139results about How to "Lighten the computational burden" patented technology

The invention provides a physical information long-short-term memory network site seismic response prediction method based on kinematics derivative constraint and data folding. The method comprises the following steps: collecting and preprocessing bedrock and ground surface seismic oscillation data; displacement, speed and acceleration supervision labels are constructed through surface acceleration integration; constructing a long-short-term memory network containing a data folding and unfolding module; physical constraint loss is constructed by utilizing kinematics derivative relations between displacement and speed and between speed and acceleration, and a network is trained by combining an adaptive weight strategy; and outputting earth surface displacement, speed and acceleration time history by adopting the trained network. According to the method, under the condition that constitutive parameters of a soil layer are not explicitly given, physical consistency constraint can be applied to the field seismic response prediction process, the long-time program train training burden is reduced through sequence folding processing, and the method is suitable for field seismic response prediction.

The invention relates to a point cloud denoising method of a photon counting laser radar and an intelligent denoising system, and belongs to the technical field of photon laser radar data processing. The technical problems that the performance of a fixed parameter denoising method is seriously unbalanced under different ground feature and complexity scenes, and high denoising rate and high signal fidelity cannot be considered at the same time are solved. Synchronously acquiring an optical image and an original photon point cloud of the target area; identifying the dominant ground feature type of the laser spot area through real-time semantic segmentation; calculating scene complexity based on the point cloud data and determining the grade of the scene complexity; querying a preset two-dimensional parameter mapping table according to the combination of the surface feature type and the scene complexity level, and dynamically obtaining an optimal denoising parameter set; and finally, the parameter set is utilized to drive adaptive denoising processing, and signal points and noise points are distinguished by calculating the local density and direction consistency of the points and constructing a two-dimensional discrimination space. The method is mainly used for real-time processing platforms such as unmanned aerial vehicles, and the point cloud denoising precision, the signal retention rate and the topographic feature fidelity are remarkably improved.

The invention discloses an unmanned aerial vehicle adaptive path planning method based on risk triggering and local optimization. The method comprises the following steps: (1) establishing an unmanned aerial vehicle motion and obstacle environment model; (2) constructing an enhanced artificial test field and calculating resultant force; (3) calculating a risk index and generating a trigger signal; (4) executing PSO local re-planning during triggering; and (5) updating the position of the unmanned aerial vehicle and outputting a planned path. The invention provides a PSO-APF hybrid single unmanned aerial vehicle path planning method based on event triggering, and the method achieves the real-time, safe, smooth and efficient path planning in an obstacle environment through the construction of an enhanced artificial potential field model, a risk index and event triggering mechanism, a self-adaptive local re-planning window and a PSO optimization module. And a closed-loop process of potential field guidance, risk assessment, event triggering, local optimization and path updating is formed.

The invention discloses an underwater target detection method based on differential attention. The method comprises the following steps: acquiring an underwater image to be detected; based on the underwater image data set, constructing a lightweight underwater target detection network model based on a differential attention mechanism; constructing a joint loss function, and performing end-to-end training on the lightweight underwater target detection network model to obtain a trained lightweight underwater target detection network model; performing mathematical equivalence fusion on the lightweight underwater target detection network model to generate a lightweight reasoning model; and inputting the obtained underwater image to be detected into the lightweight reasoning model to realize target detection of the underwater image. According to the method, a differential attention feature interaction mechanism is introduced, common-mode noise is stripped through differential calculation, and the detection precision and real-time performance of an underwater fuzzy scene and a tiny target are remarkably improved in combination with a feature pyramid network integrating dynamic scale perception and semantic gating.

The invention relates to the technical field of power system automation, and discloses a cooperative control method, system and device for inverters in a power distribution area and a medium, and the method comprises the steps: carrying out the identity authentication of the inverters in the power distribution area through constructing a distributed authentication platform, and constructing a logic neighbor network of the inverters; determining the operation capability of the inverter according to the physical constraint and power operation data of the inverter, and constructing a multi-time scale coordination framework according to the real-time demand of the power distribution area to determine a bidding strategy set of the inverter; constructing a transformer area clearing model considering the safety of the power distribution transformer area, and solving the transformer area clearing model according to the logical neighbor network and a distributed optimization algorithm to obtain a power control instruction of an inverter; the distributed authentication platform monitors the operation state of the power grid in the power distribution area, and when the risk is triggered, a security check algorithm is started to correct the power control instruction, and a power correction instruction is obtained to control the inverter to execute; and under the condition of considering the safety and the cost, the refined cooperation of the district level is realized.

