An evaluation optimization system for helium gas favorable exploration areas

Abstract

The application discloses an evaluation and optimization system for a favorable exploration area of helium gas, which comprises the following modules: a data acquisition and preprocessing module for generating a geological basic data set; a primary helium generation and discharge calculation module for calculating the primary helium generation amount and the primary helium internal pressure, and obtaining an effective helium discharge time period and a helium discharge intensity; a fault through source and leakage judgment module for restoring the ancient ground stress to judge the fault through source or the plugging state, and obtaining a through source time efficiency coefficient or a leakage loss coefficient; a caprock anti-leakage evaluation module for calculating capillary resistance and microfracture escape amount to obtain a caprock anti-leakage coefficient; a dynamic reservoir prediction module for inputting the above parameters into an improved fractional order grey prediction model embedded with a variable lower limit accumulation mechanism to obtain a helium gas dynamic reservoir probability value; and a reservoir map generation module for smoothing and interpolating the probability value and marking to obtain a spatial distribution polygon map of the favorable exploration area of helium gas. The application improves the dynamic reservoir prediction precision and the operation efficiency.

Description

Technical Field

[0001] This invention relates to the field of helium resource exploration and evaluation, and in particular to an evaluation and selection system for favorable helium exploration areas. Background Technology

[0002] Helium, as an extremely precious strategic and scarce resource, plays an irreplaceable role in aerospace, semiconductors, medicine, and cutting-edge scientific research. With the dramatic increase in global demand for high-purity helium, the traditional "helium-based exploration" model is no longer sufficient to meet the current needs for efficient exploration. Research on helium accumulation mechanisms and quantitative evaluation techniques for favorable areas are gradually becoming the core of exploration and development. Helium accumulation is a complex and dynamic geological process involving the coupling of multiple factors such as source, fault channels, and caprock sealing. Uranium and thorium decay in the basement rocks to generate helium, which is then released under the pressure of the generated helium, overcoming pore fluid pressure. Fault zones, under the influence of paleostress, can be open or closed, serving as both source channels and potential leakage pathways. The caprock seals helium through capillary resistance, but the development of microfractures can lead to escape. Therefore, accurately characterizing the temporal matching relationship between helium generation and expulsion history, fault activity history, and caprock sealing history is crucial for accurate prediction of favorable helium areas.

[0003] However, existing helium exploration and evaluation technologies often focus on static qualitative analysis of single geological elements, which has significant shortcomings. First, existing technologies typically fragment the spatiotemporal continuity of helium generation, expulsion, transport, and accumulation, lacking a quantitative calculation mechanism for dynamically comparing the internal pressure of generated helium with the pore fluid pressure of the basement over time, making it difficult to accurately define the effective helium expulsion period and its intensity. For determining fault sources and leakage, the assessment often relies solely on qualitative inferences based on current fault geometry, failing to reconstruct the paleostress state and fault gouge smearing from geological history, and failing to perform a precise timeline comparison between the fault opening period and the effective helium expulsion period, resulting in an inability to effectively distinguish between source availability and leakage risk. Furthermore, in caprock evaluation, existing technologies often neglect the influence of current geostress on the opening width of microfractures, failing to establish a quantitative game relationship between capillary resistance and microfracture escape volume, leading to an overly subjective assessment of the caprock's resistance to leakage.

[0004] Furthermore, in terms of predicting the probability of hydrocarbon accumulation, existing technologies mostly adopt traditional integer-order grey prediction models. These models rely on the equal weighting of all historical data for accumulation. They cannot identify the fluctuation characteristics of geological data sequences to adaptively adjust model parameters, and they lack boundary constraint mechanisms. This results in invalid historical data that is too far removed from the current time point interfering with the prediction results, leading to low prediction accuracy of the dynamic probability of helium accumulation and making it difficult to meet the actual needs of high-precision exploration and optimization in three-dimensional space.

[0005] Therefore, how to provide an evaluation and selection system for favorable helium exploration areas is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0006] One objective of this invention is to propose an evaluation and selection system for favorable helium exploration areas. This invention fully integrates key steps such as multi-source geological data preprocessing, quantitative determination of helium generation and expulsion pressures, dynamic assessment of fault sources and leakage, caprock anti-loss calculation, and three-dimensional spatial mapping. It constructs an intelligent evaluation process for helium accumulation with multi-element spatiotemporal coupling, dynamic threshold triggering, and spatial distribution visualization output. The core of this invention lies in introducing an improved fractional-order grey prediction model and embedding a variable lower limit accumulation mechanism with boundary constraints in the fractional-order accumulation module. By dynamically eliminating invalid historical data through a time window and using a gamma function to calculate adaptive fractional-order weights, it achieves accurate prediction of the dynamic probability of helium accumulation under complex geological evolution environments. This invention possesses advantages such as multi-dimensional coupling of data modeling, adaptive adjustment of order, dynamic constraints on historical boundaries, high prediction accuracy, and accurate spatial positioning of results. It can significantly improve the accuracy of identifying helium accumulation patterns under complex geological conditions and the reliability of exploration target area selection, thereby effectively solving problems such as fragmented evaluation of geological elements, redundant interference from historical data, and low accuracy of accumulation probability prediction in existing methods.

[0007] An evaluation and selection system for favorable helium exploration areas according to an embodiment of the present invention includes the following modules: The data acquisition and preprocessing module is used to simultaneously acquire multi-source data from the target area and preprocess it to obtain a basic geological dataset. The helium generation and helium expulsion calculation module is used to calculate the amount of helium generation and internal pressure of helium generation, as well as the pressure of basal pore fluid, based on the geological basic dataset. It compares the internal pressure of helium generation with the pressure of basal pore fluid period by period to obtain the effective helium expulsion time period and helium expulsion intensity value. The fracture source and leakage determination module is used to restore paleogeological stress results based on geological basic datasets, determine whether the fracture is in a state of open conduction or closed blockage at various periods, and obtain the source efficiency coefficient or the leakage coefficient. The caprock loss assessment module is used to calculate capillary resistance and microcrack opening width based on geological basic datasets, assess helium resistance and escape amount, and obtain the caprock loss resistance coefficient. The dynamic helium accumulation prediction module is used to input the effective helium venting time period, helium venting intensity value, source efficiency coefficient or leakage coefficient, and caprock anti-leakage coefficient into the improved fractional gray prediction model to obtain the dynamic helium accumulation probability value. The improved fractional gray prediction model includes an input module, a state triggering module, a fractional accumulation module, and a differential coupling module. The fractional accumulation module embeds a variable lower limit accumulation mechanism with boundary constraints. The hydrocarbon accumulation map generation module is used to draw polygon maps of the spatial distribution of favorable helium exploration areas based on the dynamic probability values ​​of helium accumulation.

[0008] Optionally, the data acquisition and preprocessing module specifically includes: Through geological exploration and experimental testing, multi-source data were simultaneously collected, including uranium-thorium abundance data of the basement rocks in the target area, data on the variation of basement porosity with depth, fracture geometry and mechanical parameters, test data on the distribution of micro- and nano-pore sizes in the caprock, current geostress data, data on the burial history evolution of the target area, and data on the regional paleotectonic evolution. The collected multi-source data is uniformly converted into a standard data format. The three-dimensional spatial coordinate reference system of the target area is read, and the spatial coordinates of the converted multi-source data are projected onto the three-dimensional spatial coordinate reference system. The projected multi-source data is aligned with the three-dimensional spatial grid. The multi-source data with the unified format and the three-dimensional spatial position alignment are integrated and packaged to output a geological basic dataset.

[0009] Optionally, the helium production and excretion calculation module specifically includes: Uranium and thorium abundance data of basement rocks were extracted from the geological dataset. The uranium and thorium content values ​​were read, and the time span values ​​for each geological period were retrieved from the target area's burial history evolution data. The uranium and thorium content values ​​were multiplied by the corresponding preset radioactive decay helium production rate constant and then multiplied by the time span value to obtain the uranium and thorium helium production amounts. The uranium and thorium helium production amounts were added together to calculate the helium production amount of the target area in each geological period. The preset radioactive decay helium production rate constant was set based on the physical mechanism of uranium-thorium isotope decay releasing alpha particles. The number of moles of helium atoms produced per unit mass of uranium or thorium per unit time was calculated to obtain the corresponding helium production rate constant. The uranium helium production rate constant ranged from 1.15 × 10⁻¹³ to 1.20 × 10⁻¹³ mol / (g) a) The helium production constant of thorium ranges from 3.10 × 10⁻¹⁴ to 3.15 × 10⁻¹⁴ mol / (g) a); Extract the data on the variation of basal porosity with depth and the burial history evolution data of the target area from the geological basic dataset. Read the burial depth values ​​for each geological period from the burial history evolution data of the target area. Find the porosity value corresponding to the same depth value from the data on the variation of basal porosity with depth. If no value is found, select the two values ​​with the closest depth and calculate the average value to obtain the corresponding basal porosity. The formation temperature values ​​for the corresponding geological period are read from the burial history evolution data of the target area. The helium generation amount is divided by the basement porosity value to obtain the helium generation concentration per unit pore. The helium generation concentration per unit pore is multiplied by the preset ideal gas constant and then by the formation temperature value to calculate the helium generation internal pressure for each geological period. The thickness and density of the overlying strata for each geological period are read from the burial history evolution data of the target area. The thickness of the overlying strata is multiplied by the density of the strata and then by a preset gravitational acceleration to obtain the pressure of the overlying strata. The density of the formation water is multiplied by a preset gravitational acceleration and then by the equivalent vertical height of the liquid column for the corresponding geological period, which is the thickness of the overlying strata or the paleoburial depth for the corresponding period, to obtain the hydrostatic pressure. Subtract the hydrostatic pressure value from the overlying rock pressure value and multiply it by the basement porosity value of the corresponding geological period. Add the hydrostatic pressure value to the product to calculate the basement pore fluid pressure for each geological period. The helium generation internal pressure and the basement pore fluid pressure of the same geological period are extracted to the same time point and compared period by period to determine whether the helium generation internal pressure of each geological period is greater than the basement pore fluid pressure. When the helium generation internal pressure of a certain geological period is greater than the basement pore fluid pressure, the corresponding geological period is recorded as the effective helium release period. The excess pressure is obtained by subtracting the base pore fluid pressure from the generated helium internal pressure during the effective helium purging period. The excess pressure is then multiplied by a preset helium purging efficiency coefficient to calculate the corresponding helium purging intensity.

