38results about How to "Increase inversion speed" patented technology

The invention discloses a microearthquake inversion algorithm which is a combined inversion positioning method based on grid search and Newton iteration. The algorithm is roughly divided into two parts, according to the first part, grid division is performed in the area where a microearthquake event happens, and the rough position where the microearthquake event happens is obtained through the grid search method; the second part is the Newton method, and the initial iteration value during Newton method inversion is the result of the grid search method in the first part. The defect that the Newton iteration method excessively depends on the initial iteration value is overcome, the number of iterations is reduced, and a target function can be well converged close to a theoretical true value finally.

The invention relates to a gravity satellite formation orbital stability optimization design and an earth gravity field precision inversion method on the basis of disturbing inter-satellite distance principle, in particular to an orbital stability optimization design method for a four-satellite system (FSS). In order to guarantee the stability of the four-satellite system, the quantity of satellite orbits is optimally designed, orbital semi-major axes, orbital eccentricities, orbital inclinations and right ascensions of ascending nodes keep unchanged, the difference of arguments of perigees each pair of satellites and the difference of mean anomalies of the pair of satellites are 180 degrees respectively, an initial argument of perigee of each satellite is arranged at the equator, an initial mean anomaly of each satellite is arranged at a pole, and the ratio of the semi-major axis of each elliptical orbit of the four-satellite system to a semi-minor axis of the elliptical orbit of the four-satellite system is 2:1. The gravity satellite formation orbital stability optimization design and the earth gravity field precision inversion method have the advantages that an earth gravity field is precisely and quickly inverted on the basis of a disturbing inter-satellite distance process; and the orbits are high in stability owing to the method, the earth gravity field computation precision is effectively improved, the gravity field inversion speed is increased to a great extent, and requirements on the performance of a computer are low.

The invention discloses a seismic inversion method and system based on generalized total variation regularization, so as to solve the problem that step effects are generated inside a formation when the existing seismic inversion technology adopts total variation regularization. The method uses the characteristic that generalized total variation not only uses first-order partial derivative information of a to-be-inverted parameter but also uses second or higher-order information, the formation interface can be clearly expressed and the step effects inside the formation can also be weakened; besides, a phenomenon that a product of a wavelet matrix, a difference matrix and the natural logarithm of the to-be-inverted parameter can be used to be converted to the convolution among a convolutionkernel corresponding to the wavelet matrix, a convolution kernel corresponding to the difference matrix and the natural logarithm of the to-be-inverted parameter is found out, the convolution operation in a time domain is converted to point multiplication operation in a frequency domain, and the inversion speed is greatly improved; a reflection coefficient does not need to be inverted firstly, anddirect inversion is carried out in the frequency domain to obtain the to-be-inverted parameter; and the method and the system are applicable to the technical field of seismic exploration.

The invention relates to a method for accurately detecting the earth gravity field, in particular to a method which includes: based on the load error analysis theory, accurately building the distance error between satellites of a K wave band distance meter, the satellite orbit position error and the orbital velocity error of a global positioning system (GPS) receiver and an error model in which accumulative geoid accuracy is affected by the nonconservative force error coalition of a satellite-bone accelerometer, and further accurately and rapidly inversing the earth gravity field. The method is high in inversion accuracy of the earth gravity field, simple in satellite gravity inversion process, low in performance requirements of a computer and definite in physical meanings of a satellite observation equation, and effectively improves inversion speed on the premise of ensuring calculation accuracy. The satellite gravity inversion method based on the load error analysis theory is an effective method for calculating the earth gravity field which is high in accuracy and spatial resolution.

The invention relates to a nuclear magnetic resonance echo data compression method and device. The compression method comprises the steps that nuclear magnetic resonance echo data to be analyzed are acquired; the main components of the nuclear magnetic resonance echo data are acquired by utilizing the nuclear magnetic resonance echo data; a compression matrix is constructed by utilizing the main components of the nuclear magnetic resonance echo data; and the nuclear magnetic resonance echo data is compressed by utilizing the compression matrix. According to the technical scheme, information with extremely low contribution to representation of stratum characteristics and even redundant information in the original echo data can be effectively eliminated so that the characteristics of the original data can still be maintained under the condition of high compression ratio, and inversion speed can be effectively enhanced and computational memory can be reduced under the condition of being without losing accuracy.

