2341 results about "Wave shape" patented technology

The invention provides a defect positioning accurate identification method considering the change of the section of a pile body, and the method comprises the steps: carrying out the simulation comparison of a constructed sound wave signal feature vector and a pre-constructed sound wave feature vector of a defect-free pile foundation, calculating the feature difference degree of the two, setting a difference degree threshold value, and if the feature difference degree exceeds the threshold value, carrying out the recognition of the defect-free pile foundation. If yes, preliminarily determining that the pile body has defects; on the basis of the adjusted sound wave detection, sound wave signals are collected again, signal processing and feature extraction are conducted through a self-adaptive path tracking algorithm, and the three-dimensional coordinates of the defect in the pile body are calculated according to the time delay and the propagation speed of reflected waves by combining the established multi-reflection model and an inversion sound wave propagation path; and according to the calculated three-dimensional coordinates of the defect reflection points, the influence degree of the defects on the bearing capacity of the pile foundation is calculated by combining the defect size estimated through the amplitude of reflected waves, the defect type recognized through waveform features and the strength parameter of a pile body material through a finite element method.

The invention discloses a concrete internal defect quantitative detection method and system based on reflected waves, and belongs to the technical field of nondestructive testing. A reflected wave data matrix is obtained through multi-angle excitation and synchronous receiving; calculating energy characteristics of each channel, and constructing an energy response residual field; extracting waveform offset, spectrum jitter and phase change caused by defects by adopting a disturbance comparison algorithm to form a disturbance feature vector set; a defect-response function curved surface is further constructed, a defect topological structure is inversed based on gradient and curvature analysis, and defect geometric parameters are output; and finally, inputting the multi-moment defect parameters into the recurrent neural network, and predicting a defect evolution path and a failure risk. The method has high resolution and trend prediction capability, and is suitable for detection and early warning of concrete structures in bridges, tunnels and nuclear power projects.

The invention discloses an intelligent analysis system for ultrasonic phased array detection data of a welding seam, and relates to the technical field of nondestructive testing, and the intelligent analysis system is characterized in that a four-dimensional wave field tensor is constructed by collecting full-waveform ultrasonic data under multi-channel, multi-path and multi-angle conditions; extracting reflected signals which keep time coherence in a discontinuous path, and generating a high-dimensional coherence feature matrix; inputting the features into a self-supervised contrast learning model to obtain a defect semantic embedding vector; recognizing a suspected weak defect area based on the distribution density and the boundary change trend, and performing reverse beam focusing in combination with original data to obtain a defect three-dimensional positioning map; the shape bifurcation index and the boundary stability are calculated through topological analysis, and intelligent discrimination of artifacts and microcrack defects is achieved; the method does not need label data, has high automation and robustness, and is suitable for weak defect identification of complex structure welding seams.

The invention discloses a cable partial discharge detection method and system, and the method comprises the steps: installing sampling sensor groups on a detected cable at a preset interval, and each group at least comprises a current sensor and an ultrasonic sensor; synchronously sampling and preprocessing the high-frequency current signal and the ultrasonic signal; extracting current waveform characteristics and a first time stamp of arrival of the first pulse, and extracting sound wave pulses in the ultrasonic signals and a second time stamp of arrival of the sound wave pulses; and calculating a time difference, if the time difference is smaller than a preset threshold value and the waveform amplitude similarity is within a preset range, determining that no failure partial discharge exists, otherwise, positioning a failure position. According to the method, the acoustoelectric signals are acquired cooperatively through multiple sensors, interference is eliminated in combination with time difference and waveform similarity criteria, the detection reliability and the positioning precision are effectively improved, and the method is suitable for cable insulation state monitoring and fault early warning.