[0010] Optionally, the fracture source and leakage determination module specifically includes: Extract fracture geometry and mechanical parameters, target area burial history evolution data and regional paleotectonic evolution data from the geological basic dataset. Read the stratigraphic thickness values ​​for each geological period from the target area burial history evolution data, and read the tectonic stress direction and magnitude values ​​for each geological period from the regional paleotectonic evolution data. The gravity stress value is obtained by multiplying the stratigraphic thickness by the preset rock density and the preset gravitational acceleration. This value is then added to the tectonic stress value. The sum is then projected and decomposed according to the tectonic stress direction to calculate the paleostress results for each geological period. The fracture strike, fracture dip angle and rock tensile strength values ​​are extracted from the fracture geometry and mechanical parameters. The paleostress results of each geological period are projected and calculated according to the fracture strike and fracture dip angle values ​​to obtain the normal stress and shear stress values ​​along the fracture surface. The friction coefficient value is obtained by dividing the shear stress value by the normal stress value. By comparing the friction coefficient value with the preset critical friction coefficient, when the friction coefficient value is greater than the preset critical friction coefficient, the stress state of the fracture zone is determined to be stress-induced fracture; when the friction coefficient value is less than or equal to the preset critical friction coefficient, the stress state of the fracture zone is determined to be stress-induced closure. Extract the fracture displacement and fault gouge thickness values ​​from the fracture geometry and mechanical parameters. Divide the fracture displacement value by the fault gouge thickness value to obtain the smearing ratio. Compare the smearing ratio with the preset critical smearing ratio. When the smearing ratio is greater than the preset critical smearing ratio, the fault gouge smearing is determined to be discontinuous. When the smearing ratio is less than or equal to the preset critical smearing ratio, the fault gouge smearing is determined to be continuous. Based on the combined stress state of the fault zone and the smearing of the fault gouge, when the stress state of the fault zone is stress-induced rupture and the smearing of the fault gouge is discontinuous, the fault is determined to be in a fractured and conductive state at each stage. When the stress state of the fault zone is stress-induced closure or the smearing of the fault gouge is continuous, the fault is determined to be in a closed and sealed state at each stage. Extract the effective helium release time period from the records, compare the period when the fracture was in the open and conductive state with the effective helium release time period on the time axis, and determine whether the two coincide at the time point. When the period when the fracture is in the open-conducting state coincides with the effective helium evacuation period, the time span of the period when the fracture is in the open-conducting state is divided by the time span of the effective helium evacuation period to calculate the source aging coefficient of the fracture under the corresponding effective helium evacuation period. When the period when the fracture is in the open-conducting state does not coincide with the effective helium evacuation period, the time span of the period when the fracture is in the open-conducting state does not coincide is divided by the time span of the effective helium evacuation period to calculate the loss and leakage coefficient of the fracture under the corresponding effective helium evacuation period.

[0011] Optionally, the caprock anti-loss assessment module specifically includes: Extract the micro-nano pore size distribution test data of the caprock from the geological basic dataset, read the maximum value of the pore radius from the micro-nano pore size distribution test data of the caprock, divide the preset helium interfacial tension value by the maximum value of the pore radius, and calculate the capillary resistance value. Current geostress data is extracted from the geological baseline dataset. The maximum and minimum horizontal principal stress values ​​are read from the current geostress data. The rock tensile strength and Young's modulus values ​​are read from the fracture geometry and mechanical parameters. Based on linear elastic fracture mechanics, the shear stress formed by the difference between the maximum and minimum horizontal principal stresses leads to rock tensile cracking when it exceeds the rock tensile strength. This horizontal stress difference is the equivalent driving force that drives the opening of microcracks. The stress difference value is obtained by subtracting the minimum horizontal principal stress value from the maximum horizontal principal stress value and subtracting twice the rock tensile strength value. The stress difference is multiplied by a preset microcrack height and a preset crack half-length. The product is then divided by the rock Young's modulus and a dimensional conversion factor to calculate the microcrack opening width. A dimensional conversion factor is introduced to unify the dimensions on both sides of the formula. The capillary resistance is multiplied by a preset resistance weighting factor to obtain the helium resistance. The microcrack opening width is multiplied by a preset escape weighting factor to obtain the helium escape index. Based on the cubic law of fluid dynamics, the crack escape flux is proportional to the opening width. This product is used to dynamically characterize the relative helium loss flux during the corresponding geological period. The preset microcrack height is set based on core micro-CT. The average longitudinal extension of fractures, statistically derived from scanning and thin-section observations, is assigned a value ranging from 0.2 to 0.8 m. The preset fracture half-length value is set based on the average fracture surface radius statistically derived from well logging fracture identification and field outcrop measurements, ranging from 0.5 to 1.5 m. The preset resistance weight coefficient is set based on the contribution ratio of capillary resistance to anti-loss capacity statistically derived from helium core displacement experiments, ranging from 0.6 to 0.8. The preset escape weight coefficient is set based on the contribution ratio of fracture escape volume to anti-loss capacity statistically derived from helium core displacement experiments, ranging from 0.2 to 0.4. The helium escape rate is multiplied by the preset regional geostress gradient constant to calculate the helium escape pressure. The helium resistance is subtracted from the helium escape pressure to obtain the anti-scattering difference. The anti-scattering difference is divided by the helium resistance to calculate the caprock anti-scattering coefficient.

[0012] Optionally, the dynamic hydrocarbon accumulation prediction module specifically includes: The input module extracts the effective helium release time period, helium release intensity value, source aging coefficient or leakage coefficient, and caprock anti-leakage coefficient to the same time node. The helium release intensity value, source aging coefficient or leakage coefficient, and caprock anti-leakage coefficient are concatenated into a one-dimensional array according to the chronological order of the effective helium release time period, and then normalized and input into the improved fractional gray prediction model. The state trigger module reads the values ​​of each element in the one-dimensional input vector in sequence. It subtracts the value of the previous element from the value of the next element in the one-dimensional input vector and takes the absolute value to obtain the absolute value of the adjacent difference. It adds up all the absolute values ​​of the adjacent difference and divides by the total number of the absolute values ​​of the adjacent difference to obtain the average rate of change value. It compares the average rate of change value with the preset fluctuation threshold. When the average rate of change value is greater than the preset fluctuation threshold, it outputs the first trigger signal. When the average rate of change value is less than or equal to the preset fluctuation threshold, it outputs the second trigger signal. The fractional accumulator module receives a first trigger signal or a second trigger signal. When the first trigger signal is received, the fractional accumulator order of the fractional accumulator module is set to a preset high-order value. When the second trigger signal is received, the fractional accumulator order of the fractional accumulator module is set to a preset low-order value. In the fractional accumulation module, a variable lower bound accumulation mechanism with boundary constraints is executed, a dynamic lower bound filtering operation is performed, and the fractional weight value corresponding to each valid historical element value is calculated using the gamma function, and the fractional accumulation sequence after boundary constraints is output. The differential coupling module reads the fractional-order cumulative sequence after boundary constraints, subtracts the previous fractional-order cumulative value from the next fractional-order cumulative value in the fractional-order cumulative sequence after boundary constraints to obtain the first-order difference sequence, and multiplies it by a preset development coefficient to obtain the development term value. The previous fractional-order cumulative value in the fractional-order cumulative sequence after boundary constraints is multiplied by a preset action coefficient to obtain the action term value. The development term value is added to the action term value to obtain the differential coupling prediction value. By adding the pre-set initial residual correction to the differential coupling prediction value, the dynamic helium accumulation probability value at the current time node is obtained, thus completing the prediction process of the improved fractional gray prediction model.

[0013] Optionally, the step of implementing a variable lower bound accumulation mechanism for boundary constraints in the fractional accumulation module, performing a dynamic lower bound filtering operation, calculating the fractional weight value corresponding to each valid historical element value using the gamma function, and outputting the fractional accumulation sequence after boundary constraints, specifically includes: In the fractional-order accumulation module, the variable lower limit accumulation mechanism of the boundary constraints is executed. The value of each element in the one-dimensional input vector is read in sequence as the current element value, the effective helium release time period corresponding to the current element value is read as the current time node, and all element values ​​before the current element value in the one-dimensional input vector are read as historical element values. The effective helium release period corresponding to the historical element value is read as the historical time node. The time span between the current time node and the historical time node is read from the burial history evolution data of the target area as the node time difference. The node time difference is subtracted from the preset time window constant to obtain the dynamic lower limit difference. When the difference between the dynamic lower limits is less than or equal to zero, the corresponding historical element value is removed. When the difference between the dynamic lower limits is greater than zero, the corresponding historical element value is retained as a valid historical element value, and the dynamic lower limit filtering operation is performed. Get the fractional order of the current element value in the fractional order accumulation module, set the position number of the current element value in the sequence to i, set the position number of the valid historical element value in the sequence to j, subtract j from i to get the time step value, subtract one from the fractional order accumulation to get the first parameter value, add the time step value to the first parameter value and subtract one to get the second parameter value. The gamma function is used to calculate the gamma function values ​​corresponding to the first and second parameter values ​​as the first gamma value and the second gamma value. The gamma function value corresponding to the fractional cumulative order is calculated as the third gamma value. The gamma function is multiplied from the beginning to the time step value, and the result of the multiplication is used as the time step factorial result. Multiply the first gamma value by the second gamma value to obtain the first product result, multiply the third gamma value by the time step factorial result to obtain the second product result, divide the first product result by the second product result to obtain the fractional weight value, and calculate the fractional weight value corresponding to each effective historical element value. Multiply each valid historical element value by its corresponding fractional weight value to obtain the historical weighted product result. Add all the historical weighted product results together and add the product result of the current element value multiplied by the fractional weight value corresponding to the current element value to obtain the fractional cumulative value of the current element. Arrange the fractional cumulative values ​​of all elements in the one-dimensional input vector in order to obtain the fractional cumulative sequence after boundary constraints.