The invention relates to a large inland lake and wetland water level inversion method based on satellite altimetry data. The method comprises the steps that echo data is preprocessed, a dominant riseleading edge in an echo waveform obtained after preprocessing is extracted, and echo waveform resetting is performed; continuous echo data in multiple trajectories are arbitrarily selected in a research region, different thresholds are used to perform echo resetting, and improved values corresponding to reset echoes are calculated; the maximum improved value is found out, and the threshold corresponding to the maximum improved value is used as an optimal threshold; the optimal threshold is used to perform echo waveform resetting, and water levels are calculated; all the water levels in the research region are post-processed; and last, data verification is performed on time-sequence water levels, whether precision obtained after data verification is within an allowed range or not is judged,and if yes, the time-sequence water levels are output; or otherwise echo resetting is continued. The water level inversion method is simple in calculation, high in speed and high in precision.

The invention discloses a method for increasing the inversion speed of concentration of a greenhouse-gas vertical column. The method increases the inversion speed of concentration of the greenhouse-gas vertical column through two ways of increasing a spectrum modeling speed and a continuous calculation speed for the concentration of vertical column. In the spectrum modeling process, the modeling speed is increased by ignoring the influence of some weak interference factors and discretized radiative transfer equations. In the continuous inversion process of concentration of the greenhouse-gas vertical column, a table look-up method is adopted to increase the calculation speed of the concentration of the greenhouse-gas vertical column: the solar normalized radial brightness and weighting function of all possible scenes are simulated firstly, correction factors are counted when all spectra adopt the value of simulation of one specific scene for inversion fitting, and a correction chart is made; during follow-up data processing, the measurement spectra of all scenes can adopt the value of simulation of the specific scene for fitting, only the fitting result multiplies by corresponding correction factors in the chart, and the measurement spectra of all scenes do not need modeling.

The invention provides a method and a device for quickly and automatically generating a seismic source mechanism solution. The method comprises the following steps of: performing seismic facies pickup on real-time seismic monitoring data by using a seismic facies recognition model to obtain a seismic facies category; identifying the initial motion polarity of the real-time seismic monitoring data by using the initial motion polarity identification model to obtain the initial motion polarity of the P wave; carrying out seismic correlation on the seismic phase category to obtain a seismic correlation result; obtaining an epicenter position according to an earthquake correlation result; and according to the seismic correlation result, the epicenter position and the initial motion polarity of the P wave, carrying out inversion on the seismic phase mechanism to obtain a seismic source mechanism solution. According to the method, information such as P and S arrival time and P wave initial motion polarity is efficiently and accurately obtained by using the seismic phase identification model and the initial motion polarity identification model, and the seismic correlation result and the epicenter position are obtained based on the information, so that the workload of inverting the seismic source mechanism can be greatly reduced, and the inversion speed of the seismic source mechanism solution is improved.

The invention provides a method and a device for automatically inverting a seismic source mechanism solution. The method comprises the following steps: calculating theoretical P wave arrival time and theoretical S wave arrival time of each station according to a seismic directory and a one-dimensional velocity structure model; intercepting the seismic waveform data to be studied according to the seismic directory, the theoretical P wave arrival time and the theoretical S wave arrival time to obtain intercepted seismic waveform data; performing seismic phase information extraction on the intercepted seismic waveform data by using a seismic phase identification model and an initial motion polarity identification model to obtain a seismic phase category and the initial motion polarity of a P wave; and according to the seismic directory, the seismic phase category and the initial motion polarity of the P wave, carrying out inversion on the seismic source mechanism to obtain the seismic source mechanism solution. Information such as P and S arrival time and P-wave initial motion polarity is efficiently and accurately acquired by using the seismic phase identification model and the initial motion polarity identification model, the workload of inverting a seismic source mechanism can be greatly reduced, and the inversion speed of a seismic source mechanism solution is improved.