The invention provides a power load prediction method and device, and belongs to the technical field of power load prediction.The method comprises the steps that current waveform data are obtained, and fundamental wave and harmonic components in the current waveform data are extracted; carrying out waveform spatial form geometric analysis to obtain a real-time load characteristic sequence, and then carrying out segmentation processing; the current effective value sequence of each time window is converted into a time-frequency domain energy distribution vector, and then a three-level feature library is constructed; constructing a three-dimensional tensor model through equipment start-stop event identification, inputting the three-dimensional tensor model into a multi-target optimizer to evolve feature weights, and filtering abnormal samples to obtain a feature cluster; performing random masking processing on the time sequence data of the feature cluster to generate a mask sequence, inputting the mask sequence into an encoder to reconstruct masking data, comparing, learning and judging abnormal output correction data, and inputting the corrected data into a prediction network to generate a feedback signal flow; and analyzing the feedback signal flow to update the prediction network weight. Based on the method, the invention also provides power load prediction equipment. According to the invention, the precision of power load prediction is obviously improved.

A disk-pack turbine for use, for example, in systems and methods in at least one embodiment for separating fluids including liquids and gases into subcomponents by passing the fluid through a vortex chamber into an expansion chamber and then through at least a portion of a waveform pattern present between at least two rotors and / or disks. The rotors and / or disks having waveform patterns on at least one side. In at least one embodiment, the waveform patterns include a plurality of hyperbolic waveforms travelling around an axial center of the disk with each hyperbolic waveform having a varying diameter along the waveform.

The invention relates to an internal isolated wave waveform reconstruction method and device, equipment and a medium. The method comprises the following steps: constructing a data set by using internal isolated wave waveforms of different water depths; constructing an internal isolated wave waveform reconstruction model, and training the internal isolated wave waveform reconstruction model through the data set based on a pre-constructed loss function fusing physical information to obtain a trained internal isolated wave waveform reconstruction model; wherein when the internal isolated wave waveform reconstruction model is processed, firstly, a spatial relationship among different feature values in a data set is extracted, and a first feature is obtained; analyzing time correlation feature information of the first feature to obtain a second feature; and finally, extracting important information of the second feature, and outputting a reconstructed waveform. And performing inner isolated wave waveform reconstruction through the trained inner isolated wave waveform reconstruction model. According to the method, the defects of internal isolated wave waveform observation data can be overcome, the generalization ability of the model is high, and the reconstruction precision is high.

The invention relates to a traveling wave waveform classification method based on wavelet transform and QPSO-BILSTM, and the method comprises the steps: collecting a traveling wave signal, and carrying out the preprocessing of the traveling wave signal; extracting a peripheral time domain of each wave head; collecting a large amount of labeled traveling wave waveform data of different types as labels; extracting a wave head peripheral time domain as a training set, and inputting the training set into the BiLSTM model for training and classification; parameters of the BiLSTM model are optimized through QPSO, and the classification accuracy is improved; and processing a newly collected traveling wave signal, inputting the processed traveling wave signal into the optimized BiLSTM model, and outputting a classification result. The traveling wave front calibration efficiency and precision can be remarkably improved through traveling wave waveform classification; fault types can be quickly identified through classification, and an important basis is provided for subsequent fault processing; the BiLSTM model optimized by the quantum particle swarm optimization not only can better extract key features in the traveling wave waveform, but also can better adapt to complex nonlinear features and various change modes of the traveling wave, so that the accuracy of waveform classification is improved.

The invention relates to the technical field of satellite altimetry, in particular to a height measurement method based on a feature-enhanced multi-base integrated re-tracking model, which is characterized in that the feature-enhanced multi-base integrated re-tracking model FMERM is constructed to optimize the position of a reflection waveform re-tracking point, and the FMERM model executes the following processing processes: 1, extracting input variable features by using PCA, and extracting the input variable features by using PCA; combining with an original input variable to form an enhanced feature set; secondly, optimizing the hyper-parameters of the XGBoost model by using a grid search method, so that the hyper-parameters of the XGBoost model are optimal; thirdly, utilizing the trained model to accurately calculate a re-tracking point normalization power value of the reflected signal waveform; fourthly, the sea surface height is inverted through the geometrical relationship and direct reflection signal time delay and atmospheric correction; an effective means is provided for solving the problem that a reflection waveform re-tracking method is inaccurate, and powerful support is provided for high-precision satellite-borne GNSS-R sea surface height measurement.

