11results about How to "Reduce calculation bias" patented technology

The invention discloses a long time sequence information prediction method based on a multi-scale double-source fusion network model, time domain and frequency domain features are modeled at the same time, the ability to understand a complex time sequence is enhanced, an efficient multi-scale attention module of cross-space learning can capture short-term and long-term dependency relationships at the same time, and the prediction accuracy is improved. The modeling capability of different time scale information is improved, dynamic weighted fusion of time domain and frequency domain features is realized by a gating mechanism, and the flexibility and robustness of the model are improved. And through standardization and size projection, feature deviation and memory occupation are reduced, model convergence is accelerated, and the overall memory demand is reduced. The efficient multi-scale attention module accurately captures features, weakens noise and abnormal value influences, and reduces unnecessary calculation and memory consumption. The gating mechanism dynamically adjusts the feature weight, preferentially processes key information, and reduces the memory demand.

The invention relates to the technical field of water purification equipment control, and discloses a filter element service life identification method, device and equipment, a storage medium and a program product, and the method comprises the following steps: in the working process of a filter element, obtaining the water flow, the water production time and the whole machine power-on time of the filter element; detecting whether the flow meter is abnormal or not; if the flow meter is not abnormal, whether the service life of the filter element expires or not is judged according to the water flow and the whole machine power-on time; if the flow meter is abnormal, whether the service life of the filter element expires or not is judged according to the water production time and the whole machine power-on time. The service life of the filter element can be accurately recognized by integrating the water flow, the water production time, the whole machine power-on time and the flow meter state, the situation that the service life is not due, so that the filter element is wasted, or the water quality is affected due to too long use time is avoided, and the user experience and satisfaction degree are improved.

The invention relates to the technical field of slope early warning, and discloses a slope digital twinborn monitoring model construction method based on an agent model, and the method comprises the following steps: S101, constructing a slope geometric model composed of M triangular patches through three-dimensional reconstruction software; step S102, constructing a first training data set for pre-training a pre-proxy model by using the slope geometric model; s103, acquiring monitoring data of the M triangular patches in real time, and constructing a second training data set for training the proxy model; step S104, generating a slope early warning cloud picture; according to the method, the simulation data under the multi-physics field working condition is obtained through the slope geometric model, the training data set of the pre-training pre-agent model is constructed according to the simulation data, the number of training samples of the agent model is greatly reduced, in addition, a plurality of base learners in the agent model can reduce the calculation deviation of a single time sequence analysis model, and the calculation accuracy is improved. Therefore, the prediction precision of the proxy model is improved.

The present application relates to the offshore wind power engineering field, disclose a kind of considering time-space effect's wind power pile foundation scour duration prediction method, comprising: obtaining target sea area environmental parameter and pile foundation basic parameter, calculate hydrodynamics parameter and limit equilibrium scour depth;According to the law of conservation of sediment volume and space net flux, a time-space evolution model is established;Using historical scour measured data to calibrate the undetermined coefficient in the model;Based on the calibrated model, the characteristic time scale is calculated, and the duration prediction chart of the continuous change of scour depth with time is output;The final scour depth at the end of service is compared with the safety depth threshold determined according to the stress state, and the early warning is output.The present application establishes an evolution model by combining the physical mechanism of sediment and calibrates the parameters using in-situ measured data, which improves the rigor and accuracy of the prediction, and also relates to the actual anti-overturning index of pile foundation, providing quantitative engineering basis for the structural safety evaluation and protection design of offshore wind farm.

The present application relates to the technical field of fault location of UHVDC transmission, in particular to a single-ended accurate fault location method for UHVDC transmission system, when single-pole grounding fault occurs in the line, positive and negative voltage and current signals at the installation position of the protection are collected, line mode and zero mode voltage reverse waves are obtained through Kalman-Bell transformation decoupling, the first wave head is positioned by difference and effective time domain data is intercepted, S transform is carried out for time-frequency mapping by using adaptive window width, time-frequency energy redistribution is realized by iterative optimization group delay fixed point, time direction high convergence time-frequency distribution is obtained, double-mode wave arrival time is extracted frequency by frequency and modulus time difference is calculated, initial fault distance sample is obtained by combining frequency variable propagation speed, the probability density peak value is taken as the final fault location result through non-parametric kernel density estimation, the present application does not need double-ended communication and synchronization, and greatly improves the wave arrival time extraction accuracy, anti-interference ability and fault location robustness.

This application provides a method and apparatus for determining blood flow velocity in a blood vessel, comprising: acquiring an initial contrast image and a target contrast image after injecting a contrast agent; comparing the initial contrast image and the target contrast image to determine the length of the blood vessel through which the contrast agent flows; determining the uniformity of the contrast agent in the main vessel region and / or the visualization of microvessels in the background region of the target contrast image based on the contrast agent concentration in the target contrast image; correcting the blood vessel length based on the uniformity of the contrast agent and / or the visualization of microvessels to obtain a corrected blood vessel length; and determining the blood flow velocity based on the corrected blood vessel length. This reduces calculation errors caused by the imaging angle of the contrast image, the shape of the blood vessel, or blood vessel bifurcation, thereby improving the accuracy of blood flow velocity assessment.