This invention discloses a distributed adaptive fault-tolerant control method and system for nonlinear inline systems, relating to the field of distributed control technology for nonlinear inline systems. The method includes: for each subsystem in the nonlinear inline system, constructing the actual output of each actuator in the subsystem based on the fault type of the actuator; the nonlinear inline system includes several subsystems; for each subsystem, constructing a local objective function, an error optimizer, an auxiliary system, high-gain parameters, and a control law, and generating an optimal error trajectory based on the local objective function and the error optimizer. At the upper level, this invention designs an error optimizer to generate the optimal error trajectory based on a local objective function that converges to zero; at the lower level, it designs a trajectory tracking fault-tolerant controller with the help of an auxiliary system, employing a dynamic high-gain approach to adaptively handle the effects of uncertain parameters, actuator faults, and unknown strong inline coupling.

The invention discloses a data set denoising processing method based on an empirical mode decomposition algorithm, and the method comprises the following steps: S1, data selection: obtaining an operation monitoring data set of a proton exchange membrane fuel cell, the data set at least comprising an output voltage signal; s2, degradation index determination: performing correlation analysis on the data set, and screening out an output voltage signal which is strongly and negatively correlated with the service life of the proton exchange membrane fuel cell as a performance degradation core index; s3, data sampling simplification: sampling the output voltage signals in the step S2 at equal time intervals to obtain a simplified voltage time sequence; s4, empirical mode adaptive decomposition: performing empirical mode decomposition on the simplified voltage time sequence to obtain a plurality of intrinsic mode function components and a residual component; s5, noise component elimination: eliminating high-frequency noise components from the plurality of intrinsic mode function components; and S6, signal reconstruction: superposing the residual intrinsic mode function component after elimination with the residual component, and reconstructing to obtain a denoised voltage degradation signal. According to the EMD-based denoising processing method, accurate extraction of real degradation features is realized through data screening, simplification, adaptive decomposition and noise elimination.

The invention discloses an edge terminal-oriented memory efficient model fine tuning method and system, which realize decoupling of a backbone network and a fine tuning process by constructing a parallel side network module, and avoid video memory occupation caused by backbone gradient return. A double-adapter module is arranged, modeling is carried out on same-layer features and cross-layer context information, and the feature expression precision is improved. And constructing a feature fusion module, and performing adaptive weighting on different path outputs through learnable gating. A backbone grouping module is integrated, dynamic grouping is carried out on a backbone network based on interlayer similarity, and redundant modules and calculation overhead are reduced. And through the fine tuning execution module, only back propagation and parameter updating are carried out on the side network, and a trunk freezing state is kept, so that the memory and training cost is reduced. According to the method, the memory efficiency and the training speed of the edge fine adjustment process are improved, high-precision model self-adaption is achieved with extremely low calculation burden, and application in the fields of cloud edge collaborative intelligence and the like of unmanned aerial vehicles, robots, vehicle-mounted terminals and the like is effectively supported.

This invention provides a wafer transfer device and a semiconductor processing system. The wafer transfer device includes a carrier stage and a wafer cassette. The wafer cassette includes a wafer cover and a wafer carrier. The wafer carrier is disposed on the carrier stage. The wafer cover has a downward-facing opening and is closable over the wafer carrier. The wafer cover is movable relative to the carrier stage along the Z-axis, allowing it to transition between a closed state and an open state. In the closed state, the wafer cover covers the wafer carrier; in the open state, the wafer cover moves upward to expose the wafer carrier. Therefore, the wafer transfer device of this invention not only improves positioning accuracy and system stability but also helps reduce the cost of the associated transfer robot.

The application discloses a magnetotelluric three-dimensional hybrid optimization inversion method, and steps include: updating a resistivity model; judging whether a nonlinear conjugate gradient iteration satisfies an exit condition; judging whether the gradient of the objective function of two adjacent nonlinear conjugate gradient iterations is less than a preset threshold; making the objective function jump out of a local optimal solution; judging whether a current particle swarm optimization trigger counter is greater than a maximum number of times; generating an updated resistivity model by using an optimal scaling factor; judging whether the RMS of the particle swarm optimization after optimization reaches an exit condition; and outputting a final resistivity model of a detection area. The application effectively solves the dimension disaster problem of the PSO algorithm applied to three-dimensional large-scale inversion.

