A brittle rock full life cycle Helmholtz free energy storage calculation method, system, device and medium
By constructing a full life-cycle model based on neural networks and physical methods, and combining multimodal data and Gaussian function correction, the bias problem of existing rock mechanics models when processing multimodal geological data is solved, enabling accurate calculation and stability prediction of energy storage in brittle rocks, and supporting safety analysis in deep engineering.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-04-09
- Publication Date
- 2026-07-03
AI Technical Summary
Existing rock mechanics models are difficult to effectively integrate multimodal geological data, resulting in deviations between predictions and actual conditions when dealing with heterogeneity and anisotropy. Furthermore, machine learning methods lack generalization ability.
A full life cycle model based on neural networks and physical methods is used, combined with multimodal feature data, to predict the evolution of Helmholtz free energy of brittle rocks through error analysis and physical constraint correction. The model includes an input layer, a physical constraint layer, and a prediction layer. A Gaussian function is used as a physical prior, and the model output is corrected to ensure computational stability.
It enables accurate calculation of energy storage throughout the entire life cycle of brittle rocks, improves the robustness and computational stability of the model, is applicable to rock mass engineering analysis under complex geological conditions, and supports the prediction of stability and safety in deep engineering.
Smart Images
Figure CN122334600A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of rock science, and in particular relates to a method, system, device and medium for calculating the Helmholtz free energy storage throughout the entire life cycle of brittle rocks. Background Technology
[0002] In deep rock engineering, the energy evolution of rocks under load is a core basis for judging the stability of the project. However, most existing rock mechanics constitutive models are derived from continuum mechanics or damage mechanics. While these models can reflect the basic deformation characteristics of brittle rocks, they often struggle to adapt to the heterogeneity and anisotropy of rocks when dealing with the combined effects of multimodal data (such as mineral composition, grain size, and microscopic geological features like structural planes). This leads to discrepancies between predicted results and actual conditions. Although recent studies have attempted to use machine learning to predict rock mechanical behavior, these are mostly black-box fittings with poor generalization ability, failing to provide effective predictions. Therefore, there is an urgent need for a method that can integrate multimodal geological data and accurately calculate the Helmholtz free energy storage throughout the rock's entire lifecycle. Summary of the Invention
[0003] The purpose of this invention is to provide a method, system, device, and medium for calculating the Helmholtz free energy storage of brittle rocks throughout their entire life cycle, in order to solve the problems existing in the prior art.
[0004] In a first aspect, to achieve the above objective, this invention provides a method for calculating the Helmholtz free energy storage of brittle rocks throughout their entire life cycle, including:
[0005] Acquire multimodal characteristic data of the brittle rock to be analyzed. The multimodal characteristic data includes the structural features of the brittle rock to be analyzed, mineral composition, mineral grain size, stress and strain values at different stages, and elastic energy, dissipated energy and total energy corresponding to different stress and strain values.
[0006] The multimodal characteristic data of the brittle rock to be analyzed are input into the full life cycle model for classification and prediction to obtain the Helmholtz free energy evolution law of the brittle rock to be analyzed in the whole life cycle; wherein, the full life cycle model is constructed based on neural networks and physical methods, including an input layer, a physical constraint layer and a prediction layer connected in sequence.
[0007] Optionally, the training process of the full lifecycle model specifically includes:
[0008] Acquire training data, which includes multimodal feature training data and corresponding energy evolution law labels;
[0009] Construct an initial full lifecycle model by inputting the training data into the initial lifecycle model;
[0010] Physical constraints are introduced to guide the model in learning the laws of energy evolution; the physical constraints are Gaussian function constraints on energy evolution.
[0011] Error analysis is used to identify the perturbations of multimodal data on the model output.
[0012] The prediction model is corrected based on the error analysis results to obtain the trained lifecycle model.
[0013] Optionally, the process of acquiring the training data specifically includes:
[0014] Several rock samples were prepared, and the initial static information and multimodal information of each rock sample were obtained. The obtained initial static information and multimodal information were subjected to spatiotemporal alignment processing, feature extraction and cross-modal data fusion modeling to obtain a database including multimodal feature data.
