53 results about "Ensemble Kalman filter" patented technology

A method, system and apparatus for real-time reservoir model updating using ensemble Kalman filters is described. The method includes a conforming step for bring bringing static and dynamic state variables into conformance with one another during a time step of the updating. Also, an iterative damping method is used in conjunction with the conformance step to account for nonGaussian and nonlinear features in a system. Also, a re-sampling method is described which reduces the ensemble size of reservoir models which are to be updated.

The invention discloses an airplane angular rate signal reconstruction method based on unscented kalman filter. The method comprises the following steps of: (1), denoising an attitude angle measurement signal containing measurement noise through a nonlinear tracking differentiator to obtain a differential signal of an attitude angle, and constructing a virtual measurement equation; (2), constructing a state equation of a system on the basis of natural characteristics of an airplane and according to tri-axial angular rate signals and torque characteristics of the airplane; and (3), reconstructing an angular rate signal of the system by selecting the unscented kalman filter technology as the processing manner because both the state equation and the measurement equation are nonlinear and the estimation precision can be influenced by traditional operation for linearizing a nonlinear equation.

The invention discloses an ensemble kalman filter-based particle filtering method, which comprises the following steps of: the initialization of sampled particles, the prediction of a background ensemble at a k moment, the calculation of a kalman gain, the fusion of the latest observation information, the updating of the background ensemble, the recalculation and analysis of the ensemble, the construction of a suggested Gaussian distribution function, the normalization of weights, and the like. The ensemble kalman filter-based particle filtering method provided by the invention avoids the calculation of a Jacobian matrix because a nonlinear system is not required to be linearized, and improves the calculation accuracy by adopting a sampling method to realize approximately linear distribution. In the particle filtering method, a sampling point number is heuristic and can be flexibly set, so the amount of calculation is not increased sharply with the increasing of dimensionality, and the real-time performance is effectively controlled.

The invention relates to a method for history matching a facies geostatistical model using the ensemble Kalman filter (EnKF) technique. The EnKF is not normally appropriate for discontinuous facies models such as multiple point simulation (MPS). In the method of the invention, an ensemble of realizations are generated and then uniform vectors on which those realizations are based are transformed to Gaussian vectors before applying the EnKF to the Gaussian vectors directly. The updated Gaussian vectors are then transformed back to uniform vectors which are used to update the realizations. The uniform vectors may be vectors on which the realizations are based directly; alternatively each realization may be based on a plurality of uniform vectors linearly combined with combination coefficients. In this case each realization is associated with a uniform vector made up from the combination coefficients, and the combination coefficient vector is then transformed to Gaussian and updated using EnKF.

The invention relates to a rapid collective Kalman filtering assimilating method for real-time data of high-frequency observation data. The method comprises: collecting the high-frequency observation data and controlling the quality; calculating an observation error covariance matrix; obtaining the error covariance matrix of background fields by calculating a forecast trend, i.e. the difference value of the adjacent background fields; utilizing the covariance matrix, the error covariance matrix of the background fields, the observation data and the background fields currently obtained by the calculation of a marine numerical model so as to carry out the real-time assimilation on the observation data of different moments, assigning the updated analysis field to the initial field of the next-moment integral and continuously forecasting forwards; and repeating the operations, thus realizing the real-time assimilation on the high-frequency observation data of different moments in the integral course. The assimilating method has the advantages that the real-time assimilation of the high-frequency observation data is realized; the assimilation efficiency of the data is enhanced; the defect that a large amount of collective models are simultaneously operated in the implementation course of the traditional EnKF (ensemble kalman filter) is overcome; the problem of non-convergence is avoided; and the purposes of accurate numerical simulation and marine forecasting are reached.

The invention relates to a micro-satellite attitude determination method based on ESOQPF (estimar of the quaternion particle filter ) and UKF(unscented kalman filter) master-slave filtering, which comprises the following steps: firstly establishing a micro-satellite system state space model for attitude determination and filtering initialization; taking an attitude quaternion as a state variable, taking an ESOQPF as a master filter for estimating an attitude quaternion, and converting the estimated quaternion as a corresponding attitude angle; and taking gyro drift as a state variable, and taking a UKF as a slave filter for estimating gyro drift. The micro-satellite attitude determination method reduces calculated quantity and shortens attitude determination process while ensuring attitude estimation accuracy in two aspects: firstly, the ESOQPF and the UKF are employed for master-slave filtering so as to separately estimate the attitude quaternion and the gyro drift; secondly, by combination of a QPF algorithm and an ESOQ2 algorithm while estimating the attitude quaternion by the ESOQPF, a Davenport matrix is calculated by the QPF algorithm and the characteristic vector corresponding to the maximum characteristic value of the matrix is directly calculated by the ESOQ2 algorithm. The micro-satellite attitude determination method is applicable to micro-satellite attitude determination based on vector observation in a gyro configuration mode.