[0014] Optionally, the hydrocarbon accumulation map generation module specifically includes: Import the helium dynamic accumulation probability values ​​of all grids into the three-dimensional spatial coordinate system, read the helium dynamic accumulation probability values ​​of adjacent grids, and perform three-dimensional spatial smoothing interpolation when the difference between the helium dynamic accumulation probability values ​​of adjacent grids is less than the preset smoothing difference threshold. The dynamic helium accumulation probability value after smooth replacement is compared with the preset probability threshold. When the dynamic helium accumulation probability value after smooth replacement is greater than the preset probability threshold, the corresponding grid is marked as a candidate accumulation grid. The outer edges of all adjacent candidate accumulation grids are connected end to end to form a closed polygon boundary. Read the dynamic helium accumulation probability values ​​of all candidate accumulation grids and compare them with each other. Find the grid with the largest value as the highest probability grid, and read the corresponding effective helium expulsion time period and mark it as the best accumulation time window. Read the leakage coefficients of the highest probability grid and its adjacent grids, mark the grids with leakage coefficients greater than the preset risk coefficient threshold as high-risk grids, connect the outer edges of all adjacent high-risk grids one end to the other to form a leakage risk zone, and overlay the label of the optimal accumulation time window with the polygon boundary of the leakage risk zone onto the polygon boundary of the candidate accumulation grid to generate and output a polygon map of the spatial distribution of the favorable helium exploration area.

[0015] The beneficial effects of this invention are: This invention addresses the challenges of fragmented geological element evaluation, chaotic spatiotemporal matching, and high quantitative characterization difficulty in helium exploration by constructing a multi-source data spatiotemporal alignment mechanism and deploying a helium generation and expulsion dynamics determination module. It employs a period-by-period comparison of helium generation internal pressure and basement pore fluid pressure to determine effective helium expulsion time periods. Combined with fault source and leakage determination and caprock anti-loss assessment, it performs multi-dimensional parameter quantitative calculations to generate spatiotemporally aligned hydrocarbon accumulation baseline data. Through structured encapsulation and time-series stitching of the hydrocarbon accumulation baseline data, a one-dimensional input vector is constructed. This one-dimensional input vector is then input into an improved fractional-order... The grey prediction model performs fluctuation feature identification and order adaptive setting through a state triggering module, and uses a variable lower limit accumulation mechanism with embedded boundary constraints to perform dynamic time window filtering and fractional-order weight calculation, outputting a fractional-order accumulation sequence after boundary constraints. In the hydrocarbon accumulation prediction stage, it combines a differential coupling module to perform first-order difference calculation and residual correction, outputting a dynamic helium accumulation probability value. Furthermore, the hydrocarbon accumulation map generation module performs smooth interpolation, threshold trapping, and risk zone marking on the three-dimensional spatial grid probability values ​​to complete the output of a spatial distribution map of favorable helium exploration areas. Ultimately, it achieves a closed-loop intelligent evaluation of multi-factor coupled analysis, dynamic probability prediction, and three-dimensional spatial positioning of the helium accumulation process, effectively improving the dynamic screening accuracy of hydrocarbon accumulation boundary constraints, the accuracy of dynamic hydrocarbon accumulation probability prediction, and the reliability of exploration target area selection. Attached Figure Description

[0016] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a structural diagram of an evaluation and selection system for favorable helium exploration areas proposed in this invention; Figure 2 This is a flowchart of the helium generation, emission, and leakage determination process proposed in this invention. Figure 3 This is a flowchart of the improved fractional gray prediction and hydrocarbon accumulation map generation based on the boundary constraint variable lower limit accumulation mechanism proposed in this invention. Detailed Implementation

[0017] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0018] refer to Figures 1-3 An evaluation and selection system for favorable helium exploration areas includes the following modules: The data acquisition and preprocessing module is used to simultaneously acquire data on uranium and thorium abundance in the basement rocks of the target area, data on the variation of basement porosity with depth, fracture geometry and mechanical parameters, test data on the distribution of micro and nanopores in the caprock, current geostress data, data on the burial history evolution of the target area, and data on the regional paleotectonic evolution, and preprocess the data to obtain a basic geological dataset. The helium generation and expulsion calculation module is used to extract uranium and thorium abundance data of the basement rocks from the geological basic dataset to calculate the amount of helium generation and internal pressure of helium generation in each period. Combined with the data on the change of basement porosity with depth and the evolution data of the burial history of the target area, the relationship between porosity and depth is mapped to the geological history period. The basement pore fluid pressure is calculated, and the internal pressure of helium generation and the basement pore fluid pressure are compared period by period to obtain the effective helium expulsion time period and the corresponding helium expulsion intensity value. The fault source and leakage determination module is used to extract the geometric and mechanical parameters of the fault from the geological basic dataset, combine the burial history evolution data of the target area and the regional paleotectonic evolution data to restore the paleostress results, calculate the stress state of the fault zone and the fault gouge smearing situation, determine whether the fault is in a cracked and conductive state or a closed and blocked state in each period, compare whether the period when the fault is in a cracked and conductive state is synchronized with the effective helium release time period, and obtain the source efficiency coefficient or leakage coefficient of each fault under different effective helium release time periods. The caprock loss assessment module is used to extract the caprock micro-nano pore size distribution test data from the geological basic dataset to calculate capillary resistance, extract the current geostress data from the geological basic dataset to calculate microcrack opening width, evaluate helium resistance and escape amount, and obtain the caprock loss resistance coefficient. The dynamic helium accumulation prediction module is used to input the effective helium venting time period, helium venting intensity value, source efficiency coefficient or leakage coefficient, and caprock anti-leakage coefficient into the improved fractional gray prediction model to obtain the dynamic helium accumulation probability value. The improved fractional gray prediction model includes an input module, a state triggering module, a fractional accumulation module, and a differential coupling module. The fractional accumulation module embeds a variable lower limit accumulation mechanism with boundary constraints. The hydrocarbon accumulation map generation module is used to perform three-dimensional spatial smoothing and interpolation of the helium dynamic accumulation probability values ​​of all grids, delineate the regions where the helium dynamic accumulation probability value is greater than the preset probability threshold as boundaries, and mark the effective helium expulsion time period and the region with high loss and leakage coefficient corresponding to the region with the highest helium dynamic accumulation probability value, so as to obtain a polygon map of the spatial distribution of favorable helium exploration areas marked with the best accumulation time window and leakage risk area.

[0019] This invention significantly improves the accuracy and spatiotemporal matching reliability of helium accumulation prediction. By simultaneously fusing multi-source geological data, including basement evolution, fracture mechanics, and caprock properties, it achieves dynamic reconstruction of multiple elements—source, discharge, reservoir, and conservation—within a unified time coordinate system. When processing long geological evolution time series, an improved fractional-order grey prediction model with an embedded boundary constraint variable lower limit accumulation mechanism is introduced. This model adaptively eliminates invalid historical data from the distant future, effectively overcoming data redundancy interference and computational lag problems in traditional static evaluation methods, and significantly improving model computational efficiency and the accuracy of accumulation probability fitting. By dynamically comparing the internal pressure of helium source with fluid pressure, and accurately and synchronously determining the opening and closing of fractures and the helium discharge period, it effectively solves the technical challenge of determining the source and leakage time of fractures in complex tectonic zones, significantly reducing the misjudgment and missed judgment rates. Finally, based on three-dimensional spatial smooth interpolation, polygon maps marked with the optimal reservoir formation time window and the leakage risk zone are generated, making the reservoir formation evaluation more refined and closer to reality. This provides efficient and intelligent decision support for the selection of exploration target areas and risk avoidance in helium-bearing basins, and greatly improves the success rate and economic benefits of helium resource exploration.

[0020] In this embodiment, the data acquisition and preprocessing module specifically includes: Through geological exploration and experimental testing, multi-source data were simultaneously collected, including uranium-thorium abundance data of the basement rocks in the target area, data on the variation of basement porosity with depth, fracture geometry and mechanical parameters, test data on the distribution of micro- and nano-pore sizes in the caprock, current geostress data, data on the burial history evolution of the target area, and data on the regional paleotectonic evolution. The collected multi-source data are processed by format conversion and uniformly converted into a standard data format. The three-dimensional spatial coordinate reference system of the target area is read, and the spatial coordinates of the converted multi-source data are projected onto the three-dimensional spatial coordinate reference system to eliminate spatial position deviations between different data sources. The projected multi-source data are aligned according to the three-dimensional spatial grid so that the multi-source data within the same spatial grid correspond to each other. The format-unified and three-dimensional spatially aligned multi-source data are integrated and encapsulated to output a geological basic dataset.