The invention discloses a feed proportion abnormity identification method and system, and particularly relates to the field of feed proportion abnormity identification, and the method comprises the steps: extracting the disturbance variables of vibration amplitude, feeding speed and voiceprint waveform in multi-source disturbance data, constructing the disturbance time sequence expression of a frequency, form and amplitude triple feature structure, and carrying out the recognition of the vibration amplitude, the feeding speed and the voiceprint waveform. Extracting a disturbance characteristic track with center characterization capability in a cluster through atlas clustering, and writing the disturbance characteristic track into a track benchmark expression library; constructing an aligned trajectory expression set by analyzing the frequency domain structure and the time slice expression of each channel signal in the current batch of disturbance data and combining the binding relationship between the batch number and the formula number; by constructing an offset matching and track consistency judgment mechanism based on multi-source disturbance track expression, structured recognition and path logic verification of a formula drift phenomenon are realized, and the problem that the actual feed nutrition state is misjudged due to the fact that real proportion offset caused by residual feed cannot be recognized is solved.

The invention discloses a low-water-head water turbine with a flow guide wing bionic blade, which is provided with a variable pitch system for adjusting an included angle between a runner blade and an incoming flow direction, and a deflection control system for adjusting an included angle between a guide vane arranged on a bulb body and the incoming flow direction, the water flow impact acting force on the blades can be changed through the water flow impact acting force, the water flow impact acting force on the blades can be changed through the water flow impact acting force, and the water flow impact acting force on the blades can be changed through the water flow impact acting force. In addition, the blade of the runner of the hydraulic turbine set is the blade with the wave-shaped front edge, and a section protruding outwards is arranged in a pipeline for containing the bulb body and the runner, so that the water energy utilization rate of the hydraulic turbine set is further increased. The blade adopts a sine curve-shaped waveform front edge, and the structure simulates a convex shape on a fin-shaped limb of an underwater organism, so that the resistance in water is reduced, and the water energy utilization rate of a water turbine set is increased.

The invention discloses an ion chromatography wave crest boundary extraction method based on fitting residual and tangent judgment, and belongs to the technical field of ion chromatography analysis. Comprising the following steps: S1, reading an original waveform data sequence, and executing frequency reduction sampling according to a set step length; s2, calculating the slope of the data after underclocking sampling by using a difference method, and searching positive and negative slope sections by performing first-order derivation on the waveform after underclocking sampling; s3, noise interference is eliminated, slope areas are grouped, then a chromatographic peak detection limit value is set, and an effective peak left section and an effective peak right section are screened out; s4, combining the effective wave crest left side section and the effective wave crest right side section to obtain an effective wave crest and a preliminary left boundary point and a preliminary right boundary point of the effective wave crest; s5, fitting a straight line through the preliminary left boundary point and the preliminary right boundary point to position an error integral region; and S6, updating the optimal boundary point by traversing the evaluation indexes of all points in the error integral region. The boundary search range is more accurate, and the calculation efficiency is higher.

The invention relates to the technical field of radar waveform optimization, in particular to a synthetic broadband waveform optimization method based on reinforcement learning. According to the method, a deep reinforcement learning framework is adopted, and end-to-end optimization is carried out on the center frequency and amplitude parameters of multiple sub-pulses. The method comprises the following steps: constructing a synthetic broadband waveform signal model, and parameterizing sub-waveforms to determine a search space of an optimization problem; converting a waveform optimization problem into a reinforcement learning form, and defining a state space, an action space, a reward function and a state transition mechanism; based on a deterministic soft action value iteration strategy algorithm, point spread function features are extracted in combination with a Transform encoder, and sub-pulse parameters are dynamically adjusted. Through the method, the main lobe resolution can be effectively suppressed and improved, the target detection performance of a radar system is improved, and the method has good global optimization capability and environmental adaptability and is suitable for cognitive radar and adaptive waveform design scenes.