The invention discloses a side slip angle estimation method and system based on robust unscented kalman filtering. The method includes the steps that a three-freedom-degree vehicle dynamics model related to a side slip angle is established, a state equation and an observation equation of an unscented kalman filter are determined according to the model, and the input quantity, the state quantity and the observation quantity are determined, wherein the input quantity comprises the front wheel rotation angle and the longitudinal accelerated speed, the state quantity comprises the side slip angle,the yaw velocity and the longitudinal vehicle speed, and the observation quantity comprises the lateral accelerated speed and the yaw velocity; the weight factor is calculated through an M estimatoralgorithm; and in combination with the weight factor, the state quantity at the current moment, the state equation and the observation equation, the side slip angle at the next moment is estimated through an unscented kalman filtering algorithm. The M estimator algorithm and the unscented kalman filtering algorithm are combined to restrain the influence caused by outliers, the robustness of the unscented kalman filtering algorithm to the outliers in observation signals is improved, and the estimation accuracy of the vehicle side slip angle is improved.

The invention discloses an optimum navigational parameter fusion method based on three-level filtering under redundant sensor configuration. According to the invention, aiming at the characteristics of information redundancy of a navigation sensor, a three-level filtering framework for the navigation sensor is designed; through local estimation of a first-level kalman filter, the global estimation of a second-level federated filter and the optimum global estimation of a third-level filter, the optimum global estimation of the navigational parameter of a flight management system can be performed by fully using redundant navigation sensor information. The method disclosed by the invention is beneficial to full utilization of the redundant navigation sensor information, meanwhile, improves the accuracy of navigational parameter estimation, and is such a method that facilitating engineering realization. The method has an important practical application significance on meeting navigation performance requirements required by all legs for large aircrafts flying in civilian areas.

The invention provides a river network water flow quality real-time prediction method and device based on data assimilation. According to the technical scheme, on the basis of real-time monitoring data such as the water level, the flow and the water quality of the river network, real-time monitoring data are assimilated into the river network water flow and quality model through the ensemble Kalman filter or the improved algorithm of the ensemble Kalman filter, the river network water flow and quality data assimilation model is constructed, and the computing efficiency of the data assimilationmodel is improved through a parallel computing architecture. Due to the data assimilation technology, water flow and quality model structure errors, inlet and outlet boundary errors and observation value errors are comprehensively considered; real-time observation data are fused to dynamically correct water level, flow and water quality concentration state variables and roughness and water quality parameters of the water quality model, so that the water level, flow and water quality concentration of a complex river network system can be dynamically predicted by adopting a river network waterquality data assimilation model and a parallel computing architecture, and the model prediction precision can be improved.

The invention belongs to the field of measured ocean environment data assimilation, and specifically relates to an adaptive compensation method for a static localization scheme of the ensemble Kalman filter. The method includes: pre-processing measured data of the atmosphere ocean environment; calculating subsequent required thresholds for each observed data according to a real-time observation system by employing observation errors, observation numbers, and significance levels of different observation elements; assimilating all the observation data in sequence; calculating the ensemble average and the ensemble perturbation; and calculating the observation margin and updating ensemble members. According to the method, the static localization method in the ensemble Kalman filter is improved, multi-scale information which is not extracted by the static localization method in the observation information is effectively extracted, and the assimilation precision of the ensemble Kalman filter is substantially improved.

The present invention belongs to the field of measured marine environment data assimilation, and specifically relates to an ensemble Kalman filter localization method. The present invention comprises: preprocessing measured atmospheric and marine environment data; for each piece of observation data, calculating a priori observation ensemble member of the observation data; calculating the average and variance of an observed priori ensemble; calculating an average observation increment of an ensemble; calculating a perturbed observation increment of each ensemble; projecting an average observation increment of the ensemble onto the ensemble average with a mode state; projecting the perturbed observation increment of each ensemble onto a corresponding ensemble perturbation with the mode state; and obtaining an ensemble member analysis field. According to the method provided by the present invention, the traditional localization method for ensemble Kalman filter is improved, different spatial scales represented by the ensemble average and ensemble perturbation are effectively considered, and the assimilation accuracy of ensemble Kalman filter is significantly improved.