[0015] Optionally, the step of correcting the prediction model based on the error analysis results specifically includes:
[0016] Determine whether there is oscillation in the output of the initial full life cycle model. If there is oscillation, correct the Gaussian function of the physical constraint layer and ensure that the denominator of the initial full life cycle model is not zero when calculating energy. Substitute the corrected Gaussian function into the initial full life cycle model for calculation and prediction to obtain the trained life cycle model.
[0017] Optionally, the processing procedure of the full lifecycle model specifically includes:
[0018] The multimodal characteristic data of the brittle rock to be analyzed are input into the full life cycle model, and the multimodal characteristic data are aligned and fused to obtain the feature vector;
[0019] The eigenvectors are input into the physical constraint layer. Based on the mechanical mechanism of energy drop after peak in brittle rock, the Gaussian function is used as the initial a priori form of the recoverable evolution of the Helmholtz free energy to solve the preliminary energy evolution function with strain.
[0020] By inputting the preliminary energy evolution function with strain into the prediction layer, constitutive equations that can describe the entire life cycle behavior of brittle rocks are derived.
[0021] Secondly, to achieve the above objectives, this invention provides a Helmholtz free energy storage calculation system for brittle rocks throughout their entire life cycle, comprising:
[0022] The data acquisition module is used to acquire multimodal characteristic data of the brittle rock to be analyzed. The multimodal characteristic data includes the structural features of the brittle rock to be analyzed, mineral composition, mineral grain size, stress and strain values at different stages, and elastic energy, dissipated energy and total energy corresponding to different stress and strain values.
[0023] The Helmholtz free energy storage calculation module is used to input the multimodal characteristic data of the brittle rock to be analyzed into the full life cycle model for classification and prediction, so as to obtain the evolution law of the Helmholtz free energy of the brittle rock to be analyzed in the whole life cycle; wherein, the full life cycle model is constructed based on neural network and physical derivation, including an input layer, a physical constraint layer and a prediction layer connected in sequence.
[0024] Thirdly, to achieve the above objectives, the present invention provides an electronic device, including a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to perform the Helmholtz free energy storage calculation method for the entire life cycle of brittle rocks as described in the first aspect.
[0025] Fourthly, to achieve the above objectives, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements a method for calculating the full life-cycle Helmholtz free energy storage of brittle rocks as described in the first aspect.
[0026] The technical effects of this invention are as follows:
[0027] This invention breaks through the limitations of traditional models that only consider a single mechanical parameter. It can effectively integrate multi-source geological information such as mineral composition, grain size, and structural planes, quantify the coupled influence of different factors on rock energy evolution, and is suitable for rock mass engineering analysis under complex geological conditions.
[0028] This invention enables dynamic tracking of energy storage throughout the entire process of brittle rock development, from compaction and elasticity to crack propagation and post-peak failure. Through an error back-calculation mechanism, the model can adaptively correct computational anomalies caused by data discreteness, thereby improving the robustness and computational stability of the algorithm in engineering applications.
[0029] This invention provides original technical support for predicting the confining pressure failure process in deep rock engineering, helps to establish a dynamic database for rock stability analysis in deep engineering, and has important practical engineering significance for ensuring the long-term stability and safety of deep resource extraction and underground space development. Attached Figure Description
[0030] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0031] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:
[0032] Figure 1 These are noise points in the full life cycle theoretical model in this embodiment of the invention;
[0033] Figure 2 This is a Gaussian curve in an embodiment of the invention that can store Helmholtz free energy and recover some data points;
[0034] Figure 3 This is the theoretical model curve after calculating the elastic energy function using this embodiment of the invention;
[0035] Figure 4 This describes the multimodal data content and corresponding acquisition process in an embodiment of the present invention.
[0036] Figure 5 This is a flowchart illustrating the implementation of an embodiment of the present invention. Detailed Implementation
[0037] Various exemplary embodiments of the present invention will now be described in detail. This detailed description should not be considered as a limitation of the present invention, but rather as a more detailed description of certain aspects, features, and embodiments of the present invention.
[0038] It should be understood that the terminology used in this invention is merely for describing particular embodiments and is not intended to limit the invention. Furthermore, with respect to numerical ranges in this invention, it should be understood that each intermediate value between the upper and lower limits of the range is also specifically disclosed. Every smaller range between any stated value or intermediate value within a stated range, and any other stated value or intermediate value within said range, is also included in this invention. The upper and lower limits of these smaller ranges may be independently included or excluded from the range.