The invention discloses a security constraint unit commitment method based on tensor completion approximate dynamic programming, and the method comprises the following steps: firstly, building a security constraint unit commitment model based on a Markov decision process; secondly, decoupling the multi-period security constraint unit commitment model into a single-period sub; then, obtaining a value function and a decision function by adopting an approximate dynamic programming algorithm based on tensor completion; according to the approximate dynamic programming algorithm based on tensor completion, the approximate dynamic programming algorithm is improved from the angle of tensor completion, and all value functions and decision functions are approximately obtained by sampling a small number of state variables and decision variables. Decision is made from the angle of tensor complementation, in the face of a discrete state space, the value function of the whole state space can be approximated only by sampling a small number of state points, the calculation burden of an approximate dynamic programming algorithm in the aspect of value function approximation is effectively relieved, and the calculation efficiency of the algorithm can be greatly improved.

This invention discloses an InSAR interferometric network optimization method based on a multi-factor coherence surrogate model. The method includes: constructing a multi-factor feature dataset containing SAR imagery, baseline data, and soil moisture data of the study area; constructing a candidate interferometric pair set, and randomly selecting a subset of sample interferometric pairs from the candidate set; constructing a coherence surrogate prediction model, which uses the multi-factor feature data corresponding to the sample interferometric pair subset and the ground truth values ​​of sample coherence to train the model parameters; inputting the multi-factor feature data corresponding to each interferometric pair in the candidate set into the trained coherence surrogate prediction model and outputting the predicted coherence of each interferometric pair and the optimized interferometric pair network. This invention improves the stability and accuracy of unwrapping and time-series inversion, while also reducing the difficulty of interferometric processing and coherence calculation under large-scale data, and can support the needs of large-scale, long-term, and near-real-time monitoring.

The invention discloses a hybrid control method and system for an interleaved flyback photovoltaic micro inverter. The method comprises the following steps: S1, acquiring a real-time photovoltaic panel voltage, a power grid voltage effective value, a power grid phase and the average output power of the inverter; s2, calculating a gain factor based on the voltage of the photovoltaic panel and the phase of the power grid; s3, according to the gain factor, the network voltage effective value, the transformer excitation inductance, the average output power and the transformer turn ratio, the duty ratio and the switching frequency under the BCM are calculated; s4, comparing the switching frequency with a preset limiting frequency of a BCM mode and a preset fixed working frequency of a DCM mode, and selecting a working mode according to a comparison result; and S5, generating a corresponding control signal according to the working mode, wherein the control signal is used for driving a flyback switching tube and a power frequency flip circuit switching tube. The method has the advantages of reducing the calculation burden of mode switching and the like.

The application discloses a resonant converter control method and device based on threshold voltage control, and the method comprises the following steps: comparing an input voltage with a primary side equivalent voltage to determine whether the resonant converter is currently working in a boost mode or a buck mode; determining threshold voltages of a resonant capacitor in the boost mode and the buck mode; comparing the threshold voltages with collected resonant capacitor voltages to determine switching signals of an inverter circuit; and the threshold voltages of the resonant capacitor in the boost mode and the buck mode are calculated based on P-stage cosine response parameters and N-stage cosine response parameters of the resonant capacitor voltages. By using the above technical scheme, the resonant converter has strong voltage ripple suppression capability, effectively suppresses secondary harmonic voltage ripples from a previous stage, and the control signals are directly determined by input and output states, thereby having excellent dynamic performance.

The invention provides an oil-immersed transformer fault decision-making method and system based on a three-dimensional voiceprint map, and the method comprises the steps: collecting a voiceprint signal of a transformer body through a high-sampling-rate microphone array which is arranged in a standardized manner; an optimized decoding process based on improved inverse discrete cosine transform is innovatively adopted to construct a three-dimensional voiceprint map matrix containing time, frequency and sound intensity information, and the limitation of traditional single feature analysis is broken through; space-time alignment and fusion are carried out on the voiceprint map, the real-time load rate, the oil temperature and other multi-physical field data, a unified high-dimensional feature vector is constructed, and efficient processing and feature mining are carried out on the high-dimensional data through a light-weight sliding window feature extraction mechanism of depth separable convolution reconstruction. More sensitive and robust online diagnosis of early abnormal states and composite faults is realized, and model complexity and calculation burden are remarkably reduced through deep adaptation of a core algorithm and a hardware architecture.