[0021] In this embodiment, the helium production and expulsion calculation module specifically includes: Extract uranium and thorium abundance data of basement rocks from the geological basic dataset, read the uranium and thorium content values, read the time span values ​​of each geological period from the burial history evolution data of the target area, multiply the uranium and thorium content values ​​by the corresponding preset radioactive decay helium production rate constant and multiply by the time span value to obtain the uranium and thorium helium production amount, add the uranium helium production amount and thorium helium production amount to calculate the helium production amount of the target area in each geological period; Extract the data on the variation of basal porosity with depth and the burial history evolution data of the target area from the geological basic dataset. Read the burial depth values ​​for each geological period from the burial history evolution data of the target area. Find the porosity value corresponding to the same depth value from the data on the variation of basal porosity with depth. If no value is found, select the two values ​​with the closest depth and calculate the average value to obtain the corresponding basal porosity. The formation temperature values ​​for the corresponding geological periods are retrieved from the burial history data of the target area. The helium generation amount is divided by the basement porosity value to obtain the helium generation concentration per unit pore. The derivation logic of the helium generation concentration per unit pore is as follows: the basement porosity value represents the proportion of pore volume in a unit total volume of rock. Since the generated helium is free and stored in the pore space of the rock, dividing the helium generation amount representing the molar amount of helium generated per unit volume by the basement porosity value representing the proportion of pore volume, that is, the total helium generation amount is evenly distributed in the actual pore space, thereby deriving the molar amount of helium generated per unit pore volume, i.e., the helium generation concentration per unit pore. The helium generation concentration per unit pore is multiplied by a preset ideal gas constant and then by the formation temperature value to calculate the helium generation internal pressure for each geological period. The preset ideal gas constant is set based on the physical laws of the ideal gas law in thermodynamics, and the constant value published by the International Scientific Committee is directly adopted, which is 8.314 J / (mol·K). The calculation formula is as follows: ; in, For the pore fluid pressure of the substrate, For the pressure of the overlying strata, For hydrostatic pressure, This represents the basement porosity corresponding to different geological periods.

[0022] The thickness and density of the overlying strata at various geological periods are retrieved from the burial history data of the target area. The overlying strata thickness is multiplied by the strata density and then by a preset gravitational acceleration to obtain the overlying strata pressure. The formation water density is multiplied by a preset gravitational acceleration and then by the equivalent vertical height of the liquid column at the corresponding geological period, i.e., the overlying strata thickness or paleoburial depth at that period, to obtain the hydrostatic pressure. The derivation logic of the formula is as follows: During the evolution of the sedimentary basin, the overlying strata at a certain geological period of the target layer... The thickness is numerically equal to the paleoburial depth of the target layer in that historical period. Since underground water is usually connected to the surface or upper strata, the paleoburial depth of the target layer represents the equivalent vertical height of the hydrostatic pressure column acting on the target layer in that period. Therefore, using this equivalent vertical height of the liquid column to calculate the hydrostatic pressure conforms to the conventional engineering calculation standards in the field of basin simulation. The preset gravitational acceleration is based on the physical calculation of Newton's law of universal gravitation at a normal height on the Earth's surface. The standard gravitational acceleration constant value is directly taken, which is 9.8 m / s². The effective stress value is obtained by subtracting the hydrostatic pressure value from the overlying strata pressure value. The effective stress value is then multiplied by the basement porosity value of the corresponding geological period. The product is then added to the hydrostatic pressure value to calculate the basement pore fluid pressure for each geological period. The derivation of this formula is based on the Terzaghi effective stress principle. The basement porosity value is based on the physical mechanism in Biot's effective stress law that the proportion of pore pressure to effective stress is directly related to porosity. That is, the actual pressure increment borne by the fluid is equivalently distributed by the effective stress of the overlying strata after deducting the hydrostatic pressure according to the dynamic porosity value of the current geological period. The dynamic basement porosity value is used as the distribution coefficient for converting effective stress into pore fluid pressure. The product after distribution is added to the background hydrostatic pressure value to calculate the basement pore fluid pressure for each geological period. The helium generation internal pressure and the basement pore fluid pressure of the same geological period are extracted to the same time point and compared period by period to determine whether the helium generation internal pressure of each geological period is greater than the basement pore fluid pressure. When the helium generation internal pressure of a certain geological period is greater than the basement pore fluid pressure, the corresponding geological period is recorded as the effective helium release period. The excess pressure is obtained by subtracting the basement pore fluid pressure from the generated helium internal pressure during the effective helium purging period. The excess pressure is then multiplied by a preset helium purging efficiency coefficient. This preset helium purging efficiency coefficient is set based on the pore connectivity coefficient obtained from the pore structure and connectivity test of the basement rock in the target area, and is assigned a value in combination with the empirical proportional relationship of rock hydrocarbon and liquid purging efficiency measured in the laboratory. The value ranges from 0.6 to 0.8, and the corresponding helium purging intensity is calculated.

[0023] The helium generation and expulsion calculation module constructs a dynamic equilibrium model that mutually constrains the hydrocarbon generation pressurization and the formation confining pressure. During the calculation, the basement porosity is not a fixed value but a dynamic variable that evolves with burial history, directly determining the helium generation concentration per unit volume, thus calculating the real-time changing internal pressure of helium generation. Simultaneously, the system couples the overlying strata pressure with hydrostatic pressure to calculate the basement pore fluid pressure and compares it with the real-time difference of the internal pressure of helium generation over time. This coupling effect realistically simulates the physical process in geological history where fluid pressure accumulates with increasing hydrocarbon generation and dissipates with increasing confining pressure. This ensures that the acquisition of the effective helium expulsion time period is no longer a simple comparison of single pressure values, but a precise capture of the dynamic evolution of formation energy, thereby guaranteeing the scientific rigor and reliability of the helium expulsion kinetic analysis.

[0024] In this embodiment, the fracture source and leakage determination module specifically includes: Extract fracture geometry and mechanical parameters, target area burial history evolution data and regional paleotectonic evolution data from the geological basic dataset. Read the stratigraphic thickness values ​​for each geological period from the target area burial history evolution data, and read the tectonic stress direction and magnitude values ​​for each geological period from the regional paleotectonic evolution data. The gravity stress value is obtained by multiplying the stratigraphic thickness by a preset rock density and then by a preset gravitational acceleration. The derivation logic of the formula is as follows: In conventional engineering calculations for paleogeographic stress restoration in basins, a method of superimposing and reconstructing segmented uniformly distributed loads is usually used to handle the variation of rock density with depth. That is, the overlying strata are divided into multiple lithological segments according to the stratigraphic sedimentary sequence and lithological combination. An equivalent average rock density based on well logging and core testing calibration is assigned to each lithological segment. The thickness of each lithological segment is multiplied by its corresponding equivalent average rock density and gravitational acceleration, and then summed to obtain the gravity stress value of the target layer in the corresponding geological history period. This segmented equivalent density accumulation calculation method effectively conforms to the objective laws of the vertical heterogeneity of rock density and its variation with depth, ensuring the accuracy and reliability of the calculation results. The preset rock density is set based on the following: density logging data and core physical property test results measured in the target area, compaction correction calibration is performed in combination with the burial depth corresponding to the geological history period in the burial history evolution data of the target area, and the equivalent average density value is assigned to each lithological segment. The gravity stress value and the tectonic stress value are added together, and the added result is projected and decomposed according to the tectonic stress direction value to calculate the paleostress results of each geological history period. The fracture strike, fracture dip angle and rock tensile strength values ​​are extracted from the fracture geometry and mechanical parameters. The paleostress results of each geological period are projected and calculated according to the fracture strike and fracture dip angle values ​​to obtain the normal stress and shear stress values ​​along the fracture surface. The friction coefficient value is obtained by dividing the shear stress value by the normal stress value. The friction coefficient is compared with a preset critical friction coefficient. When the friction coefficient is greater than the preset critical friction coefficient, the fracture zone is determined to be in a state of stress-induced fracture. When the friction coefficient is less than or equal to the preset critical friction coefficient, the fracture zone is determined to be in a state of stress-induced closure. The preset critical friction coefficient is set based on Byerlee's friction law, combined with friction strength data obtained from triaxial friction tests of rocks in the target area. The ratio of critical shear stress to normal stress at which rock fractures due to frictional slip is calculated, and its value ranges from 0.5 to 0.7. The accuracy and reliability of the judgment logic are ensured by the following mechanism: based on the Coulomb-Mohr friction slip criterion, the friction coefficient is the ratio of shear stress to normal stress, which reflects the mechanical critical state of relative slip fracture of the fracture zone, and is related to the shear stress or normal stress. The absolute numerical values ​​are irrelevant; that is, regardless of whether the shear stress and normal stress values ​​are both large or small, as long as their ratio crosses the critical friction coefficient threshold, the fracture zone will undergo mechanical rupture. Simultaneously, to address the extreme critical ambiguity where values ​​in the geological data sequence might exactly equal the critical value due to rounding errors, the system introduces a mechanical safety margin determination mechanism. This expands the original single critical friction coefficient into a determination zone including the upper and lower limits of the critical interval. When the friction coefficient value falls within the critical interval, the system automatically marks it as a 'critical undetermined state' and performs a secondary verification based on the fault gouge smearing in the fracture geometry and mechanical parameters. If the smearing is discontinuous, it is confirmed as a stress-induced fracture; if the smearing is continuous, it is confirmed as a stress-induced closure. This effectively eliminates the ambiguity in the judgment under a single ratio critical state, ensuring the accuracy and reliability of subsequent fracture conduction state determination. The fracture displacement and fault gouge thickness values ​​are extracted from the fracture geometry and mechanical parameters. The fracture displacement value is divided by the fault gouge thickness value to obtain the smearing ratio. The smearing ratio is compared with a preset critical smearing ratio. When the smearing ratio is greater than the preset critical smearing ratio, the fault gouge smearing is determined to be discontinuous. When the smearing ratio is less than or equal to the preset critical smearing ratio, the fault gouge smearing is determined to be continuous. The preset critical smearing ratio is set based on the mudstone smearing potential evaluation theory of fault gouge smearing. It is set by measuring the geometric ratio between the critical thickness of the plastic flow coating of mudstone in the fracture zone and the fracture displacement. Its value range is between 3.0 and 5.0. Based on the combined stress state of the fault zone and the smearing of the fault gouge, when the stress state of the fault zone is stress-induced rupture and the smearing of the fault gouge is discontinuous, the fault is determined to be in a fractured and conductive state at each stage. When the stress state of the fault zone is stress-induced closure or the smearing of the fault gouge is continuous, the fault is determined to be in a closed and sealed state at each stage. Extract the effective helium release time period from the records, compare the period when the fracture was in the open and conductive state with the effective helium release time period on the time axis, and determine whether the two coincide at the time point. When the period when the fracture is in the open-conducting state coincides with the effective helium evacuation period, the time span of the period when the fracture is in the open-conducting state is divided by the time span of the effective helium evacuation period to calculate the source aging coefficient of the fracture under the corresponding effective helium evacuation period. When the period when the fracture is in the open-conducting state does not coincide with the effective helium evacuation period, the time span of the period when the fracture is in the open-conducting state does not coincide is divided by the time span of the effective helium evacuation period to calculate the loss and leakage coefficient of the fracture under the corresponding effective helium evacuation period.

