BiLSTM model is trained and tested; and 7) inputting the time-frequency characteristics of a to-be-detected sample into the model to obtain a fault diagnosis result. The off-line and on-line identification methods are combined, the turn-to-turn short circuit fault of the switched reluctance motor and the degree of the turn-to-turn short circuit fault of the switched reluctance motor can be accurately diagnosed, and a feasible fault diagnosis method is provided for the turn-to-turn short circuit fault of the switched reluctance motor.

The application provides a wind tunnel pressure regulating valve opening set fault diagnosis method based on a composite criterion, relates to the technical field of equipment fault diagnosis, and comprises the following steps: training a fault diagnosis model by using seen fault sample, extracting sample features by using an encoder, and obtaining classification information and reconstruction information of the sample by using a classifier and a decoder; then, according to the output result of the model on training data, a composite criterion is constructed in a feature space, a probability space and a sample space, and is used for identifying unknown faults; finally, when a new fault sample is input, the composite criterion is used to judge the sample, accurate diagnosis of known faults and effective identification of unknown faults are realized. The composite criterion method has a clear structure, a simple implementation process, can be extended on the basis of an existing fault diagnosis model, and has good engineering application value.

The application claims a small sample bearing fault diagnosis method based on Mamba and double granularity cross-domain alignment, which is used to solve the problem of low diagnosis accuracy caused by large distribution difference of bearing monitoring data across working conditions and scarcity of target working condition samples. The method constructs a double branch space-channel feature extraction network CA-Mamba, adopts an improved Mamba module and channel attention to enhance key response, and fuses deep features through a selective bridging module. Then, a domain discriminator and a group discriminator are constructed to realize cross-domain feature confusion and class boundary constraint; relying on the confidence weighted pseudo label PSMMD and dynamic memory enhanced contrast learning DMECL, the target domain feature is constrained from the class-instance double granularity, so that the target domain sample is aligned to the homologous source domain prototype, and finally the covariance measurement is used to complete the target domain classification. The application can extract domain-invariant and high-discriminative features under small sample conditions, and improve the diagnosis accuracy under cross-domain conditions.

The invention provides a method for generating performance degradation data of a train air braking system. The method comprises the following steps: on the basis of a brake system of a preset model, analyzing three stages of pressure rise, maximum pressure and pressure stability of pressure change of a brake cylinder, and extracting a plurality of typical fault modes; based on an extended random process model with non-monotonicity and independent incremental property, three stages of pressure change of a brake cylinder and multiple typical fault modes are combined, and linear, nonlinear or stepped performance degradation process modeling of the train air brake system is carried out; overlapping normal distribution random noise on a performance degradation track generated by modeling, simulating a control precision error of an actual braking system, and optimizing a performance degradation process model of the train air braking system; and utilizing the train air braking system performance degradation process model to generate train air braking system performance degradation data in batches.

The application provides a curvature-enhanced large displacement image variational optical flow method for motion estimation, relates to the field of image processing, and aims to depict the complexity of local structures of an image by introducing the curvature of the image contour line, and to restrict the adaptive weighting adjustment of the brightness invariable constraint and the gradient invariable constraint on the basis of the curvature in the data item level of the variational model of the optical flow, so as to inhibit the interference of unreliable matching pairs on the optical flow estimation in the complex structure area, improve the robustness of the optical flow estimation under the scenes of illumination change, complex texture and large displacement, and guarantee the numerical stability and convergence of the model in the multi-scale calculation process.

The application relates to the technical field of new energy vehicles and vehicle detection equipment, and discloses a portable device for heavy truck axle housing liquid level measurement, which is suitable for electric drive axle and traditional drive axle. The device comprises a shell support structure, an electronic control assembly, a posture adjusting execution mechanism, a quick locking connection mechanism and a liquid level measurement assembly; the shell support structure is used for supporting the device and stretching and retracting the measurement assembly; the electronic control assembly is used for power supply, control and result display; the posture adjusting execution mechanism is used for adjusting the posture of the measurement assembly to keep it vertical; the quick locking connection mechanism is used for quickly fixing the device and the oil filling hole of the axle housing; and the liquid level measurement assembly realizes liquid level detection by cooperating a magnetic float and a Hall sensor. The application takes into account the application requirements of the electric drive axle and the traditional drive axle, and can especially meet the requirements of the electric drive axle on liquid level detection precision and reliability.