The invention provides a deep sea convergence area characteristic forecasting method and system based on a normal wave theory. The method comprises the following steps: determining a normal wave eigenvalue equation; solving a normal wave eigenequation, and calculating a normal wave eigenvalue of each order from the first order to a set normal wave order upper limit; calculating a normal wave eigenfunction value according to the sound source and the receiving depth; superposing and synthesizing a sound field by using the normal wave eigenvalue and the normal wave eigenfunction value; and calculating the sound propagation loss of the convergence area by using the synthesized sound field so as to obtain the energy and position of the deep sea convergence area. Compared with the prior art, the method has the advantages that the analytical solution form of the normal wave eigenequation is obtained in advance, compared with pure numerical calculation, certain calculation time can be saved, in addition, some function expansion and frequency interpolation technologies can be conveniently combined, a basis is provided for broadband sound propagation and signal waveform forecasting of the deep sea convergence area, and the method has a wide application prospect. The sound field calculation speed of the convergence area is greatly improved, and the method has good calculation precision.

The invention discloses a spraying control method of a waveform guardrail spraying manipulator, which relates to the technical field of industrial automation and robot control, and comprises the following steps: acquiring relative position change data of the manipulator and the surface of a waveform guardrail in a rotation process through a pre-established manipulator rotation track model; and for each time node, recording the spraying angle change and distance dynamic adjustment information, establishing a dynamic space mapping relationship under a manipulator rotation state for spraying control, and triggering a secondary spraying supplementing control instruction according to the coating quality consistency feedback, so that the reliability and stability of the spraying supplementing effect can be further ensured, and the spraying efficiency is improved. Therefore, the equipment utilization rate of the manipulator and the automation and intelligence level of spraying operation are remarkably improved while the uniformity of the coating and the consistency of spraying quality are improved.

The embodiment of the invention discloses a waveform identification and correction method and system for winding deformation diagnosis, and relates to the technical field of transformer winding monitoring, and the method comprises the steps: synchronously collecting voltage and current signals, and carrying out the preprocessing; performing waveform disturbance type identification on the preprocessed signal, and dividing the disturbance type into harmonic dominant disturbance, voltage sag disturbance or compound disturbance, and harmonic dominant disturbance, voltage sag disturbance or compound disturbance; dynamically selecting a corresponding waveform processing path according to a waveform disturbance type identification result; processing the signal by adopting the selected waveform processing path so as to correct and reconstruct the signal and obtain compensated voltage and current waveform data; by constructing a disturbance identification and path adaptive processing mechanism, the problems of processing lag and compensation distortion of a general algorithm during dynamic signal disturbance are solved, the identification precision is improved, and real-time error correction is realized.

The invention discloses a heat-proof skirt which is of a single-layer or multi-layer structure, each single-layer material is formed by compositing quartz fiber cloth and coating when multiple layers are arranged, the heat-proof skirt is made of 6-8 single-layer materials, and the skirt edge is in a wave shape. The coating is prepared from the following raw materials in parts by weight: 4 to 6 parts of vinyl polysiloxane, 15 to 20 parts of hydrogen-containing silicone oil, 15 to 20 parts of ammonium polyphosphate, 10 to 15 parts of aluminum hydroxide, 5 to 8 parts of silicon dioxide, 0.005 to 0.02 part of a platinum catalyst and 30 to 40 parts of a solvent. The heat-proof skirt has contractility, and the influence on normal swinging resistance of the engine is reduced to the maximum extent.