The application discloses a multi-agent material calculation consistency calibration method, belonging to the cross field of material science, artificial intelligence and computational physics. The method adopts a master-slave structure multi-agent system, containing one master agent and at least three slave agents. The master agent is responsible for overall coordination, embedding physical mechanism constraints, consistency calibration and output results. The method steps are: constructing a calculation system, introducing physical mechanism constraints and embedding slave agent calculation models, collecting calculation data, performing bias analysis, adjusting parameters or correcting results through iterative calibration, and finally verifying and outputting the final results. The application introduces physical mechanism constraints, combines bias analysis and iterative calibration, and solves the problems of poor consistency of existing multi-agent material calculation results and easy violation of physical laws. Experiments show that the method reduces the deviation of the calculation results and the experimental measured values to below 3%, improves the calibration efficiency by more than 40%, and improves the calculation accuracy by more than 35% compared with the prior art, thereby improving the accuracy and engineering practicability of high-throughput calculation of material performance.

This invention relates to the field of image data processing technology and discloses a real-time generation, comparison, and tracing method for document handwriting for forensic identification. The method involves performing a two-dimensional optical scan of the sample to be identified, calculating the paper-based fiber topology matrix; performing physical medium classification and calculating the microscopic spatial quantization feature sequence; calling a multi-state physical model for inverse mapping to reconstruct instantaneous velocity and acceleration time-series signals; performing frequency domain analysis on the acceleration time-series signals to extract biological tremors and mechanical harmonic energy to calculate the initial screening judgment factor and locate the core physical frequency of vibration; combining the core frequency to extract the actual spatial physical resonance component and generate a theoretical vibration waveform; performing a space-frequency dual-domain closed-loop verification to obtain the maximum phase cross-correlation coefficient; and finally, combining the judgment factor and cross-correlation coefficient to perform logical judgment and output the identification conclusion and parameters of the forged device. This invention can effectively distinguish between genuine handwriting and high-precision mechanical imitation, achieving accurate tracing of handwriting evidence.

The invention relates to a flight data-oriented calculation model and parameter calling management method and system, and the method comprises the steps: obtaining flight test data and to-be-calculated flight subject information; determining a subject calculation model corresponding to the to-be-calculated flight subject information according to the to-be-calculated flight subject information and a calculation model library pre-stored in a system; extracting test parameter values related to the subject calculation model from the flight test data, and / or obtaining self-defined parameter values related to the subject calculation model; and determining an analysis calculation result of the flight subject to be calculated according to the subject calculation model in combination with the test parameter value and / or the custom parameter value. The flight data-oriented calculation model and parameter calling management method and system have the characteristics of easiness in operation, high accuracy and high processing efficiency, and can effectively assist operators in analyzing and calculating flight data.

The application provides a method, device, equipment, medium and product for predicting the adsorption amount of coal bed methane under water-containing conditions, and relates to the technical field of oil and gas field development. The method comprises the following steps: determining the actual water saturation of each pore diameter in a target coal sample based on the pore structure characteristic data of the target coal sample and the nuclear magnetic signal amplitudes of the target coal sample under multiple different water-containing conditions; determining the length of methane occupying in the pores corresponding to each pore diameter in the target coal sample based on the actual water saturation of each pore diameter in the target coal sample; calculating the total volume of methane adsorption and the total surface area of methane adsorption; and determining the total adsorption amount of methane of the target coal sample under the actual water saturation based on the total volume of methane adsorption and the total surface area of methane adsorption. The method of the application fully considers the influence of the water saturation in the coal bed on the adsorption amount of methane, and combines different adsorption modes of methane in the pores, so that the predicted adsorption amount of methane is more accurate.

The present invention relates to the technical field of slope early warning, and discloses a method for constructing a digital twin monitoring model of a slope based on an agent model, comprising the following steps: step S101, constructing a slope geometric model composed of M triangular facets by means of three-dimensional reconstruction software; step S102, constructing a first training data set for pre-training a pre-agent model by means of the slope geometric model; step S103, acquiring monitoring data of the M triangular facets in real time, and constructing a second training data set for training the agent model; step S104, generating a slope early warning cloud map; the present invention obtains simulation data under multi-physical field conditions by means of a slope geometric model, and constructs a training data set for pre-training a pre-agent model according to the simulation data, thereby greatly reducing the number of training samples of the agent model; in addition, multiple base learners in the agent model can reduce the calculation deviation of a single time series analysis model, thereby improving the prediction accuracy of the agent model.