This invention provides an active, compliant end-effector and its control method for laying insulating boards. The end-effector includes a robotic arm connection base, a six-dimensional force sensor, a two-dimensional orthogonal movement platform, and a vacuum sponge suction cup adsorption mechanism. The six-dimensional force sensor is installed below the end flange of the robotic arm and is used to measure the force and torque of the gripper end in three-dimensional space in real time and with precision. The two-dimensional orthogonal movement platform includes X-axis and Y-axis movement units. This end-effector, through mechanical decoupling design and force sensing information fusion control, endows the robotic arm with a high degree of environmental adaptability. By automatically compensating for assembly path errors and adapting to changes in the shape and position of the workpiece and surface in real time, it completes the laying operation with compliant and controllable force. This fundamentally solves the problems of inaccurate alignment, uneven pressing, and damage caused by excessive local stress in traditional laying methods, significantly improving the automation level of insulating board laying and possessing good engineering application value.

The present application relates to a kind of spot I type nugget fracture failure constitutive model construction method, belong to the fracture mechanics and numerical simulation method technical field of spot welding joint.The technical scheme of the present application is: the parameter construction of constitutive model is carried out, parameter is directly used for spot welding joint related finite element simulation;Spot welding joint is bonded with uniaxial tensile fixture, and the tensile experiment of nugget fracture form is completed under tensile testing machine and the load-displacement curve and maximum load are recorded;Finite element model is constructed and tensile displacement is calculated;Convert and correct stress-displacement curve in combination with measured nugget area, solve maximum tensile stress and fracture energy, parameter is substituted into model and the precision of replacement connection mode is verified.The beneficial effects of the present application are: the accurate construction of spot I type nugget fracture failure constitutive model is realized, which provides standardized mechanical parameter support for spot welding joint structure reliability design and welding process optimization, and significantly improves the accuracy and engineering application value of finite element simulation.

The invention relates to the technical field of chemical synthesis, and discloses a PMMA impact modifier and a preparation method and application thereof.The PMMA impact modifier comprises a core layer, a transition layer and a functional shell layer, the core layer is sequentially coated with the transition layer and the functional shell layer, the core layer is a soft monomer crosslinking layer, the transition layer is a soft monomer and hard monomer crosslinking layer, the functional shell layer is a hard monomer crosslinking layer, the functional shell layer contains a photochromic component. Through the multi-layer structural design of the core layer, the transition layer and the functional shell layer, the shock resistance, the optical performance and the functional performance are synergistically improved; the core layer endows the material with good toughness and energy absorption capability, the transition layer effectively buffers performance difference between layers and keeps transparency, and the functional shell layer introduces a photochromic component while ensuring structural rigidity and durability, so that the material has a peep-proof function under the change of illumination conditions; the method is suitable for application scenes with high requirements on safety and functionality, such as PMMA products, automobile glass and the like.

The utility model discloses a water -cooled coal bed gas hydraulic fracturing system, this system passes through integrated high -efficient water -cooling unit, effectively solved traditional coal bed gas hydraulic fracturing system in air -cooled heat dissipation efficiency low, energy consumption is high, loud and frequent maintenance technical problem. System adopts plate heat exchanger and circulating cooling water system, has realized the accurate control of hydraulic oil temperature, has improved system energy efficiency and operating stability significantly, especially suitable for coal mine underground limited space and high dust environment. The system has heat dissipation efficiency high, compact structure, energy consumption is low and maintenance cycle long etc. advantage, provides strong support for the efficient exploitation of coal bed gas.

This invention discloses a method, system, and storage medium for multivariate time-series feature extraction and grade prediction in flotation processes. The method includes the following steps: Step S1: Raw data input and encoding; encoding and structured input of time-series data composed of various process variables; Step S2: Extracting dynamic features; Step S3: Condition-guided encoding modeling; Step S4: Enhancing the saliency of key variables and important time segments; employing a multi-head attention mechanism to match key-value pairs generated from target-guided query vectors and multi-scale features; Step S5: Outputting a prediction module; performing feature fusion and nonlinear mapping on the attention mechanism output to output the predicted concentrate grade and recovery rate for future times. The system and storage medium are both based on the above method. This invention has advantages such as higher intelligence, better controllability, and improved prediction accuracy and model adaptability for key indicators in the flotation process.

The invention relates to a building block type ice melting device active current sharing system and a control method, and the system comprises a plurality of ice melting rectification devices which are connected in parallel to a to-be-ice-melted line, and a bidirectional Buck-Boost current sharing module is connected in series between the DC output end of each ice melting rectification device and the to-be-ice-melted line; and the current sharing controller is used for collecting the output voltage value and the output current value of each ice-melting rectifying device in real time, judging whether the plurality of ice-melting rectifying devices share the current or not, calculating the current sharing control quantity and generating a compensation voltage instruction when the plurality of ice-melting rectifying devices do not share the current, and adjusting the corresponding bidirectional Buck-Boost current sharing module through the compensation voltage instruction, so that the plurality of ice-melting rectifying devices realize the current sharing. High-precision current sharing and effective circulating current suppression can be realized under the condition of not changing the structure of the existing main circuit, and the circuit is suitable for a large-capacity and multi-machine parallel mobile ice melting equipment scene and has a remarkable engineering application value.