The invention provides a carbon dioxide concentration inversion optimization method and system. The method includes the steps of 1, obtaining data from a greenhouse gas detecting instrument and a cloud detection and aerosol detecting instrument, removing cloud pixels in H2 carbon dioxide L1B data by utilizing CAPI L2 level cloud detection product and aerosol product, and screening out pixels whichcan be used for carbon dioxide concentration inversion; 2, screening pixel data of different earth surface types for inversion according to carbon satellite observation modes; 3, equally distributingthe screened-out pixels which can be used for carbon dioxide concentration inversion to multiple computational nodes for inversion; 4, enabling a single computational node to start a number of subprocesses through a multi-process parallel technology and meanwhile conduct carbon dioxide concentration inversion, and writing inversion results into a Redis in-memory database; 5, reading inversion results in the Redis in-memory database and generating a carbon dioxide concentration inversion product file and a browsing graph.

The invention discloses a rapid microwave imaging method. In each iteration process of a DBIM algorithm, an inversion speed is improved by calling an efficient forward problem solver to calculate a forward problem. Meanwhile, in order to enable a whole inversion algorithm to stably carry out iterative inversion, a truncated singular value decomposition (TSVD) technology is adopted, and a large number of numerical tests are prevented from being used for determining regularization terms. In order to further reduce a calculated amount, in a cost function of an inversion algorithm, an incident field is used to replace a total field. Therefore, the method can achieve quick and stable inversion, and an inversion speed and quality of an image are ensured.

The embodiment of the invention provides an inversion method of a transient electromagnetic resistivity. The method is characterized in that the method comprises: (1), an initial inversion model is set, inversion is carried out on the initial inversion model, and a corrected inversion model is obtained (2), layer number refining, inversion, and correction are carried out on the corrected inversion model; (3), the step (2) is repeated until the number of times of the layer number refining of the corrected inversion model reaches a preset value or the fitting error of the corrected inversion model response and the observation data is less than a first threshold value. According to the embodiment of the invention, the method has the following beneficial effects: refining is carried out on the inversion model gradually, so that the dependency on the initial model by the inversion process is reduced and the convergence of the inversion process is improved. Moreover, the number of times of iterations is reduced to the certain extent and the inversion speed is accelerated.

The invention provides a compression method and device of multidimensional nmr logging data. The method comprises that multidimensional nmr echo string data and a multidimensional nmr logging kernel matrix are obtained; a window-averaging method is used to compress the echo string data and the kernel matrix preliminarily; one-dimensional discrete cosine transform is carried out on the preliminarily compressed echo string data to obtain a first discrete cosine transform coefficient and first low-frequency data; two-dimensional discrete cosine transform is carried out on the preliminarily compressed kernel matrix to obtain a second discrete cosine transform coefficient and second low-frequency data; and one-dimensional discrete cosine inverse transformation is carried out on the first low frequency data, two-dimensional discrete cosine inverse transformation is carried out on the second low frequency data, and compressed multi-dimensional nmr logging data is obtained. The compression method is high in compression ratio, the computational complexity of an inversion process of the nmr logging data is reduced, and the inversion speed is improved.

The invention discloses a formation interface real-time extraction method based on while-drilling azimuth electromagnetic wave logging data. The formation interface real-time extraction method comprises the steps that s1, sliding windowing is conducted on the logging data of a large-slope well / a horizontal well; s2, if shallow detection geological signals exist, the s3 is executed, and otherwise the s5 is executed; s3, single-interface inversion is conducted through the shallow detection geological signals; s4, the optimal solution of the single-interface inversion is input into the s7 to be used as known constraint; s5, a dual-interface inversion model is established; s6, whether proximity interface information exists or not is judged, if yes, the s7 is executed, and otherwise the s8 is executed; s7, multi-initial-value selection is conducted on far formation interface positions and far surrounding rock resistivity in the model in the s5, and then the s9 is executed; s8, multi-initial-value assignment is conducted on all to-be-inverted parameters in the model in the s7; s9, on the basis of deep detection geological signals, the optimal solution of double-interface inversion is obtained; and s10, real-time imaging processing is conducted on inversion results. According to the formation interface real-time extraction method, fast and accurate interface information can be provided for real-time geological guidance.