The invention discloses an offshore wind power typical fault transient waveform rapid simulation method and system, and relates to the technical field of offshore wind power unit grid connection and power grid fault simulation, and the method comprises the steps: obtaining an original typical fault voltage transient waveform at an offshore wind power unit grid connection point; a wavelet analysis method is adopted to carry out filtering processing on the fault voltage transient waveform, decomposition is carried out to obtain an approximate signal and a detail signal, a fundamental wave and a main oscillation waveform in the detail signal are extracted, an RLC second-order fault simulation circuit is established, waveform superposition is carried out to form a reconstructed waveform, a reconstructed waveform analytic expression is obtained, and fault transient characteristic parameters needing to be solved are determined; preliminarily fitting the initial values of the characteristic parameters, and iteratively solving the final values of the characteristic parameters by adopting a nonlinear least square method; fitting and solving equivalent line parameter values in combination with a fault feature expression; and substituting the waveform into a fault simulation circuit, and carrying out analog simulation to generate a reproduction waveform. According to the method, the typical fault transient waveform of offshore wind power can be rapidly and accurately reproduced.

The invention provides a metal mask plate and a manufacturing method thereof. The manufacturing method of the metal mask plate comprises the following steps: respectively determining a length deviation variable quantity and a wave shape variable quantity of a base material based on length deviation values and wave shape values of a plurality of partitions of the base material; creating an association model for representing an association relationship among the length deviation variable quantity, the wave shape variable quantity and the pixel hole variable quantity of the base material; the length deviation variable quantity and the wave shape variable quantity of the base material are substituted into the correlation model, and a predicted value of the pixel hole variable quantity of the base material is obtained; and under the condition that the predicted value of the pixel hole variation of the base material is not smaller than the target threshold value, the manufacturing process of the metal mask plate is adjusted. According to the method, the predicted value of the pixel hole variable quantity of the metal mask plate is predicted before the metal mask plate is manufactured, and the standard manufacturing process is adjusted according to the predicted value, so that the pixel hole variable quantity of the metal mask plate manufactured based on the adjusted manufacturing process meets the design requirement, and the product consistency can be improved.

The invention belongs to the technical field of solid-state batteries, and particularly relates to a battery cell structure of a solid-state battery, the solid-state battery, battery manufacturing equipment and a battery manufacturing process, and the battery cell structure comprises a positive plate, a solid-state electrolyte layer and a negative plate which are arranged in sequence; the solid electrolyte layer is of a waveform structure, micro-nano salient point arrays are arranged on the surfaces of the two sides of the solid electrolyte layer, and self-repairing dynamic interface layers are arranged on the surfaces of salient points; one side, facing the solid electrolyte layer, of the positive plate is provided with a first waveform surface, and the first waveform surface has a gradient pore structure; the negative plate is provided with a second wave-shaped surface, and the second wave-shaped surface is provided with a gradient pore structure; wherein the repairing dynamic interface layer can automatically repair microcracks generated by circulation, and mechanical anchoring of the salient point array is matched, so that the problem of impedance increase caused by interface cracks of a traditional solid-state battery is solved, and the cycle life is remarkably prolonged.

The invention relates to a method for displaying sixteen gray scales by an electronic paper price tag type display module, which comprises the following steps of: performing refreshing by utilizing a condition that at most four waveforms are controlled during refreshing each time in a new and old data comparison mode by utilizing a 1-bit data IC (Integrated Circuit) of electronic paper, identifying the first three gray scales and pure white gray scales in sixteen gray scale data during the first refreshing, and displaying the sixteen gray scales in the sixteen gray scale data; converting into data of corresponding waveforms, and brushing four corresponding gray scales through the debugged gray scale waveforms; in the second to fifth refreshing, three pieces of gray scale data are identified each time, and the debugged corresponding gray scale waveform is called to brush out three gray scales; the method is used for sixteen-gray-scale display through five times of superposition of different-gray-scale pictures.

A waveform analytical device 4 which analyzes a target waveform which is a chromatogram or an optical spectrum includes a waveform division unit 54 configured to divide the target waveform into a plurality of partial waveforms, a determination unit 55 configured to determine whether each of the plurality of partial waveforms of the target waveform is a peak portion using a learned model created by machine learning using a plurality of sets of a plurality of partial waveforms created by dividing a reference waveform having a peak portion whose position is known, and a classification unit 56 configured to classify the target waveform into a peak region where the peak portion continues and a non-peak region other than the peak region based on a determination result from the determination unit.