[0039] Various modifications and variations can be made to the specific embodiments described in this specification without departing from the scope or spirit of the invention, as will be apparent to those skilled in the art. Other embodiments derived from this specification will also be obvious to those skilled in the art. This application specification and embodiments are merely exemplary.
[0040] The terms “include,” “including,” “have,” “contain,” etc., used in this article are all open-ended terms, meaning that they include but are not limited to.
[0041] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.
[0042] Example 1
[0043] like Figure 1 - Figure 5 As shown in the figure, this embodiment provides a method for calculating the energy storage of brittle rocks throughout their entire life cycle, including: acquiring multimodal characteristic data of the brittle rock to be analyzed, wherein the multimodal characteristic data includes the structural features of the brittle rock to be analyzed, mineral composition, mineral grain size, stress and strain values at different stages, and elastic energy, dissipated energy and total energy corresponding to different stress and strain values; inputting the multimodal characteristic data of the brittle rock to be analyzed into a life cycle model for classification and prediction, thereby obtaining the evolution law of the Helmholtz free energy of the brittle rock to be analyzed throughout its entire life cycle; wherein the life cycle model is constructed based on neural networks and physical methods, including an input layer, a physical constraint layer and a prediction layer connected in sequence.
[0044] The decrease in the Helmholtz free energy of a rock is equal to the maximum work done by the rock on the surroundings. The Helmholtz free energy can be regarded as the ability of a rock to do work under isothermal conditions. Its main function is to be used as a derivation of theoretical models to predict the stability of rocks.
[0045] (1)
[0046] (2)
[0047] (3)
[0048] In the formula, This represents the stress experienced by the rock throughout its entire life cycle. This represents the strain experienced by a rock throughout its entire life cycle. This represents the Helmholtz free energy of rock. It indicates the degree of rock fracture under load. This represents the energy dissipated when rocks fracture. This represents the recoverable portion of the Helmholtz free energy. It is the unrecoverable portion of the Helmholtz free energy, which can be obtained through subsequent derivations of the full life cycle model. , These are weight parameters. It is a parameter reflecting the properties of rocks. Unlike the elastic energy storage previously studied by scholars, elastic energy can be measured experimentally.
[0049] This embodiment introduces the Gaussian function as a physical prior into the machine learning model, so that the prediction of energy storage strictly follows the physical law of rock failure of rising-peak-fall, avoiding the physical paradox that may be generated by a pure data-driven model. The introduction of the correction function further eliminates the noise interference brought by multimodal data, making the prediction results more consistent with the actual rock fracture process.
[0050] This embodiment discloses a method for calculating the full life-cycle energy storage of multimodal brittle fracture rocks based on machine learning. The calculation method includes the following three aspects:
[0051] (1) Establish a multimodal database characterizing the features of brittle rocks, including: stress and strain values of rocks at different stages; elastic energy, dissipated energy, and total energy corresponding to different stress and strain values; prefabricated structural features of the sample; mineral composition of the rock; and the size of rock mineral grains.
[0052] (2) Model training, including: combining data model with physical prediction, based on the full life cycle characteristics of brittle fracture rocks, rather than simply data fitting, predicting that the stress peak will fall to near zero, and further predicting that the energy storage change law is to rise to the peak and then fall to near zero, predicting that the function is a Gaussian function;
[0053] (3) Error analysis and reverse deduction, including:
[0054] To address the instability and noise points in constitutive curves caused by different factors in rock multimodal databases, the target algorithm in this embodiment focuses on the constitutive equation ( The recoverable portion of the Helmholtz free energy in ) Regarding the functional relationship of strain, a correction function for the energy storage Gaussian function is given based on whether oscillations exist in the constitutive relation, to ensure that the denominator does not reach zero. This correction function is applicable to the entire life cycle of rock under load, from the compaction stage, elastic stage, stable crack propagation stage, unstable crack propagation stage to the failure stage, such as... Figure 3 As shown in the figure, this embodiment achieves accurate calculation of the highly nonlinear energy storage law generated by the combined effect of multimodal data under complex stress fields. It has positive guiding significance for predicting the confining pressure failure process in deep rock mass engineering and provides technical support for the stability and safety of the project.
[0055] The specific processing steps of an implementable, full lifecycle model include:
[0056] The multimodal characteristic data of the brittle rock to be analyzed are input into the full life cycle model. Feature extraction is performed on the multimodal characteristic data, and the original mechanical response signal (stress and strain) is upgraded to energy characteristics (elastic energy, etc.) with clear physical meaning.