The invention discloses a utility tunnel gas leakage concentration field prediction and correction and leakage rate estimation method. The utility tunnel gas leakage concentration field prediction andcorrection and leakage rate estimation method integrates real-time monitoring data of a gas sensor, a wind speed and direction sensor and a temperature and humidity sensor inside a utility tunnel, structures a numerical calculation model of gas leakage and diffusion inside the utility tunnel and applies a data assimilation method to predict and estimate of a gas leakage and diffusion process andthe gas pipe leakage rate, wherein the data assimilation method is an improved ensemble Kalman filter algorithm, and the state vectors of the improved ensemble Kalman filter algorithm are composed ofgas concentration and the leakage rate. The utility tunnel gas leakage concentration field prediction and correction and leakage rate estimation method combines monitoring data of the wind speed and direction sensor and the temperature and humidity sensor, can inhibit the influence of uncertainty of wind flow speed and diffusion coefficient on results after adding noise meeting practical situations into the wind flow speed and the diffusion coefficient, and can acquire the internal gas concentration distribution inside the utility tunnel and meanwhile accurately perform inverse estimation on the gas pipe leakage rate, thereby well meeting the practical situations.

The invention discloses a dynamic data driven forest fire spreading prediction method. A disturbance generation set is applied to the position of a live wire; forward prediction of the set is achievedthrough FARSITE software; an Ensemble Kalman Filter algorithm is used for carrying out dynamic correction on a prediction result by integrating observation data, so that the accuracy of simulation and prediction of FARSITE forest fire spreading software is improved. At the same time, the Ensemble Kalman Filter algorithm calculates a prediction error covariance of the state through a Monte-Carlo method; the problem of difficulty in estimating and forecasting a background error covariance matrix in the practical application is solved with the idea of the set, thus program codes can be compiled,debugged and maintained easily, and the program codes can be calculated in parallel; and the method is simple and efficient in implementation.

The invention discloses a numerical weather forecast mixed data assimilation method based on triple multi-layer perceptrons, and the method comprises the following steps: building a first perceptron model based on a three-dimensional variational data assimilation method, and carrying out the training of the first perceptron model; establishing a second perceptron model based on an ensemble Kalmanfilter data assimilation method, and training the second perceptron model; establishing a third perceptron model of hybrid assimilation; training the third perceptron model; and calculating by using the trained first perceptron model, second perceptron model and third perceptron model to obtain analysis field data. According to the method, the physical law of atmospheric motion and the time characteristic of atmospheric state variables are utilized; a traditional data assimilation method is simulated, optimized and mixed, the flow dependence of the atmospheric variables is fully considered, time characteristics contained in atmospheric data are introduced accordingly, and the assimilation performance of the method is effectively improved compared with a traditional method.

The invention proposes a flood forecasting method based on ensemble Kalman filtering. The method comprises the steps of interfering with an intermediate state quantity, rainfall and evaporation amountof a selected basin to form a sample set, selecting an average value of a part of set closest to a measured value in a model prediction set to substitute a real value to calculate and obtain a prediction covariance matrix, then obtaining a Kalman gain matrix with the combination of the model prediction set according to the variance of the measured value, at the same time, updating a predicted value to obtain a model analysis value. According to the method, the certain improvement of the set Kalman filtering is carried out, the correction effect is better, a better flood prediction result is achieved, and the flood prediction accuracy is improved.

The invention provides an ensemble Kalman filter dynamic reservoir history matching method based on hyper-sphere transform. The method comprises the following steps: 1, initializing a set of reservoir models, wherein the set comprises static parameters, dynamic parameters and oil well production data; 2, performing hyper-sphere transformation on permeability in the static parameters, and constructing a novel state vector set; 3, inputting each state vector in the novel state vector set into a reservoir simulator for predicting so as to obtain a state prediction value, wherein each state vector and the state prediction value thereof form a prediction set; 4, calculating a Kalman gain matrix according to the prediction set; 5, updating the prediction set according to the prediction set, the Kalman gain matrix and the measured oil well production data, thereby obtaining the updated static parameters, dynamic parameters and oil well production data. According to the method disclosed by the invention, the accuracy of reservoir history matching precision can be improved, and the blindness of manual history matching can be reduced.

The invention discloses a data assimilation method based on adaptive covariance expansion, and the method comprises the steps of obtaining an atmospheric observation value, and carrying out the mode integration based on a t-1 analysis value, and obtaining a forecast field at an analysis moment t; according to the forecast field set, estimating to obtain an ensemble forecast error covariance matrix Pt and an expansion factor vector at the moment t; updating ensemble members of ensemble Kalman filtering in the assimilation initial process by using the expansion factor vector to increase the ensemble variance of new ensemble Kalman filtering to form new ensemble members, and performing iterative updating on the new ensemble members by using an ensemble Kalman filtering method to obtain final analysis ensemble members; and carrying out pattern forecasting by using analysis set members as initial fields. According to the invention, the expansion factor is adjusted, so that the updated prediction error covariance matrix conforms to the statistical relationship of the prediction error covariance, the information quantity and the observation error covariance, the calculated expansion factor is more reasonable, and the assimilation performance is obviously improved.