The application discloses an online identification method for fractional order parameters of a transformer based on an Adam-SGD fractional order neural network, and comprises the following steps: acquiring a fractional order state equation representing dynamic characteristics of capacitors and inductors of a DC-DC transformer; establishing a fractional order discrete recursive model based on GL fractional order calculus definition; constructing a neural network structure according to a recursive relationship of capacitor voltage and inductor current in the discrete model, taking actually collected voltage and current values as reference values, taking neural network output as estimated values, and constructing a mean square error loss function; performing gradient calculation and weight updating on the loss function by using an Adam-SGD optimization algorithm, and online adjusting neural network weights; and calculating and outputting parameters of the capacitors and the inductors in real time according to the established fractional order discrete model and the obtained weights. When the weights are updated, the application adopts an adaptive switching mechanism of Adam-SGD, and has the rapid convergence ability of Adam and the stable and fine optimization ability of SGD, so that the efficiency and reliability of online training of the fractional order neural network are significantly improved.

The present application belongs to the technical field of photovoltaic power generation, and discloses a photovoltaic output smoothing regulation method, system, device and medium. First, the original photovoltaic output signal is low-pass filtered to generate a reference action, a state vector is input to a pre-trained PPO reinforcement learning model to output a residual action, and after superposition, a combined action instruction is formed. The combined action instruction is then corrected by a power limit and SOC boundary constraint mechanism and output by a teacher model. Subsequently, based on a distillation loss function, a lightweight student model is trained using the teacher output as a soft label for real-time regulation. In this method, low-pass filtering provides a stable reference, reinforcement learning residual compensation compensates for high-frequency fluctuations and non-stationary characteristics, and the constraint mechanism protects the battery life. The distillation process migrates the knowledge of the complex model to the lightweight model, reducing the computational burden. This method significantly improves the smoothing performance, avoids control failure, effectively protects the battery health, and achieves efficient edge deployment, balancing control performance, life management and engineering feasibility.

This invention belongs to the technical field of motor control and discloses a model predictive control method and system for effectively suppressing the second harmonic of a permanent magnet synchronous motor. The method includes the following steps: obtaining the three-phase input current of the permanent magnet synchronous motor at the current moment, converting the three-phase input current into a sum-i current in a coordinate system; extracting the two orthogonal components of the second harmonic at the current moment from the transfer function of the current input to a second-order generalized integrator; calculating the second harmonic component at the next moment using the two orthogonal components of the second harmonic at the current moment; compensating the fundamental current reference command at the next moment using the second harmonic component at the next moment, obtaining the compensated fundamental current reference command at the next moment, thus achieving the suppression of the second harmonic current at the next moment. This invention solves the problem of not suppressing the second harmonic in the control of permanent magnet synchronous motors.

This invention provides a model predictive control method for a six-phase electric drive reconfigurable on-board charging system, belonging to the field of power electronics and electric drive technology. It solves the problems of high current ripple and heavy computational burden on the system. The technical solution includes the following steps: S1: Obtaining the q-axis current setpoint; S2: Obtaining the real-time 0 / 1 axis current setpoint; S3: Acquiring the six-phase current of the six-phase symmetrical permanent magnet synchronous motor; S4: Calculating the predicted current for the next cycle; S5: Selecting the voltage vector for the next cycle; S6: Obtaining the on / off state of the switching transistors in the next cycle; S7: Realizing the operation mode of the six-phase electric drive reconfigurable on-board charging system based on the drive signal of the six-phase inverter. The beneficial effects of this invention are: it can improve the dynamic and steady-state performance of the system while reducing the computational burden and lowering zero-sequence current ripple, and has broad application prospects.