The invention relates to a multi-dimensional geological background model construction method for optimizing electrical prospecting inversion. According to the method, electrical stratification is carried out on a stratum, and a longitudinal geological background model with layer resistivity, layer thickness and layer interface parameters at a drilled hole is constructed; geological background model grid units are divided by taking the coordinate positions of the drill hole and adjacent drill holes in an electrical prospecting area as base points; and a weight coefficient in each direction is calculated according to distance between unknown points and surrounding base points so as to solve parameters such as layer resistivity, layer thickness, layer interface and the like at each unknown point, establish a corresponding geological background model, and form a geological background model with a single-point set, a grid unit set and a whole measurement area which represent the continuouschange characteristics of a sedimentary stratum.

The invention discloses a rapid processing method for array lateral logging data of an anisotropic formation of an inclined shaft. The method comprises the following steps: establishing a forward model based on a three-dimensional finite element forward algorithm, analyzing the sensitivity of an array lateral logging response to anisotropic formation parameters of the inclined shaft, constructingan array lateral logging response database, correcting an actual measurement curve and carrying out layering; selecting a corresponding database according to block formation information, establishingan inversion initial value selection plate in combination with sensitivity analysis, rapidly determining initial values of various parameters, establishing an inversion model, simulating an array lateral logging inversion curve after initialization, calculating a two-norm between the array lateral logging inversion curve and the corrected actual measurement curve, judging whether the precision requirement is met or not, if the precision requirement is met, determining the optimal anisotropic formation parameter of the inclined shaft, and if the precision requirement is not met, constraining the inversion model according to reservoir prior information, iteratively updating the formation parameters by adopting a regularized Levenberg-Marquardt algorithm until the precision requirement is met, and determining the optimal anisotropic formation parameter of the inclined shaft. The method lays a foundation for fine evaluation of complex reservoirs.

The invention discloses a Two-dimensional nuclear magnetic resonance rapid inversion algorithm based on LSQR-RSVD, and belongs to the technical field of nuclear magnetic resonance inversion. The method comprises the following steps: designing and generating a forward modeling model suitable for two-dimensional nuclear magnetic inversion; setting a certain number of iterations to invert the forwardmodeling model by using the LSQR; setting the sub-matrix dimension of the RSVD algorithm, and performing inversion by taking the output of the LSQR as the input of the RSVD algorithm; and calculatingthe inversion time, the occupied memory and the relative error and generating a two-dimensional spectrum. The invention provides a novel fast two-nuclear magnetic resonance spectrum inversion methodbased on random singular value decomposition (RSVD). A calculation result of a non-negative least squares (LSQR) is taken as an inversion initial value of the RSVD, so that the memory requirement canbe remarkably reduced and the inversion speed can be increased when two-dimensional nuclear magnetic inversion is carried out, the inversion precision can be ensured, the calculation time and the hardware cost are effectively reduced, the design is reasonable, and a good effect is achieved.

The invention discloses a power consumption attack resistance method based on a d+1-order mask, which comprises the following steps of: splitting a plaintext into a plurality of parts, enabling all mask exclusive-OR results after splitting to be equal to the plaintext, and enabling any one part to be independent from an unsplit plaintext; generating a random value mask, and updating the mask to obtain a random mask containing plaintext information; And constructing a d + 1-order security S box replacement algorithm on the basis of security multiplication to construct an integral d + 1-order mask implementation method. The method has the advantages that high-order power consumption attack protection is carried out on the lightweight block cryptographic algorithm, so that the lightweight block cryptographic algorithm can effectively resist high-order power consumption attack, and meanwhile, required calculation and space resources are few, the efficiency is high, and the method is suitable for cryptographic algorithm hardware implementation.

The invention relates to an inversion method and system of temperature and pressure profiles. The method includes: retrieving corresponding data from a data set according to selected month information and longitude and latitude information, and carrying out classification and library storage on the retrieved data to obtain a corresponding training sample library; obtaining a corresponding training sample set from the training sample library by matching according to occultation observation information, and generating a corresponding simulation light spectrum according to the training sample set obtained by matching and a model invoked for selection; and calculating and obtaining a regression coefficient of a corresponding observation spectrum according to the training sample set obtained by the matching, the corresponding simulation spectrum and a selected statistical regression model, and calculating and obtaining the corresponding temperature and pressure profiles according to a regression system and the satellite observation spectrum. The inversion method of the temperature and pressure profiles of the invention is suitable for use in a sensor of occultation observation geometry, and meets requirements of the service running system timeliness at the same time.