The invention discloses an intermediate frequency electro-therapeutic apparatus, which comprises a controller and a waveform synthesis circuit, and is characterized in that the waveform synthesis circuit comprises a first amplitude adjusting circuit, a second amplitude adjusting circuit and a synthesis adjusting circuit, and the controller generates a first PWM signal according to the overall amplitude information; the first amplitude adjusting circuit performs amplitude modulation processing on an input original waveform according to the first PWM signal to generate a first analog waveform signal; the controller obtains amplitude information corresponding to the frequency of the modulation waveform by using the phase accumulator, and generates a second PWM signal; a second amplitude adjusting circuit performs multiplication operation on the second PWM signal and the first analog waveform signal to obtain a second analog waveform signal and obtain an inverted waveform signal; the controller obtains amplitude information corresponding to the frequency of the fundamental frequency waveform by using a phase accumulator, and generates a third PWM signal; and the synthesis adjusting circuit performs multiplication operation on the third PWM signal and the second analog waveform signal or the inverted waveform signal to generate an intermediate frequency treatment waveform.

The invention discloses a method for testing the compactness and stress of an anchor rod. The method comprises the following steps: constructing an experimental model for testing the compactness and stress of the anchor rod; performing nondestructive testing on the experimental model by using an anchor rod nondestructive testing system, and calculating the compactness of the anchor rod according to a waveform time-distance graph transmitted by the obtained excitation waves in the anchor rod; carrying out multiple anchor rod drawing experiments and anchor rod nondestructive testing experiments on the experimental model to respectively obtain anchoring surrounding rock body parameters, anchoring quality parameters and anchor rod stress states; and the analysis anchoring force is calculated according to the anchoring surrounding rock parameters and the anchoring quality parameters. According to the method, a nondestructive testing result and a result obtained by an anchor rod drawing force experiment are mutually verified by constructing an experimental model for testing the compactness and the stress of the anchor rod, and a calculation rule of the anchor rod drawing force is summarized, so that mathematical calculation is performed, the working efficiency is improved, and reliable data are provided for tunnels, geotechnical engineering supports and the like.

This invention relates to a traveling wave waveform morphology classification method based on wavelet transform and QPSO-BILSTM, comprising: acquiring traveling wave signals; preprocessing the traveling wave signals; extracting the time domain of each wavefront periphery; collecting a large amount of labeled traveling wave waveform data of different types as labels; extracting the time domain of the wavefront periphery as a training set and inputting it into a BiLSTM model for training and classification; optimizing the parameters of the BiLSTM model through QPSO to improve the classification accuracy; processing newly acquired traveling wave signals and inputting them into the optimized BiLSTM model to output the classification results. The classification of traveling wave waveform morphology can significantly improve the efficiency and accuracy of traveling wave front calibration; classification can quickly identify fault types, providing an important basis for subsequent fault handling; the BiLSTM model optimized by the quantum particle swarm optimization algorithm can not only better extract key features in the traveling wave waveform, but also better adapt to the complex nonlinear characteristics and diverse variation patterns of the traveling wave, thereby improving the accuracy of waveform classification.

The invention provides a blasting vibration waveform feature extraction and damage range intelligent calibration system. The method comprises the following steps: calculating a waveform irregularity sequence based on original vibration waveform data; then, a fiber grating sensor network is used for collecting strain vibration waves on the surface of the structure, and spatial distribution information of the strain change rate of the structure is mapped into a feature enhancement channel; learning a feature distribution mode based on the waveform data processed by the feature enhancement channel by using a coding and decoding structure, and performing reconstruction operation on the current blasting vibration waveform according to the feature distribution mode; identifying and amplifying abnormal feature differences by comparing an actual waveform irregularity sequence with a reconstruction sequence, and generating reconstruction error peak distribution; and finally, when the strain gradient abrupt change point position detected by the optical fiber sensor network is overlapped with the reconstruction error peak point space, generating a blasting damage area thermodynamic diagram based on the overlapped point space density. The precision and reliability of blasting damage range calibration are improved.