The present application relates to the technical field of multi-sensor system state estimation, and particularly relates to a kind of cluster multi-sensor network robust state estimation method and system resisting compound interference. Including the following steps: constructing the cluster multi-sensor network comprising multiple estimators and multiple sensors, using encryption algorithm to encrypt sensor measurement data and generate authentication tag, calculate sensor attack strength based on the hamming distance of authentication tag, each estimator based on maximum correlation entropy criterion, using the measurement data of trusted sensor in cluster and sensor attack strength to carry out local state update, obtain posterior state estimation;Each estimator exchanges posterior state estimation with neighbor estimator based on information consistency algorithm, and realizes the consistent fusion state estimation of inter-cluster state estimation through iterative communication.The present application guarantees the confidentiality and integrity of data transmission by integrating lightweight AES-CTR encryption and GMAC authentication mechanism.

The invention relates to the technical field of adsorbent performance optimization, in particular to a CO2 humidifying adsorbent adsorption performance optimization method and system based on multi-feature matching, and the method comprises the steps: constructing a synthetic functional group set, selecting adsorbents based on the synthetic functional group set to obtain a pre-selected adsorbent set, sequentially extracting pre-selected adsorbents from the pre-selected adsorbent set, and extracting the pre-selected adsorbents from the pre-selected adsorbent set; the method comprises the following steps: extracting a pre-selected adsorbent, determining a pre-selected adsorption characteristic parameter group of the extracted pre-selected adsorbent, constructing a simulation model according to the pre-selected adsorption characteristic parameter group to obtain a simulation adsorbent, performing adsorption condition optimization on the extracted pre-selected adsorbent based on the simulation adsorbent to obtain target adsorption data, and summarizing the target adsorption data to obtain a target adsorption data set. According to the method, the efficiency of optimizing the adsorption performance of the CO2 adsorbent can be improved, and the dependence on experimental resources in the optimization process is reduced.

The embodiment of the invention relates to an equipment control method and device, a controller and a storage medium, and the method comprises the steps: obtaining a control parameter sequence of target equipment from a digital base system at a first time interval, and storing the control parameter sequence in a storage module, the control parameter sequence is a plurality of control parameters for the target equipment predicted by the digital base system in a second time period in the future; and under the condition that the communication with the digital base system is abnormal, controlling the target equipment according to the control parameter sequence in the storage module. Therefore, the PLC controller can continuously execute control based on the pre-stored predictive control parameter sequence when the communication is abnormal, so that the influence of network instability on the control continuity is reduced under the condition that the calculation burden of the PLC is not increased, and the problem that accurate intelligent control is difficult to support during the network disconnection period of the PLC controller is solved.

The application discloses a kind of energy-saving control method of frequency conversion speed regulation and energy recovery cooperation rig plunger pump, and relates to industrial control technical field.The method includes: monitoring the current working power of rig plunger pump at first working frequency, when power exceeds preset first working power interval, first working power time sequence is collected;According to the time sequence data, the geological adaptability coefficient is calculated using the pre-trained geological adaptability analyzer;Determine whether the working frequency needs to be adjusted, and determine the second working frequency and the corresponding second working power interval accordingly;Based on the second working frequency, frequency converter, energy storage device and energy recovery device are cooperatively controlled.The application realizes intelligent perception and self-adaptive adjustment to formation change, and significantly improves the energy efficiency and operation economy of rig plunger pump under complex geological conditions through the collaborative management of energy.

The invention discloses a data driving control method based on a sliding window and feature point selection in a time-varying scene, and relates to the technical field of adaptive control of data driving, and the method comprises the steps: collecting input and output signals of a time-varying system in real time through a sensor; dynamically updating an input / output track data structure based on a sliding window mechanism to ensure that the data structure always reflects the current dynamic characteristics of the system; feature point selection is carried out on the trajectory data structure through singular value decomposition, data information is quantized, dimension reduction processing is carried out, and a dimension reduction data structure is generated; constructing a control optimization objective function based on the dimension reduction data structure, solving a decision variable and generating a control instruction; and balancing the data structure dimension and the information completeness through an adaptive order adjustment mechanism. According to the method, the data is updated in real time through a sliding window mechanism, and the data dimension is selected and optimized based on the feature points of singular value decomposition, so that the timeliness, the calculation efficiency and the control precision of time-varying system data driving control are remarkably improved.