[0057] Data of different dimensions and types are normalized and spliced together to form a feature vector describing the current state of the rock.
[0058] The eigenvectors are input into the physical constraint layer. Based on the mechanical mechanism of energy drop after peak in brittle rock, the Gaussian function is used as the initial a priori form of elastic energy evolution to solve the preliminary Helmholtz free energy evolution function with strain.
[0059] The obtained elastic energy evolution function is substituted into the pre-set constitutive model framework of rock life cycle, and through mechanical derivation or parameter solving, the constitutive equation that can fully describe the stress-strain behavior of brittle rock is finally output.
[0060] The feasible training process for the aforementioned full lifecycle model specifically includes:
[0061] Training data was acquired, several rock samples were prepared, and initial static information and multimodal information of each rock sample were obtained. The acquired initial static information and multimodal information were spatiotemporally aligned, feature extracted, and cross-modal data fusion modeled to obtain a database including multimodal feature data and corresponding energy evolution law labels.
[0062] Construct an initial full lifecycle model by inputting the training data into the initial lifecycle model;
[0063] Physical constraints are introduced to guide the model in learning the laws of energy evolution; the physical constraints are Gaussian function constraints on energy evolution.
[0064] Error analysis is used to identify the perturbation of multimodal data on the model output, and to determine whether there is oscillation in the output of the initial full life cycle model. If oscillation exists, the Gaussian function of the physical constraint layer is corrected, and the denominator of the initial full life cycle model is not zero when calculating energy. The corrected Gaussian function is substituted into the initial full life cycle model for calculation and prediction to obtain the trained life cycle model.
[0065] The feasible process for constructing the aforementioned database specifically includes:
[0066] Phase 1: Sample Preparation and Initial Characterization
[0067] 1) Sample selection and design:
[0068] Lithology: At least three typical deep rock samples (e.g., granite, marble, sandstone) should be included, with no fewer than three parallel samples for each lithology to avoid randomness in experimental results.
[0069] Stress path: The design includes a combination of experimental sequences with different confining pressures (e.g., 0 MPa, 10 MPa, 20 MPa, 25 MPa, 30 MPa) and different cyclic loading and unloading gradients.
[0070] Sample preparation: Samples taken from the same rock block were processed into standard cylinders, and then processed into Φ50mm×H100mm and Φ25mm×H50mm according to the requirements of the experimental instruments.
[0071] 2) Initial static information acquisition (unstructured / semi-structured data entry):
[0072] Macroscopic geological description: High-resolution photography was performed on each sample to record visible textures and joints. Corresponding electronic archives of rock sample appearance characteristics were established.
[0073] Micromineralogical analysis:
[0074] Mineral distribution maps, grain size and shape statistics, and mineral composition quantification tables of the sample profiles were obtained by means of scanning electron microscopy (SEM) scattering imaging, energy dispersive spectroscopy (EDS), and X-ray diffraction (XRD).
[0075] Initial internal structure CT scan:
[0076] Using high-precision nano-CT, an initial scan of the sample is performed in a non-destructive manner to obtain a three-dimensional spatial distribution model of initial porosity, microfractures, and mineral phases. This serves as the "zero-state" baseline for all dynamic evolutions. These data will serve as the starting point for understanding rock heterogeneity.
[0077] Phase Two: Experiment on Synchronous Acquisition of Multimodal Information During Loading Process
[0078] 1) Experimental platform setup: True triaxial rigid testing machine, equipped with a high-precision servo control system to achieve stress / strain control for complex paths.
[0079] 2) Multimodal data acquisition, the specific process is as follows: Figure 4 As shown.
[0080] Phase 3: Data Processing, Feature Extraction, and Spatiotemporal Fusion Modeling
[0081] 1) Data preprocessing and spatiotemporal alignment:
[0082] Time alignment: Timestamps of all data streams (mechanics, AE, DIC image sequences) are synchronized to the same absolute time axis.
[0083] Spatial alignment: Establish a unified spatial coordinate system with the geometric center of the specimen as the origin. Unify all spatial data (CT data, AE positioning points, DIC coordinates) under this coordinate system.