The invention discloses a method for increasing the inversion speed of concentration of a greenhouse-gas vertical column. The method increases the inversion speed of concentration of the greenhouse-gas vertical column through two ways of increasing a spectrum modeling speed and a continuous calculation speed for the concentration of vertical column. In the spectrum modeling process, the modeling speed is increased by ignoring the influence of some weak interference factors and discretized radiative transfer equations. In the continuous inversion process of concentration of the greenhouse-gas vertical column, a table look-up method is adopted to increase the calculation speed of the concentration of the greenhouse-gas vertical column: the solar normalized radial brightness and weighting function of all possible scenes are simulated firstly, correction factors are counted when all spectra adopt the value of simulation of one specific scene for inversion fitting, and a correction chart is made; during follow-up data processing, the measurement spectra of all scenes can adopt the value of simulation of the specific scene for fitting, only the fitting result multiplies by corresponding correction factors in the chart, and the measurement spectra of all scenes do not need modeling.

The invention discloses a dual-mode non-destructive testing method for dielectric functional gradient insulation. First, the boundaries of different density regions are obtained by segmenting the density changes at different positions inside the three-dimensional insulator, and then the geometric information is used as input, according to the measured boundary plate. The change of capacitance obtains training samples, which are trained after normalization; the functionally gradient insulator boundary capacitance matrix is ​​substituted into the trained model, and the value of the dielectric constant in the corresponding material partition inside the insulator is obtained to complete the non-destructive testing. The invention combines the advantages of high spatial resolution of ICT and direct dielectric inversion of ECT to realize the non-destructive testing requirements of dielectric functional gradient insulation with fixed shape and parameters, and solves the parameter inversion process while improving the inversion speed through machine learning algorithm nonlinear and ill-conditioned problems.

The invention discloses an aviation transient electromagnetic inversion method based on LSTM network. The LSTM network is established by dimension, the input data is the aviation transient electromagnetic time domain signal, the sensor height is spliced ​​into the feature layer to participate in the calculation, and the output data is the resistivity model parameters; 3) According to the error reduction curve, select the appropriate training set size and training period, Ensure the training effect and training efficiency, and complete the network training; 4) Judge the generalization ability of the network according to the fitting of the network inversion results and its forward response; 5) Use the LSTM network to complete the inversion of aviation transient electromagnetic data and obtain the The resistivity distribution structure of the medium. The inversion method is stable and accurate, and has fast calculation speed, which can improve the interpretation efficiency of aviation transient electromagnetic data, and then provide technical support for real-time imaging of aviation transient electromagnetic data.

The invention discloses a novel calculation method for the surface soil moisture content of CBERS-02B satellite image data. The novel calculation method constructs a new inversion CBERS-02B satelliteimage data soil moisture content calculation method according to the relationship between two data sources and spectral radiance related waveband ranges, ground object reflectivity, atmospheric transmittance and incident solar energy. Besides, a large number of inspection results show that the calculation method not only improves the accuracy of soil moisture content inversion of the CBERS-02B satellite image data and accelerates the inversion speed, but also solves the problem that an algorithm for inverting the surface soil moisture content of the image data is difficult to implement and thelike, and is suitable for estimating the soil moisture content of a large-area region.

The invention discloses a fast microwave imaging method. In each iteration of the DBIM algorithm, the inversion speed is improved by calling this efficient direct problem solver to calculate the direct problem. At the same time, in order to make the whole inversion algorithm perform iterative inversion stably, the truncated singular value decomposition (TSVD) technique is adopted, which avoids a large number of numerical tests to determine the regularization term. In order to further reduce the amount of calculation, in the cost function of the inversion algorithm, the incident field is used instead of the total field. Therefore, the invention can perform inversion quickly and stably, ensuring the image inversion speed and quality.