[0025] The fracture source and leakage determination module abandons independent static evaluation and directly uses the "effective helium release time period" output by the helium generation and release calculation module as the sole criterion for comparison along a timeline accurate to geological history. This mechanism generates a source timeliness coefficient by calculating the overlap ratio between the time span of the fracture in the fractured and conductive state and the effective helium release time period. Effective migration is only confirmed when the fracture opening window and the peak of helium generation and pressurization are precisely synchronized in time. This coupling mechanism effectively eliminates invalid noise generated by fracture activity before or after the hydrocarbon accumulation period, ensuring that only fractures with time matching are considered hydrocarbon accumulation channels. This solves the problem of interference in determining the hydrocarbon accumulation period in complex tectonic areas and significantly improves the accuracy of identifying key moments of hydrocarbon accumulation.

[0026] In this embodiment, the caprock anti-loss assessment module specifically includes: The micro- and nano-pore size distribution test data of the caprock were extracted from the geological basic dataset. The maximum pore radius value was read from the caprock micro- and nano-pore size distribution test data. The capillary resistance value was calculated by dividing the preset helium interfacial tension value by the maximum pore radius value. The derivation logic of the formula is as follows: based on the Young-Laplace equation in capillary physics, the critical pressure difference for non-wetting phase gas to overcome capillary resistance and enter the pore is directly proportional to the gas-liquid interfacial tension and inversely proportional to the pore radius. The pore cross section is approximated as a circle, and the capillary resistance value is directly calculated by dividing the interfacial tension by the maximum pore radius value. The preset helium interfacial tension value is set based on the helium surface tension constant under standard conditions obtained by thermodynamic experiments of the gas-water two-phase interface. The measured formation temperature value of the target area is substituted into the linear empirical formula of interfacial tension changing with temperature to calculate the interfacial tension value under the corresponding temperature conditions and assign it a value. Its value range is between 0.030 and 0.050 N / m. Extract the current geostress data from the geological basic dataset, read the maximum and minimum horizontal principal stress values ​​from the current geostress data, read the rock tensile strength and rock Young's modulus values ​​from the fracture geometry and mechanical parameters, and concatenate the helium expulsion intensity values, source aging coefficient or leakage coefficient, and caprock anti-leakage coefficient into a one-dimensional array according to the chronological order of the effective helium expulsion time period. The state trigger module reads the values ​​of each element in the one-dimensional input vector in sequence, subtracts the value of the previous element from the value of the next element in the one-dimensional input vector and takes the absolute value to obtain the absolute value of the adjacent difference, adds up all the absolute values ​​of the adjacent difference and divides by the total number of the absolute values ​​of the adjacent difference to obtain the average rate of change value, and compares the average rate of change value with the preset fluctuation threshold to obtain the stress difference value. The stress difference is multiplied by a preset microfracture height and a preset fracture half-length. The product is then divided by the Young's modulus of the rock to calculate the microfracture opening width. The capillary resistance is multiplied by a preset resistance weighting coefficient to obtain the helium resistance. The microfracture opening width is multiplied by a preset escape weighting coefficient to obtain the helium escape rate. The preset microfracture height is 0.5 m, and the preset fracture half-length is set based on fracture identification in the target area through well logging. The results were compared with field outcrop measurements. The mean radius of the micro-fracture surface in the caprock was obtained and assigned a value, ranging from 0.5 to 1.5 m. The resistance weight coefficient and escape weight coefficient were set based on the physical game relationship between helium sealing and loss in the caprock. Through helium core displacement and capillary breakthrough pressure experiments, the contribution ratio of capillary resistance and fracture escape to the overall anti-loss capability was statistically analyzed and assigned a value. The resistance weight coefficient ranged from 0.6 to 0.8, and the escape weight coefficient ranged from 0.2 to 0.4. The helium escape pressure is calculated by multiplying the helium escape amount by a preset regional geostress gradient constant. The helium escape pressure is obtained by subtracting the helium escape pressure from the helium resistance value. The anti-dissipation difference value is calculated by dividing the anti-dissipation difference value by the helium resistance value. The preset regional geostress gradient constant is set based on the current geostress test data of the target area, and the growth rate of the maximum horizontal principal stress with depth is calculated and assigned a value ranging from 0.01 to 0.03 MPa / m.

[0027] In this embodiment, the dynamic hydrocarbon accumulation prediction module specifically includes: The input module extracts the effective helium release time period, helium release intensity value, source aging coefficient or leakage coefficient, and caprock anti-leakage coefficient to the same time node. The helium release intensity value, source aging coefficient or leakage coefficient, and caprock anti-leakage coefficient are concatenated into a one-dimensional array according to the chronological order of the effective helium release time period, and then normalized and input into the improved fractional gray prediction model. The state trigger module reads elements in a one-dimensional array sequentially, subtracts the value of the previous element from the value of the next element, and takes the absolute value to obtain the absolute value of the adjacent difference. The physical meaning of the absolute value of the adjacent difference is the dynamic fluctuation amplitude of the helium accumulation parameters in adjacent geological periods. All the absolute values ​​of the adjacent differences are added together and divided by the total number to obtain the average rate of change. The average rate of change is compared with a preset fluctuation threshold. The preset fluctuation threshold is set based on statistical variance analysis of historical accumulation evolution data, and the average standard deviation limit of the normal fluctuation of the data sequence is calculated and set. The value range is 0.3 to 0.7. When the average rate of change is greater than the preset fluctuation threshold, a first trigger signal is output. When the average rate of change is less than or equal to the preset fluctuation threshold, a second trigger signal is output. The fractional-order accumulator receives a first trigger signal or a second trigger signal. When the first trigger signal is received, the fractional-order accumulator is set to a preset higher-order value. When the second trigger signal is received, the fractional-order accumulator is set to a preset lower-order value. The preset higher-order and lower-order values ​​are set based on the following: In fractional calculus theory, the accumulator order is adaptively allocated according to the intensity of historical data fluctuations. Higher-order values ​​are suitable for strongly fluctuating sequences to enhance historical dependence, while lower-order values ​​are suitable for stationary sequences to highlight current data. The range of higher-order values ​​is 0.7 to 0.9, and the range of lower-order values ​​is 0.2 to 0.4. In the fractional accumulation module, a variable lower bound accumulation mechanism with boundary constraints is executed, a dynamic lower bound filtering operation is performed, and the fractional weight value corresponding to each valid historical element value is calculated using the gamma function, and the fractional accumulation sequence after boundary constraints is output. The differential coupling module reads the fractional-order cumulative sequence after boundary constraints. The first-order difference sequence is obtained by subtracting the previous fractional-order cumulative value from the next fractional-order cumulative value in the boundary-constrained fractional-order cumulative sequence. The first-order difference sequence is multiplied by a preset development coefficient to obtain the development term value. The first-order cumulative value in the boundary-constrained fractional-order cumulative sequence is multiplied by a preset action coefficient to obtain the action term value. The development term value is added to the action term value to obtain the differential coupling prediction value. The preset development coefficient and action coefficient are set based on the parameter identification method of differential equations in grey prediction theory. The least squares method is used to fit and solve the historical input vector and the fractional-order cumulative sequence. The development coefficient ranges from -0.5 to -0.3, and the action coefficient ranges from 0.5 to 0.7. The differential coupling prediction value is added to the preset initial residual correction value to obtain the dynamic helium accumulation probability value at the current time node, thus completing the prediction process of the improved fractional gray prediction model. The preset initial residual correction value is set based on the average initial deviation of the system calculated by comparing the model prediction value with the historical actual geological evaluation results, and the value range is 0.05 to 0.15.

[0028] In this embodiment, a variable lower bound accumulation mechanism with boundary constraints is executed in the fractional accumulation module. A dynamic lower bound filtering operation is performed, and the fractional weight value corresponding to each valid historical element value is calculated using the gamma function. The fractional accumulation sequence after boundary constraints is then output, specifically including: In the fractional-order accumulation module, the variable lower limit accumulation mechanism of the boundary constraints is executed. The value of each element in the one-dimensional input vector is read in sequence as the current element value, the effective helium release time period corresponding to the current element value is read as the current time node, and all element values ​​before the current element value in the one-dimensional input vector are read as historical element values. The effective helium release period corresponding to the historical element value is read as the historical time node. The time span between the current time node and the historical time node is read from the burial history evolution data of the target area as the node time difference. The node time difference is subtracted from the preset time window constant to obtain the dynamic lower limit difference. The preset time window constant is set based on the time resolution scale of the geological history evolution of the target area, combined with the smallest effective time unit of the hydrocarbon accumulation period division, and the value range is 1 to 3 million years. Determine if the dynamic lower limit difference is greater than zero. If the dynamic lower limit difference is less than or equal to zero, it means that the historical element value is too far away from the current time node, and the corresponding historical element value is removed. If the dynamic lower limit difference is greater than zero, it means that the historical element value is within the valid time window, and the corresponding historical element value is retained as a valid historical element value. Perform the dynamic lower limit filtering operation. Get the fractional order of the current element value in the fractional order accumulation module, set the position number of the current element value in the sequence to i, set the position number of the valid historical element value in the sequence to j, subtract j from i to get the time step value, subtract one from the fractional order accumulation to get the first parameter value, add the time step value to the first parameter value and subtract one to get the second parameter value. The gamma function value corresponding to the first parameter value is used as the first gamma value, the gamma function value corresponding to the second parameter value is used as the second gamma value, and the gamma function value corresponding to the fractional cumulative order is used as the third gamma value. The gamma function is multiplied from the beginning to the time step value, and the result of the multiplication is used as the time step factorial result. Multiplying the first gamma value by the second gamma value yields the first product. Multiplying the third gamma value by the time step factorial yields the second product. Dividing the first product by the second product yields the fractional weight value. The fractional weight value corresponding to each valid historical element value is then calculated. The formula for calculating the fractional weight value is as follows: ; in, These are fractional weight values. This is the cumulative order of the fractional order, which is the value of the first parameter + 1. The position number of the current element. This is the position index of the valid historical element value. This is a gamma function.