The invention discloses a low-sensitively ensemble Kalman filtering-based induction motor state monitoring method. The invention aims to solve the technical problem of the low precision of an induction motor monitoring result obtained by an existing method. The method includes the following steps that: the state equation of an induction motor system is established; the measurement equation of theinduction motor system is established; the state equation and measurement equation are a discretized; and low-sensitively ensemble Kalman filtering is performed on the discretized state equation and measurement equation, so that the stator current, rotor flux and angular velocity of the induction motor are outputted. Compared with EnKF, the low-sensitively ensemble Kalman filtering-based inductionmotor state monitoring method of the invention adopts a low-sensitivity optimal control technology to eliminate the uncertainty of parameters (such as rotor resistance and stator resistance) in the induction motor system, so as to decrease the sensitivity of the estimation of the states (stator current, rotor flux and angular velocity) of the induction motor system to uncertain parameters, and improve the state monitoring accuracy of the induction motor. The method can be applied to the state PID control process of the induction motor.

The invention relates to a an unmanned ship water quality detection work path dual-radar accurate positioning method comprising the steps that step 1) the geometrical position of the radar observationpoint is set and the distance and radial speed data of the unmanned ship are acquired in real time through dual radar; step 2) an observation equation of unmanned ship position state and dual-radar information sharing is established; and step 3) the unmanned ship movement trajectories of dual-radar information sharing are recursively solved based on the idea of ensemble Kalman filter, and the unmanned ship water quality detection work path is accurately positioned. The beneficial effects are that high-accuracy positioning of the unmanned ship water quality detection work path can be realized.

The invention provides a ground surface deformation monitoring method and system based on multi-source monitoring data fusion. According to the invention, an ensemble Kalman filtering assimilation mode taking high-precision level settlement data as a main body is adopted; fusion of a radar image differential surface element settlement result of the mining subsidence area and discrete actual measurement level data is realized; high-precision monitoring of the mining area is realized; the high precision of the level inversion value in the center of the basin is maintained, and the D-InSAR differential data is fused, so that the superiority of the D-InSAR data in boundary monitoring is further retained; high-precision seamless subsidence monitoring of the whole mining area is realized, and the transitivity of the fused surface subsidence deformation data in space is more consistent with the actual mining area deformation area.

The invention provides a soil temperature and humidity data assimilation method based on EnPF. The method comprises the following steps that firstly, the weight thought in the particle filtering is used in ensemble Kalman filtering by analyzing Ensemble Kalman filtering and particle filtering, and dual-sampling ensemble particle filtering is built; based on the magnitude of the weight, the particle degeneracy conditions are judged, if particles are degenerated, the particles which are extremely low in weight are removed, the remaining particles are subjected to weighted averaging, and samplingis conducted again in the analysis stage; if the particles are not degenerated, the predicted values of all the particles and the corresponding weights are reserved, and the predicted values of the particles are updated in the analysis stage; secondly, combined with a land hydrological model, soil temperature and humidity data assimilation is conducted, and the simulated overview of the soil temperature and humidity is summarized. In the method, the weight thought in the particle filtering is used in ensemble Kalman filtering, the distribution of observation errors is not specifically assumed, the land hydrological data assimilation of the soil temperature and humidity is conducted, and the simulating precision of the model is improved.

Disclosed is a method of conditioning one or more parametric models. The method comprises obtaining a plurality of candidate parametric models, each describing a sequence of domains characterising a subsurface region and determining whether each sequence of domains described by one or more of said candidate parametric models is a valid sequence of domains. For each valid sequence of domains, each candidate parametric model describing that valid sequence of domains (or a subset of these models) is conditioned simultaneously, for example by using an Ensemble Kalman Filter or artificial neural network.

The invention discloses a tide numerical model water depth estimation method based on ensemble Kalman filtering. The tide numerical model water depth estimation method comprises the following steps: (1) determining ocean numerical model water depth parameters; (2) generating a set and carrying out numerical simulation: superposing an unbiased Gaussian random number on each water depth increment parameter to generate a water depth increment parameter set so as to generate a water depth parameter set, and substituting the water depth parameter set into the ocean numerical mode to carry out free integration until the ocean numerical mode is stable; (3) assimilating by adopting an enhanced parameter correction data assimilation method; (4) after assimilation is finished, freely integrating the ocean numerical mode to be stable by using assimilated parameters so as to obtain an optimized ocean numerical mode state variable; and (5) carrying out harmonic analysis on the ocean numerical mode state variable to obtain an optimized tide harmonic constant analysis result for tide forecasting. According to the method, observation data are assimilated into the tide numerical model through the EAKF, the water depth parameters are optimally estimated, and the tide simulation precision is improved.