The application belongs to the technical field of wireless communication, and is especially a pinch antenna system (PASS) assisted downlink communication resource allocation method and system. The method comprises the following steps: S101: constructing a system model; S102: establishing a double-time scale optimization problem with the goal of maximizing the long-term average system sum rate; S103: constructing a resource optimization framework; S104: performing transmission resource optimization according to the number of medium waveguides in the system; S105: constructing a sample average approximation target of the long-term average system sum rate based on multiple small-scale fading samples; and S106: updating the PA deployment position based on the long-term average performance index until the convergence condition is met or the maximum number of iterations is reached. The application can avoid frequent adjustment of the pinch antenna position, thereby effectively reducing the system implementation complexity and calculation overhead, significantly improving the long-term average sum rate and multi-user communication performance of the system, improving the transmission quality of users, and having good engineering application value.

The invention provides a cast-in-place box girder pre-camber automatic regulation and control method and device. The device comprises a first-stage jacking system, a second-stage jacking system, a third-stage jacking system and a transverse moving system. After the pouring maintenance of the current span box girder is completed, the hydraulic oil cylinders of the first-stage jacking systems of the front and rear main supporting legs retract, the whole machine falls down, and the formwork is demoulded; the hanging outer rib is disconnected, a three-stage jacking system oil cylinder at the top of the cross beam retracts, and a transverse moving system is started to achieve mold opening; the whole machine moves forwards by one span, and the jacking amount is adjusted according to the longitudinal slope of a main beam through a front and rear main supporting leg first-stage jacking system; correcting the section position of the template through a three-stage jacking system at the top of the cross beam, and controlling the bottom die to be positioned at the same elevation; a transverse moving system is started to complete outer mold bottom plate connection, and the actual position line of the mold plate is determined; calculating to obtain the elevation adjustment amount of the outer mold of each block based on an influence matrix method, and performing synchronous and automatic regulation and control by utilizing a secondary jacking system, so that a template system reaches a designed pre-camber line position; the problems that the pre-camber regulation and control efficiency of a traditional cast-in-place box girder movable formwork is low, errors are uncontrollable and the like are solved.

The application discloses in-situ growth coated catalytic proppant and a preparation method and application thereof, and the preparation method of the proppant comprises the following steps: dispersing a proppant matrix in a salt solution containing one or more transition metal precursors to form a solid-liquid suspension system; adding a precipitator to the suspension system to induce in-situ heterogeneous nucleation and growth of the metal precursors on the surface of the proppant matrix, thereby forming a precursor-coated proppant with a core-shell structure; and performing high-temperature calcination on the precursor-coated proppant to obtain the in-situ growth coated catalytic proppant. The application utilizes in-situ growth and high-temperature reactivity sintering technology to construct a firm high-entropy ceramic catalytic film on the surface of the proppant, thereby solving the problems that a traditional impregnated proppant coating is easy to fall off and has poor temperature resistance; meanwhile, the high-entropy effect and lattice oxygen vacancies are utilized to achieve a technical breakthrough that the proppant can efficiently catalyze thick oil viscosity reduction without an organic modifier.

This utility model discloses an ultrasonically consolidated graphite-reinforced copper-aluminum gradient composite functional structure, comprising a layered matrix formed by alternating consolidation of aluminum and copper layers, and a graphite rod embedded in the layered matrix. The layered matrix has a three-dimensional flow channel structure internally, and heat dissipation fins are disposed on its surface. The layered matrix is ​​fabricated using an ultrasonic consolidation device. By alternately consolidating copper and aluminum foil to form the layered matrix, and integrating internal cooling channels and external heat dissipation fins, it achieves coordinated heat dissipation through liquid and air cooling. This structure utilizes ultrasonic low-temperature solid-state bonding technology to avoid the formation of intermetallic compounds, and possesses high thermal conductivity, low density, and multifunctional heat dissipation characteristics. It is suitable for efficient thermal management in fields such as electronic equipment and new energy devices, and its manufacturing process is simple with significant engineering application value.

The invention relates to a wavelet neural network dynamic construction method, which comprises the following steps of: setting a nonlinear system, and processing an unknown nonlinear function in the nonlinear system by adopting a self-adaptive scaling wavelet network approximation framework; processing the non-linear function, namely determining an initial wavelet space for approaching the non-linear function by utilizing an initial wavelet frequency estimator of the adaptive scaling wavelet network approaching framework; on the basis of a scaling parameter self-adaptive adjustment mechanism, scaling parameters are dynamically updated on line so as to adjust the frequency bandwidth of the wavelet basis function and enable the frequency bandwidth to be matched with the spectrum distribution of the nonlinear function; and based on a wavelet basis function increasing and cutting mechanism of the adaptive scaling wavelet network approximation framework, increasing and cutting a wavelet basis function on the basis of the initial wavelet space so as to further approach the nonlinear function.