[0029] Multiply each valid historical element value by its corresponding fractional weight value to obtain the historical weighted product result. Add all the historical weighted product results together and add the product result of the current element value multiplied by the fractional weight value corresponding to the current element value to obtain the fractional cumulative value of the current element. Arrange the fractional cumulative values ​​of all elements in the one-dimensional input vector in order to obtain the fractional cumulative sequence after boundary constraints.

[0030] This invention achieves dynamic optimization and accurate prediction of hydrocarbon accumulation probability for multi-source parameters in geological time series by introducing an improved fractional-order grey prediction model combined with a variable lower limit accumulation mechanism for boundary constraints. Helium expulsion intensity, source availability, and caprock loss resistance coefficient are concatenated into an input vector in chronological order. A state-triggered mechanism adaptively matches the fractional-order accumulation order to construct a dynamically optimized structure. During accumulation, a dynamic time window is introduced to automatically remove invalid historical data that is too far removed from the current node, and a gamma function is used to accurately calculate the fractional-order weights of valid historical elements. This invention maintains prediction continuity and matching accuracy under complex geological conditions such as multi-phase tectonic evolution, drastic data fluctuations, or redundant interference. The hydrocarbon accumulation probability sequence, after differential coupling and residual correction, significantly suppresses noise and anomalous drift in long-term data, greatly improving the timeliness and reliability of dynamic helium accumulation prediction. Simultaneously, helium expulsion intensity, source availability coefficient, and caprock loss resistance coefficient are constructed as a unified time series vector input to the improved fractional-order grey prediction model. Through state-triggered and differential coupling mechanisms, the model can capture the nonlinear correlation between various elements, overcome the one-sidedness of single-element evaluation, and achieve accurate prediction of hydrocarbon accumulation probability under the synergistic effect of multiple elements.

[0031] The improved fractional grey prediction model of this invention is similar to the original model in that both retain the core architecture of grey prediction, namely, reducing randomness by accumulating the input sequence, extracting the development term and action term using differential coupling equations, and combining residual correction to output the final predicted value. Both rely on the gamma function to calculate the fractional weights of historical elements for sequence accumulation.

[0032] The difference lies in that this invention breaks the limitation of the original model, which uses a single fixed order for the accumulation calculation of all input sequences, and introduces a dual dynamic control mechanism of state triggering and boundary constraints. While the original model directly sets the global order, this invention adds a state triggering module. This module obtains the average rate of change by calculating the absolute value of the difference between adjacent elements and compares it with a preset threshold. When fluctuations are severe, a higher-order trigger signal is output; when fluctuations are stable, a lower-order signal is output, thereby achieving dynamic adaptive allocation of the accumulation order. Furthermore, in the core accumulation stage, the original model indiscriminately utilizes all historical data, while this invention embeds a variable lower limit accumulation mechanism with boundary constraints. By calculating the node time difference and subtracting the time window constant, a dynamic lower limit difference is obtained, forcibly eliminating invalid historical data that is too far from the current node, and retaining only the valid elements within the window for gamma weight calculation.

[0033] Based on the above improvements, the beneficial effects of this invention are that, through the state triggering mechanism, the model can adaptively match the optimal order according to the data fluctuation characteristics of different geological evolution stages, breaking the limitation of the original model statically processing all sequences and realizing the dynamic and refined prediction strategy; the boundary constraint mechanism effectively blocks the interference of long-term invalid geological data on the current prediction, significantly reduces computational redundancy, and while improving the model's fitting accuracy for the dynamic helium accumulation probability, it greatly enhances the system's computational efficiency and robustness in complex multi-stage tectonic evolution scenarios.

[0034] In this embodiment, the hydrocarbon accumulation map generation module specifically includes: The helium dynamic accumulation probability values ​​of all grids are imported into a three-dimensional spatial coordinate system. The helium dynamic accumulation probability values ​​of adjacent grids are read. When the difference between the helium dynamic accumulation probability values ​​of adjacent grids is less than a preset smoothing difference threshold, the helium dynamic accumulation probability values ​​of adjacent grids are added together and divided by a preset smoothing divisor constant to calculate the average probability value. The helium dynamic accumulation probability values ​​of adjacent grids are then replaced with the average probability value to complete the three-dimensional spatial smoothing interpolation. The logic of the smoothing replacement process is as follows: In the field of three-dimensional geological modeling and spatial data visualization, the original predicted values ​​of single-point grids are often limited by local discretization errors of geological parameters or small perturbations of inversion algorithms, resulting in unreasonable abrupt jump values. The smoothing difference threshold is used to identify and filter out such non-real geological abrupt changes in data noise reduction. The threshold is set when the difference between adjacent grids is less than the threshold value, indicating that the two predicted values ​​are within the same reasonable probability band. In this case, the mean replacement is not to negate the original prediction result of a single point, but to use the spatial neighborhood weighted smoothing algorithm to reduce noise and reconstruct the local data field to eliminate the numerical jump caused by discretization error, so that the probability distribution field of helium accumulation in three-dimensional space can be continuously transitioned, thereby truly reflecting the objective physical law of continuous migration and accumulation of helium in underground reservoirs, ensuring the overall accuracy of the prediction results in three-dimensional space and the rationality of geological significance. The preset smoothing difference threshold value is set based on the continuity characteristics of the probability distribution of three-dimensional spatial grids, and the normal gradual change limit of the probability values ​​between adjacent grids is calculated and set. The value range is 0.03 to 0.08. The preset smoothing divisor constant is 2. The dynamic helium accumulation probability value after smooth replacement is compared with a preset probability threshold. When the dynamic helium accumulation probability value after smooth replacement is greater than the preset probability threshold, the corresponding grid is marked as a candidate accumulation grid. The outer edges of all adjacent candidate accumulation grids are connected end to end to form a closed polygon boundary. The preset probability threshold is set based on the probability distribution characteristics of the known helium accumulation areas in the target area. The lower probability quantile that can distinguish between effective accumulation areas and non-accumulation areas is selected and set. The value range is 0.6 to 0.8. Read the dynamic helium accumulation probability values ​​of all candidate accumulation grids and compare them with each other. Find the grid with the largest value as the highest probability grid. Read the effective helium expulsion time period corresponding to the highest probability grid from the dynamic accumulation prediction module and mark it as the best accumulation time window. Read the leakage coefficients corresponding to the highest probability grid and adjacent grids. Grids with leakage coefficients greater than a preset risk coefficient threshold are marked as high-risk grids. Connect the outer edges of all adjacent high-risk grids one end to the other to form a leakage risk zone. The preset risk coefficient threshold is set based on the distribution characteristics of leakage coefficients of known leakage fault zones in the target area. The lower limit quantile that can define significant leakage risk is selected and set, with a value range of 0.3 to 0.5. The label of the optimal reservoir formation time window and the polygon boundary of the leakage risk zone are superimposed on the polygon boundary of the candidate reservoir formation grid. Generate and output a polygon map of the spatial distribution of the favorable helium exploration area labeled with the optimal reservoir formation time window and the leakage risk zone.

[0035] Example 1: To verify the feasibility of this invention in practice, it was applied to a three-dimensional spatial prediction of hydrocarbon accumulation probability and optimization of exploration target areas in a high-helium exploration area of ​​the Ordos Basin, a key helium-bearing natural gas basin in a certain province. This exploration area spans three municipal-level administrative regions, with a total exploration area of ​​approximately 18,000 square kilometers. It has a complex geological structure, densely developed basement faults, and has undergone multiple phases of tectonic evolution. Helium, as an extremely precious strategic and scarce resource, has a very complex accumulation mechanism, characterized by wide sources, strong diffusion, and easy loss. Traditional helium exploration and evaluation methods mainly rely on qualitative analysis of single static geological elements, separating helium generation calculation, fault source determination, and caprock sealing evaluation, lacking a spatiotemporal dynamic matching mechanism. Especially when faced with the frequent switching of fault opening and closing states caused by multiple tectonic movements in the exploration area, traditional methods cannot accurately define the synchronous relationship between the effective helium expulsion period and the fault source period. Moreover, when processing massive geological time series data, the accuracy of hydrocarbon accumulation probability prediction is often low due to interference from invalid historical data from the distant past, which cannot meet the actual needs of high-precision exploration and optimization in three-dimensional space.