The invention relates to a method for accurately detecting the earth gravity field, in particular to a method which includes: based on the load error analysis theory, accurately building the distance error between satellites of a K wave band distance meter, the satellite orbit position error and the orbital velocity error of a global positioning system (GPS) receiver and an error model in which accumulative geoid accuracy is affected by the nonconservative force error coalition of a satellite-bone accelerometer, and further accurately and rapidly inversing the earth gravity field. The method is high in inversion accuracy of the earth gravity field, simple in satellite gravity inversion process, low in performance requirements of a computer and definite in physical meanings of a satellite observation equation, and effectively improves inversion speed on the premise of ensuring calculation accuracy. The satellite gravity inversion method based on the load error analysis theory is an effective method for calculating the earth gravity field which is high in accuracy and spatial resolution.

The invention discloses a fast two-dimensional density inversion method based on a physical property function. The method comprises the following steps of: S1, collecting gravity data by field actual measurement; S2, determining the physical property function order of an inversion area according to the gravity data or the existing data; S3, transversely subdividing the inversion area in the underground half space according to the distribution of the physical property function; S4, establishing a corresponding kernel function matrix; and S5, performing forward modeling on the parameter inversion result, and checking the inversion fitting degree. The fast two-dimensional density inversion method introduces the physical property function, reduces the memory occupation during operation and improves the operation speed; a construction mode of an underground half-space density model when a physical property function is introduced; the inversion speed is improved; and less memory space is occupied in the calculation process.

The invention discloses a formation anisotropic resistivity extraction method based on electromagnetic wave logging data while drilling, comprising the steps of: s1. performing environmental correction on the input electromagnetic wave logging curve while drilling; s2. acquiring block geological structure information; s3. Sliding window processing of logging data; s4. Establishing a single-interface inversion stratigraphic model; s5. Judging whether there is an inversion result of the previous sliding window, and if so, execute step s6; otherwise, execute step s8; s6. The inversion result of the previous sliding window is used as the initial value of the inversion of the current sliding window; s7. Use the nonlinear optimization algorithm to invert the electromagnetic logging data while drilling. If not, execute step s9; otherwise, execute step s8; s8. Adopt adaptive multi-initial value selection method, combined with nonlinear optimization algorithm to invert electromagnetic logging data while drilling; s9. Optimize and select multiple inversion results . The invention can provide accurate resistivity information for real-time geosteering and reservoir evaluation.

The invention discloses a real-time extraction method of formation interface based on azimuth electromagnetic wave logging data while drilling. The extraction method includes the steps: s1. Sliding windowing on the logging data of highly deviated wells / horizontal wells; s2. If there are shallow detection geological signals, then perform step s3, otherwise perform step s5; s3. Use shallow detection geological signals signal, perform single-interface inversion; s4. input the optimal solution of single-interface inversion into step s7 as a known constraint; s5. establish a double-interface inversion model; s6. judge whether there is adjacent interface information; Execute step s7, otherwise execute step s8; s7. select multiple initial values ​​for the remote formation interface position and far surrounding rock resistivity in the model described in s5, and then go to step s9; s8. Assign multiple initial values ​​to the inversion parameters; s9. Obtain the optimal solution for dual-interface inversion based on the deep detection geological signals; s10. Perform real-time imaging processing on the inversion results. The invention can provide fast and accurate interface information for real-time geosteering.

The invention relates to a gravity satellite formation orbital stability optimization design and an earth gravity field precision inversion method on the basis of disturbing inter-satellite distance principle, in particular to an orbital stability optimization design method for a four-satellite system (FSS). In order to guarantee the stability of the four-satellite system, the quantity of satellite orbits is optimally designed, orbital semi-major axes, orbital eccentricities, orbital inclinations and right ascensions of ascending nodes keep unchanged, the difference of arguments of perigees each pair of satellites and the difference of mean anomalies of the pair of satellites are 180 degrees respectively, an initial argument of perigee of each satellite is arranged at the equator, an initial mean anomaly of each satellite is arranged at a pole, and the ratio of the semi-major axis of each elliptical orbit of the four-satellite system to a semi-minor axis of the elliptical orbit of the four-satellite system is 2:1. The gravity satellite formation orbital stability optimization design and the earth gravity field precision inversion method have the advantages that an earth gravity field is precisely and quickly inverted on the basis of a disturbing inter-satellite distance process; and the orbits are high in stability owing to the method, the earth gravity field computation precision is effectively improved, the gravity field inversion speed is increased to a great extent, and requirements on the performance of a computer are low.