This invention belongs to the field of power system automation and energy storage technology. It provides a hybrid energy storage configuration and scheduling method based on digital twins. This invention constructs a digital twin and establishes a two-layer collaborative optimization framework. The upper layer optimizes the energy storage capacity configuration, and the lower layer trains the near-end strategy optimization scheduling PPO intelligent agent. At the same time, it introduces an energy storage aging cost model. Finally, it deploys online and uses digital twin advanced simulation to correct scheduling instructions, thereby reducing the annual comprehensive cost, extending the energy storage life, and realizing safe and efficient hybrid energy storage collaborative operation.

The application discloses a mud generator torque control method and system, and belongs to the technical field of well drilling engineering. The method comprises the following steps: performing phase identification on three-phase alternating current output by a mud generator through a digital processor, generating a PWM driving signal based on phase information, controlling the conduction and shutdown of a power switch tube connected with a generator winding, adjusting the current of a generator stator winding, and completing the control of the output torque of the generator. The application performs phase identification on three-phase alternating current through a digital processor and generates a PWM driving signal, realizes accurate control of the output torque of the generator, directly adjusts the current of the stator winding, and avoids the heat dissipation and space problems caused by the traditional resistance dissipation mode.

This invention provides an intelligent examination scheduling method based on time consumption prediction and dynamic optimization, belonging to the technical field of medical service management. The method includes: removing extreme data using a quantile filtering strategy and preprocessing equipment and examination time information; obtaining reasonable time consumption through random forest regression model training combined with departmental percentile truncation; integrating examination time with equipment resource status, calculating queuing time based on total time / number of equipment, and accumulating time for patients with multiple examinations; traversing operating time at 1-minute granularity, processing patient check-in and departmental queues in real time; generating a full permutation of examination items and filtering by walking time, selecting the optimal order with the earliest completion time; and finally outputting standardized scheduling results. This invention significantly reduces patient waiting time, improves hospital equipment utilization, has a significant optimization effect on high-frequency departments, and the scheduling results closely match actual hospital operations, enabling data-driven management decisions.

This application provides a method and related equipment for detecting small target vehicles from a low-altitude perspective, belonging to the fields of computer vision and intelligent transportation technology. The method includes: acquiring low-altitude images and performing multi-level feature extraction, performing feature preservation operations during the extraction process to retain high-resolution detail information; adaptively enhancing the extracted features to improve target-related feature responses; constructing a high-resolution feature response path, fusing shallow high-resolution features with deep semantic features to form enhanced features; and performing target detection based on the enhanced features. This application also optimizes feature representation through feature compression and path pruning to reduce computational complexity. Through a collaborative design of feature information preservation and reconstruction during multi-scale propagation, this application achieves a simultaneous improvement in the performance and computational efficiency of small target vehicle detection in complex low-altitude scenarios.

The utility model belongs to tobacco technical field, specifically disclose a kind of storage cabinet anti-mixing control device, the device includes shell, and the lower end surface of shell is equipped with the long hole being communicated with inner chamber, and the outer wall of front and back sides is equipped with fixed seat respectively, and shell front side wall right side is equipped with motor. In shell, the inner wall of front and back sides is rotatably connected with two rotating rods by bearing, and one rotating rod front end is fixedly connected with motor output end by coupling, and rotating rod outer wall is connected with gear by interference fit. Two baffle synchronous rotation are driven by motor driving rotating rod and gear, device is installed in the discharge end in stem and filament storage cabinet, can accurately accept the stem and filament falling at cabinet head, prevent its mixing into discharge conveyor belt, fundamentally avoid mixing phenomenon, ensure subsequent cigarette processing stem and filament quality pure, accurate specification, effectively guarantee the stable quality of cigarette product.

The invention discloses a reservoir area water surface photovoltaic physical-LSTM collaborative early warning algorithm model and method, and aims to solve the problems of single data dependence, disjunction of physical and data driving, threshold stiffness and the like in the existing early warning technology. The method comprises the following steps: firstly, collecting meteorological, hydrological and equipment multi-source data, and carrying out fusion preprocessing to generate a standardized tensor; a floating body dynamic physical model is constructed, a bidirectional LSTM network is optimized, and a cooperation mechanism is established through parameter bidirectional transmission and dynamic weight distribution; and finally, based on an entropy weight method and dynamic threshold calibration, triggering four-level early warning. The model F1 score reaches 85.9%, the response time delay is 8.7 minutes, the false alarm rate is reduced by 41%, the cross-reservoir generalization ability is remarkably improved, and the operation and maintenance cost is reduced by 30%. The wind wave load and electrical fault coupling risk high-precision early warning is achieved, and the method is suitable for active defense of a photovoltaic system in a typhoon high-incidence area and a complex water area with a large water level amplitude.