[0036] In practical deployment, the method of this invention divides the study area into a 500m x 500m three-dimensional grid, fully integrating helium generation and expulsion dynamics determination with an improved fractional-order grey prediction model to perform intelligent evaluation. The system first extracts uranium-thorium abundance data and burial history time span data from the basement rocks, combines the radioactive decay yield constant to calculate the helium generation amount and internal pressure at each stage, and simultaneously reads the burial depth and maps it to obtain the basement porosity value. Then, it calculates the overlying strata pressure and hydrostatic pressure to obtain the basement pore fluid pressure. The system compares the internal pressure of helium generation and the basement pore fluid pressure period by period to accurately determine the effective helium expulsion time period and calculate the helium expulsion intensity. Regarding fracture determination, the system extracts burial history and paleotectonic evolution data to reconstruct paleostress results, calculates the friction coefficient and smear ratio according to the fracture strike and dip angle projection, comprehensively determines whether the fracture is in an open, conductive, or closed state at each stage, and performs a time-axis synchronous comparison with the effective helium expulsion time period to calculate the source efficiency coefficient and leakage coefficient. In the core dynamic hydrocarbon accumulation prediction stage, the system concatenates the aforementioned parameters into a one-dimensional input vector, which is then fed into an improved fractional-order grey prediction model. The model identifies data fluctuation characteristics by calculating the differences between adjacent elements and adaptively adjusts the fractional-order accumulation order. Particularly when processing historical data, the system introduces a variable lower limit accumulation mechanism embedded with boundary constraints. By calculating the difference between the node time difference and the dynamic lower limit, it automatically eliminates invalid historical elements that are too far removed from the current time node, retaining only valid historical elements to participate in the gamma function weight calculation, and outputting a fractional-order accumulation sequence after boundary constraints. Subsequently, the development term and action term are calculated through a differential coupling module, and residual corrections are added to obtain the helium dynamic hydrocarbon accumulation probability value for the current time node. Finally, the system imports the probability value into three-dimensional coordinates, performs smooth interpolation on adjacent grids, and combines preset probability thresholds and risk coefficient thresholds to form polygonal boundaries for candidate hydrocarbon accumulation areas and missed risk areas. The optimal hydrocarbon accumulation time window label is overlaid to generate and output a polygonal map of the spatial distribution of favorable helium exploration areas. Table 1 below shows the comparative data of the method of this invention and the traditional static evaluation method in the helium accumulation prediction task in different geological zones of this exploration area over a period of six months: Table 1. Comparison of the overall performance of the present invention and traditional methods in three-dimensional hydrocarbon accumulation prediction of helium-bearing basins.

[0037] Based on the comparative data shown in Table 1, it can be seen that the improved fractional gray prediction method based on the boundary constraint variable lower limit accumulation mechanism proposed in this invention shows significant performance advantages over the traditional static evaluation method in the three-dimensional hydrocarbon accumulation prediction of helium-bearing basins. In particular, it has achieved comprehensive improvement in key indicators such as prediction accuracy, computation time, probability level matching degree, and control of misjudgment and omission.

[0038] In terms of prediction accuracy, this invention maintained a high level of over 91% in all five typical evaluation zones, far exceeding the average accuracy of approximately 56% for traditional methods. For example, in the "II complex fault block area," traditional methods rely on qualitative inference based on static fault geometry, which suffers from mismatches between fault opening and closing periods and helium expulsion periods, achieving only 55.0% accuracy. In contrast, this invention achieves 92.9% accuracy by dynamically comparing helium-generating internal pressure and basement pore fluid pressure over time, combined with paleogeological stress recovery to accurately determine the timeliness of fault source and leakage, effectively improving the ability to identify helium accumulation patterns in complex zones.

[0039] Regarding computational efficiency, this invention, through a variable lower limit accumulation mechanism embedded with boundary constraints, achieves automatic removal of invalid historical data and dynamic time window filtering, significantly shortening the model computation time. In the five evaluation zones, the average computation time is reduced from 16.9 hours using traditional methods to approximately 2.4 hours, a nearly seven-fold speedup, ensuring rapid processing of massive amounts of 3D mesh data. For example, the traditional computation for the "V-type multi-phase redevelopment zone" takes an average of 19.5 hours, while this invention only requires 2.6 hours, greatly improving the real-time processing and iterative capabilities of hydrocarbon accumulation probability prediction.

[0040] Regarding the probability level matching accuracy, this invention maintains a accuracy rate above 87% across all zones, while traditional methods typically achieve around 50%. Traditional methods usually employ fixed thresholds for level classification, which can easily lead to slight deviations or short-term anomalies being misjudged as serious risks, or significant losses being underestimated. This invention, however, utilizes an improved fractional-order grey prediction model with adaptive order adjustment and gamma function weighting mechanisms to make level judgments more refined and closer to reality, ensuring the accuracy and differentiation of subsequent exploration target area selection.

[0041] The invention also demonstrates significant advantages in controlling the false positive and false negative rates. The average false positive rate is controlled below 6%, and the false negative rate is around 3%. Compared with the traditional method's approximately 23% false positive and nearly 16% false negative, this invention significantly reduces redundant calculations and omissions, lowers exploration and deployment costs, and improves the reliability of target area delineation.

[0042] Overall, this invention achieves efficient, accurate, and interpretable hydrocarbon accumulation probability prediction in key geological scenarios such as "Key Zone I," "Complex Fault Block Zone II," and "Deep Hidden Zone III" by integrating dynamic helium generation and expulsion determination with an improved grey prediction mechanism. It possesses broad practical value and promising prospects for deployment and promotion. Particularly in multi-phase alteration zones and complex fault block zones, the improvement in accuracy and computational timeliness is particularly significant, filling the technical shortcomings of traditional methods in such dynamic risk control.

[0043] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A system for evaluating and selecting favorable helium exploration areas, characterized in that, Includes the following modules: The data acquisition and preprocessing module is used to simultaneously acquire multi-source data from the target area and preprocess it to obtain a basic geological dataset. The helium generation and helium expulsion calculation module is used to calculate the amount of helium generation and internal pressure of helium generation, as well as the pressure of basal pore fluid, based on the geological basic dataset. It compares the internal pressure of helium generation with the pressure of basal pore fluid period by period to obtain the effective helium expulsion time period and helium expulsion intensity value. The fracture source and leakage determination module is used to restore paleogeological stress results based on geological basic datasets, determine whether the fracture is in a state of open conduction or closed blockage at various periods, and obtain the source efficiency coefficient or the leakage coefficient. The caprock loss assessment module is used to calculate capillary resistance and microcrack opening width based on geological basic datasets, assess helium resistance and escape amount, and obtain the caprock loss resistance coefficient. The dynamic helium accumulation prediction module is used to input the effective helium venting time period, helium venting intensity value, source efficiency coefficient or leakage coefficient, and caprock anti-leakage coefficient into the improved fractional gray prediction model to obtain the dynamic helium accumulation probability value. The improved fractional gray prediction model includes an input module, a state triggering module, a fractional accumulation module, and a differential coupling module. The fractional accumulation module embeds a variable lower limit accumulation mechanism with boundary constraints. The hydrocarbon accumulation map generation module is used to draw polygon maps of the spatial distribution of favorable helium exploration areas based on the dynamic probability values ​​of helium accumulation.

2. The evaluation and selection system for favorable helium exploration areas according to claim 1, characterized in that, The data acquisition and preprocessing module specifically includes: Through geological exploration and experimental testing, multi-source data were simultaneously collected, including uranium-thorium abundance data of the basement rocks in the target area, data on the variation of basement porosity with depth, fracture geometry and mechanical parameters, test data on the distribution of micro- and nano-pore sizes in the caprock, current geostress data, data on the burial history evolution of the target area, and data on the regional paleotectonic evolution. The collected multi-source data is uniformly converted into a standard data format. The three-dimensional spatial coordinate reference system of the target area is read, and the spatial coordinates of the converted multi-source data are projected onto the three-dimensional spatial coordinate reference system. The projected multi-source data is aligned with the three-dimensional spatial grid. The multi-source data with the unified format and the three-dimensional spatial position alignment are integrated and packaged to output a geological basic dataset.

3. The evaluation and selection system for favorable helium exploration areas according to claim 1, characterized in that, The helium generation and excretion calculation module specifically includes: Extract uranium and thorium abundance data of basement rocks from the geological basic dataset, read the uranium and thorium content values, read the time span values ​​of each geological period from the burial history evolution data of the target area, multiply the uranium and thorium content values ​​by the corresponding preset radioactive decay helium production rate constant and multiply by the time span value to obtain the uranium and thorium helium production amount, add the uranium helium production amount and thorium helium production amount to calculate the helium production amount of the target area in each geological period; Extract the data on the variation of basal porosity with depth and the burial history evolution data of the target area from the geological basic dataset. Read the burial depth values ​​for each geological period from the burial history evolution data of the target area. Find the porosity value corresponding to the same depth value from the data on the variation of basal porosity with depth. If no value is found, select the two values ​​with the closest depth and calculate the average value to obtain the corresponding basal porosity. The formation temperature values ​​for the corresponding geological period are read from the burial history evolution data of the target area. The helium generation amount is divided by the basement porosity value to obtain the helium generation concentration per unit pore. The helium generation concentration per unit pore is multiplied by the preset ideal gas constant and then by the formation temperature value to calculate the helium generation internal pressure for each geological period. The thickness and density of the overlying strata for each geological period are read from the burial history evolution data of the target area. The thickness of the overlying strata is multiplied by the density of the strata and then by a preset gravitational acceleration to obtain the pressure of the overlying strata. The density of the formation water is multiplied by a preset gravitational acceleration and then by the equivalent vertical height of the liquid column for the corresponding geological period, which is the thickness of the overlying strata or the paleoburial depth for the corresponding period, to obtain the hydrostatic pressure. Subtract the hydrostatic pressure value from the overlying rock pressure value and multiply it by the basement porosity value of the corresponding geological period. Add the hydrostatic pressure value to the product to calculate the basement pore fluid pressure for each geological period. The helium generation internal pressure and the basement pore fluid pressure of the same geological period are extracted to the same time point and compared period by period to determine whether the helium generation internal pressure of each geological period is greater than the basement pore fluid pressure. When the helium generation internal pressure of a certain geological period is greater than the basement pore fluid pressure, the corresponding geological period is recorded as the effective helium release period. The excess pressure is obtained by subtracting the base pore fluid pressure from the generated helium internal pressure during the effective helium purging period. The excess pressure is then multiplied by a preset helium purging efficiency coefficient to calculate the corresponding helium purging intensity.