The invention belongs to the technical field of data processing, and particularly discloses a displacement data acquisition control method and device, electronic equipment, a medium and a program product.The method comprises the steps that multi-channel displacement data acquisition is started based on a detection task, an FPGA module is triggered to record a reference position through sensing head original point positioning, and the reference position is obtained; the collected binomial square wave signals are conditioned to obtain waveform data; analyzing the waveform data in real time, identifying first preset waveform types such as a unilateral waveform, a special symmetric waveform and a sawtooth waveform and other second preset waveform types, and triggering a corresponding execution plan according to an identification result; and aiming at a second preset waveform type, adopting a preset synthesis algorithm to calculate the multichannel data, and generating characteristic parameters such as depth, thickness, roundness and taper of the detected object. According to the method, high-precision and high-robustness displacement data real-time acquisition and processing can be realized, and the reliability and the automation level of a detection result are improved.

The invention relates to the technical field of intelligent sensing systems, in particular to an algorithm and system for avoiding water surface floating objects and improving monitoring precision based on radar water level monitoring, and the method comprises the following steps: collecting radar echoes, constructing spatial-temporal characteristics based on waveform asymmetry and coherent attenuation gradient, and calculating the spatial-temporal characteristics; according to the method, the central moments of the front edge and the rear edge of the echo are calculated, the waveform asymmetry is quantified, the dynamic stability is analyzed in combination with the attenuation gradient of the multi-time-lag coherence coefficient, the floating object interference is recognized from the double dimensions of the spatial form and the time memorability, and the actual water level is inverted. An effective scattering trailing interval is positioned, nonlinear approximation operation is performed on trailing data in combination with an index physical model, a shielded water surface energy distribution curve is reconstructed, a derivative zero point is solved, distance measurement deviation caused by floating objects is corrected, the real water level height is effectively restored, and the monitoring anti-interference capability and the data accuracy in a complex hydrological environment are improved.

The invention discloses an intrusive rock alteration zone reservoir prediction method comprising the following steps: S1, analyzing and identifying sensitive logging parameters of an intrusive rock alteration zone, and establishing a dominant lithology identification combination chart; s2, establishing a model, and determining seismic response characteristics of intrusive rocks; s3, quantitative evaluation is carried out on the intrusive rock; s4, reservoir prediction of the intrusive rock alteration zone is carried out; and S5, quantitatively predicting the reservoir distribution and thickness of the intrusive rock alteration zone. According to the method, through different lithologic logging response characteristics, a multi-parameter intersection advantage lithologic chart is established, wave group characteristics of intrusive rocks are determined, and the distribution range and thickness of the intrusive rocks are identified by utilizing wave impedance inversion and optimizing attributes of a product of a reflection intensity alternating current component and a phase cosine; and waveform indication inversion of the invaded rock mud shale alteration zone is carried out by using natural potential as a characteristic curve, and reservoir distribution and thickness of the invaded rock alteration zone are predicted and are basically consistent with actual drilling thickness.

An object is to accurately detect peaks of various compositions, even in a case of unseparated peaks in which peaks of a plurality of compositions are superimposed. A computer acquires waveform data D1 having a peak P1 in a composition A measured by a data analysis device (S10). Next, the computer acquires waveform data D2 having a peak P2 in a composition B measured by the data analysis device (S20). Next, waveform data D12 including unseparated peaks by superimposing the waveform data D1 including the acquired peak P1 and the waveform data D2 including the acquired peak P2 (S30) is generated. Next, the generated waveform data D12 of the unseparated peaks is input as learning data, and the waveform data D1 and D2 corresponding to the waveform data D12 are input as training data in Step S40. Next, machine learning is performed using the waveform data D12, D1, and D2, and a learned model for estimating an accurate separation method of unseparated peaks is constructed based on the trained result (S50).