This invention discloses an adaptive self-triggered optimal control method for catalytic rod reactors, specifically comprising: a temperature field model of the catalytic rod reaction process described by a high-dissipation partial differential equation system based on the control input of the cooling material under time triggering, and obtaining a low-order slow subsystem characterizing the high-dissipation partial differential equation system after order reduction processing; constructing a corresponding Hamiltonian function based on the performance index of the low-order slow subsystem, obtaining the optimal control strategy under time triggering control based on the Hamiltonian function, and then obtaining the optimal control strategy under self-triggered control, converting the Hamiltonian function into HJB equations; constructing a self-triggered controller based on an evaluation-execution dual-network structure using data-driven adaptive dynamic programming, and outputting the adaptive self-triggered optimal strategy based on the optimal value function of the HJB equations.

This application relates to a method and apparatus for substation equipment condition monitoring based on edge computing. The method includes: deploying edge computing nodes at the substation side, which aggregate real-time condition data collected by monitoring equipment; preprocessing and extracting features from the real-time condition data at the edge computing nodes to obtain equipment feature vectors representing the health status of the equipment; dynamically determining high-risk monitoring areas within the substation based on historical operation and maintenance data and real-time power grid operation requirements, and deploying corresponding lightweight intelligent diagnostic models at the edge computing nodes according to these high-risk monitoring areas; using the lightweight intelligent diagnostic models to perform real-time analysis of the equipment feature vectors to obtain equipment condition diagnostic results including fault types and fault probabilities; and executing tiered early warning or rapid control operations at the edge side based on the equipment condition diagnostic results. This method can improve the diagnostic results of substation equipment condition monitoring.

The application discloses a kind of mission critical monitoring method, device and equipment and storage medium, the method comprises: obtaining task dependency relationship and baseline configuration information and the time information of each task in task dependency relationship;Time information includes scheduling time, running time length and actual start time and actual end time;Task dependency relationship is divided into several connected components;For each connected component, the earliest start time of each task is calculated according to scheduling time and actual start time;The earliest end time of each task is calculated according to the earliest start time and running time length and actual end time;According to each connected component and the earliest end time and baseline configuration information, the predicted completion time of each baseline and critical path are determined;According to predicted completion time and critical path and baseline configuration information, it is judged whether to carry out timeout early warning to each baseline.The method of the application can reduce the computing power of monitoring critical tasks, reduce invalid or repeated alarm.

The invention relates to the field of small target detection, in particular to an improved unmanned aerial vehicle small target detection method based on YOLO26s, and the method comprises the steps: constructing a YOLO26s-SOD model for the small target detection of an unmanned aerial vehicle, replacing a first C3k2 module and a second C3k2 module of an original backbone network with a context anchor point residual attention feature fusion module, replacing a third C3k2 module with an A2C2f module, and carrying out the detection of the small target of the unmanned aerial vehicle. Deleting modules behind the windmill-shaped wavelet transform convolution module and connecting the modules with a first convolution module; a first Upsample module in an original neck network is replaced by a first connection module, a context anchor point residual attention feature fusion module and a DySample module which are connected in sequence, a second Upsample module is replaced by the DySample module, and a C3k2 module in the neck network is replaced by the context anchor point residual attention feature fusion module. A DySample module, a second connection module and a context anchor point residual attention feature fusion module which are connected in sequence are added behind the second C3k2 module; and the space focusing capability on a tiny target is obviously improved.

The application discloses a drill-and-blast tunnel dynamic geological modeling method, medium, equipment and product, relates to the technical field of tunnel geological modeling, and comprises the following steps: according to the tunnel length and the hole diameter range of different construction periods, a tunnel three-dimensional geological model with different precision and resolution is constructed, the tunnel three-dimensional geological model is constructed in the following process: based on geological exploration data and drilling data, a stratum interface three-dimensional model is constructed; based on drilling core data, advanced drilling lithology information and geophysical prospecting data, lithology information is coupled to the stratum interface three-dimensional model to form a lithology three-dimensional distribution; the physical and mechanical parameters of rock-soil bodies are associated with the lithology three-dimensional distribution to realize three-dimensional assignment of the physical and mechanical parameters and form an attribute parameter field model; and the tunnel three-dimensional geological model is dynamically incrementally updated along with construction progress. The application adopts hierarchical models, takes into account the high-precision requirement of a current construction section and the efficiency requirement of a long-term planning section, incrementally updates the models, and improves the working efficiency of the whole modeling process.