4. The evaluation and selection system for favorable helium exploration areas according to claim 1, characterized in that, The fracture source and leakage determination module specifically includes: Extract fracture geometry and mechanical parameters, target area burial history evolution data and regional paleotectonic evolution data from the geological basic dataset. Read the stratigraphic thickness values ​​for each geological period from the target area burial history evolution data, and read the tectonic stress direction and magnitude values ​​for each geological period from the regional paleotectonic evolution data. The gravity stress value is obtained by multiplying the stratigraphic thickness by the preset rock density and the preset gravitational acceleration. This value is then added to the tectonic stress value. The sum is then projected and decomposed according to the tectonic stress direction to calculate the paleostress results for each geological period. The fracture strike, fracture dip angle and rock tensile strength values ​​are extracted from the fracture geometry and mechanical parameters. The paleostress results of each geological period are projected and calculated according to the fracture strike and fracture dip angle values ​​to obtain the normal stress and shear stress values ​​along the fracture surface. The friction coefficient value is obtained by dividing the shear stress value by the normal stress value. By comparing the friction coefficient value with the preset critical friction coefficient, when the friction coefficient value is greater than the preset critical friction coefficient, the stress state of the fracture zone is determined to be stress-induced fracture; when the friction coefficient value is less than or equal to the preset critical friction coefficient, the stress state of the fracture zone is determined to be stress-induced closure. Extract the fracture displacement and fault gouge thickness values ​​from the fracture geometry and mechanical parameters. Divide the fracture displacement value by the fault gouge thickness value to obtain the smearing ratio. Compare the smearing ratio with the preset critical smearing ratio. When the smearing ratio is greater than the preset critical smearing ratio, the fault gouge smearing is determined to be discontinuous. When the smearing ratio is less than or equal to the preset critical smearing ratio, the fault gouge smearing is determined to be continuous. Based on the combined stress state of the fault zone and the smearing of the fault gouge, when the stress state of the fault zone is stress-induced rupture and the smearing of the fault gouge is discontinuous, the fault is determined to be in a fractured and conductive state at each stage. When the stress state of the fault zone is stress-induced closure or the smearing of the fault gouge is continuous, the fault is determined to be in a closed and sealed state at each stage. Extract the effective helium release time period from the records, compare the period when the fracture was in the open and conductive state with the effective helium release time period on the time axis, and determine whether the two coincide at the time point. When the period when the fracture is in the open-conducting state coincides with the effective helium evacuation period, the time span of the period when the fracture is in the open-conducting state is divided by the time span of the effective helium evacuation period to calculate the source aging coefficient of the fracture under the corresponding effective helium evacuation period. When the period when the fracture is in the open-conducting state does not coincide with the effective helium evacuation period, the time span of the period when the fracture is in the open-conducting state does not coincide is divided by the time span of the effective helium evacuation period to calculate the loss and leakage coefficient of the fracture under the corresponding effective helium evacuation period.

5. The evaluation and selection system for favorable helium exploration areas according to claim 1, characterized in that, The caprock anti-scattering assessment module specifically includes: Extract the micro-nano pore size distribution test data of the caprock from the geological basic dataset, read the maximum value of the pore radius from the micro-nano pore size distribution test data of the caprock, divide the preset helium interfacial tension value by the maximum value of the pore radius, and calculate the capillary resistance value. Extract the current geostress data from the geological basic dataset, read the maximum and minimum horizontal principal stress values ​​from the current geostress data, read the rock tensile strength and rock Young's modulus values ​​from the fracture geometry and mechanical parameters, subtract the minimum horizontal principal stress value from the maximum horizontal principal stress value and subtract twice the rock tensile strength value to obtain the stress difference value. Multiply the stress difference value by the preset microcrack height value and the preset crack half length value, divide the product result by the rock Young's modulus value to calculate the microcrack opening width value, multiply the capillary resistance value by the preset resistance weight coefficient to obtain the helium resistance value, and multiply the microcrack opening width value by the preset escape weight coefficient to obtain the helium escape amount value. The helium escape rate is multiplied by the preset regional geostress gradient constant to calculate the helium escape pressure. The helium resistance is subtracted from the helium escape pressure to obtain the anti-scattering difference. The anti-scattering difference is divided by the helium resistance to calculate the caprock anti-scattering coefficient.

6. The evaluation and selection system for favorable helium exploration areas according to claim 1, characterized in that, The dynamic hydrocarbon accumulation prediction module specifically includes: The input module extracts the effective helium release time period, helium release intensity value, source aging coefficient or leakage coefficient, and caprock anti-leakage coefficient to the same time node. The helium release intensity value, source aging coefficient or leakage coefficient, and caprock anti-leakage coefficient are concatenated into a one-dimensional array according to the chronological order of the effective helium release time period, and then normalized and input into the improved fractional gray prediction model. The state trigger module reads the values ​​of each element in the one-dimensional input vector in sequence. It subtracts the value of the previous element from the value of the next element in the one-dimensional input vector and takes the absolute value to obtain the absolute value of the adjacent difference. It adds up all the absolute values ​​of the adjacent difference and divides by the total number of the absolute values ​​of the adjacent difference to obtain the average rate of change value. It compares the average rate of change value with the preset fluctuation threshold. When the average rate of change value is greater than the preset fluctuation threshold, it outputs the first trigger signal. When the average rate of change value is less than or equal to the preset fluctuation threshold, it outputs the second trigger signal. The fractional accumulator module receives a first trigger signal or a second trigger signal. When the first trigger signal is received, the fractional accumulator order of the fractional accumulator module is set to a preset high-order value. When the second trigger signal is received, the fractional accumulator order of the fractional accumulator module is set to a preset low-order value. In the fractional accumulation module, a variable lower bound accumulation mechanism with boundary constraints is executed, a dynamic lower bound filtering operation is performed, and the fractional weight value corresponding to each valid historical element value is calculated using the gamma function, and the fractional accumulation sequence after boundary constraints is output. The differential coupling module reads the fractional-order cumulative sequence after boundary constraints, subtracts the previous fractional-order cumulative value from the next fractional-order cumulative value in the fractional-order cumulative sequence after boundary constraints to obtain the first-order difference sequence, and multiplies it by the preset development coefficient to obtain the development term value. The previous fractional-order cumulative value in the fractional-order cumulative sequence after boundary constraints is multiplied by the preset action coefficient to obtain the action term value. The development term value is added to the action term value to obtain the differential coupling prediction value. By adding the pre-set initial residual correction to the differential coupling prediction value, the dynamic helium accumulation probability value at the current time node is obtained, thus completing the prediction process of the improved fractional gray prediction model.

7. The evaluation and selection system for favorable helium exploration areas according to claim 6, characterized in that, The step of implementing a variable lower bound accumulation mechanism with boundary constraints in the fractional accumulation module involves performing a dynamic lower bound filtering operation, calculating the fractional weight value corresponding to each valid historical element value using a gamma function, and outputting the fractional accumulation sequence after boundary constraints. Specifically, this includes: In the fractional-order accumulation module, the variable lower limit accumulation mechanism of the boundary constraints is executed. The value of each element in the one-dimensional input vector is read in sequence as the current element value, the effective helium release time period corresponding to the current element value is read as the current time node, and all element values ​​before the current element value in the one-dimensional input vector are read as historical element values. The effective helium release period corresponding to the historical element value is read as the historical time node. The time span between the current time node and the historical time node is read from the burial history evolution data of the target area as the node time difference. The node time difference is subtracted from the preset time window constant to obtain the dynamic lower limit difference. When the difference between the dynamic lower limits is less than or equal to zero, the corresponding historical element value is removed. When the difference between the dynamic lower limits is greater than zero, the corresponding historical element value is retained as a valid historical element value, and the dynamic lower limit filtering operation is performed. Get the fractional order of the current element value in the fractional order accumulation module, set the position number of the current element value in the sequence to i, set the position number of the valid historical element value in the sequence to j, subtract j from i to get the time step value, subtract one from the fractional order accumulation to get the first parameter value, add the time step value to the first parameter value and subtract one to get the second parameter value. The gamma function is used to calculate the gamma function values ​​corresponding to the first and second parameter values ​​as the first gamma value and the second gamma value. The gamma function value corresponding to the fractional cumulative order is calculated as the third gamma value. The gamma function is multiplied from the beginning to the time step value, and the result of the multiplication is used as the time step factorial result. Multiply the first gamma value by the second gamma value to obtain the first product result, multiply the third gamma value by the time step factorial result to obtain the second product result, divide the first product result by the second product result to obtain the fractional weight value, and calculate the fractional weight value corresponding to each effective historical element value. Multiply each valid historical element value by its corresponding fractional weight value to obtain the historical weighted product result. Add all the historical weighted product results together and add the product result of the current element value multiplied by the fractional weight value corresponding to the current element value to obtain the fractional cumulative value of the current element. Arrange the fractional cumulative values ​​of all elements in the one-dimensional input vector in order to obtain the fractional cumulative sequence after boundary constraints.

8. The evaluation and selection system for favorable helium exploration areas according to claim 1, characterized in that, The hydrocarbon accumulation map generation module specifically includes: Import the helium dynamic accumulation probability values ​​of all grids into the three-dimensional spatial coordinate system, read the helium dynamic accumulation probability values ​​of adjacent grids, and perform three-dimensional spatial smoothing interpolation when the difference between the helium dynamic accumulation probability values ​​of adjacent grids is less than the preset smoothing difference threshold. The dynamic helium accumulation probability value after smooth replacement is compared with the preset probability threshold. When the dynamic helium accumulation probability value after smooth replacement is greater than the preset probability threshold, the corresponding grid is marked as a candidate accumulation grid. The outer edges of all adjacent candidate accumulation grids are connected end to end to form a closed polygon boundary. Read the dynamic helium accumulation probability values ​​of all candidate accumulation grids and compare them with each other. Find the grid with the largest value as the highest probability grid, and read the corresponding effective helium expulsion time period and mark it as the best accumulation time window. Read the leakage coefficients of the highest probability grid and its adjacent grids, mark the grids with leakage coefficients greater than the preset risk coefficient threshold as high-risk grids, connect the outer edges of all adjacent high-risk grids one end to the other to form a leakage risk zone, and overlay the label of the optimal accumulation time window with the polygon boundary of the leakage risk zone onto the polygon boundary of the candidate accumulation grid to generate and output a polygon map of the spatial distribution of the favorable helium exploration area.

Images