[0084] 2) Core Feature Extraction:
[0085] Crack extraction from CT data: Image processing techniques are used to extract cracks from CT data at each loading stage. This includes ① 3D reconstruction: generating a 3D model of the crack network; ② Geometric quantization: calculating the total volume, surface area, number, length, width, and inclination angle of the cracks.
[0086] Extract parameters characterizing damage features from AE data: calculate the temporal changes of AE event rate, cumulative energy, and b-value, and analyze the spatial evolution of the source mechanism (tension / shear ratio).
[0087] Calculating energy from mechanical data: Numerical integration of the stress-strain curve to calculate the total energy, elastic energy, and dissipated energy at the time points to be studied.
[0088] 3) Cross-modal data fusion modeling (the "intelligent" core of the database): Establish a core "spatiotemporal event association table," where each row represents a key physical event, for example:
[0089] [Timestamp, Axial strain, Current stress state, AE event ID and its spatial coordinates (X,Y,Z), Fracturing type, Total volume increment of the local crack network at this moment, Average strain tensor of the region calculated by DVC, Current elastic energy, Current total dissipated energy].
[0090] Phase 4: Database Architecture and Service Platform Implementation
[0091] 1) Storage files include:
[0092] ①L0 raw files, stored in a standard directory (.raw CT projection, .tdms waveform, .tiff image sequence, .csv mechanical data).
[0093] ②L1 alignment and basic features: spatiotemporally aligned data packets, such as crack models, acoustic emission events, and stress-strain points with uniform timestamps and coordinates.
[0094] ③L2 feature data: Extract quantitative indicators from L1 data, such as crack volume, acoustic emission b-value, and energy evolution curve.
[0095] ④L3 fusion / correlation data: cross-modal correlation of data, such as "crack propagation volume corresponding to each elastic energy value".
[0096] Metadata module: Records the initial characteristics of each rock, including rock type, initial density, initial joint porosity, experimental loading path (confining pressure, loading control, loading speed, etc.).
[0097] Database selection: For unstructured data, such as CT images and waveforms, use object storage to manage files and record metadata indexes. For structured data (feature tables, relational tables), use a relational database (e.g., SQLite for personal use, MySQL for groups of 3-5 people).
[0098] 2) MATLAB Access and Services:
[0099] The process of accessing MATLAB programming:
[0100] data=get Sample Metadata (sample ID);
[0101] [ctVolume,info] = load CT Volume (sample ID, load Step);
[0102] aeTable=query AE(sample ID,time Range, Energy Threshold);
[0103] Energy Curves =calculate Energy (sample ID);
[0104] These functional programming data paths and query methods allow researchers to efficiently access data for analysis and modeling without needing to understand complex database storage structures.
[0105] As natural geological materials and engineering environments, rock masses often have their original stress equilibrium disrupted by disturbances during construction, leading to changes in their internal stress and energy fields. In the applicant's previous research, the plastic deformation of brittle-failed rocks was not significant during the fracturing process. The Helmholtz free energy of the rock sample system under load exhibited nonlinear separation characteristics. Based on the nonlinear separation parameters and analytical solutions of the fracturing evolution, a full-life-cycle constitutive model that can predict the entire damage and failure process was derived. This model accurately reflects the nonlinear mechanical properties of brittle rocks. However, with further research, after considering multimodal data reflecting the complex characteristics of brittle rocks, the study revealed a striking phenomenon: individual noise points appear on the constitutive curve of brittle-failed rocks, such as... Figure 1As shown. Through examination and reverse engineering of the theoretical model, it was found that elastic energy storage needs to be based on the full life cycle characteristics of brittle fracture rocks, rather than simply data fitting. Based on the stress peak falling to near zero, the variation law of Helmholtz free energy storage is further predicted to be rising to a peak and then falling to near zero. The predicted recoverable storage function of Helmholtz free energy is a Gaussian function, as shown. Figure 2 As shown. Through error analysis and back-calculation, the elastic energy storage function is corrected to obtain a predictive model that reflects the characteristics of rock deformation and failure. Figure 3 As shown in the figure, this embodiment enables a better integration of physical laws and artificial intelligence multimodal data, better solving practical engineering problems and having greater practical engineering significance. This embodiment can promote the advancement of rock mass fracture mechanics towards more complex multiphase and multimodal hybrid systems, establish a dynamic database for rock stability analysis in deep engineering, and provide original support for the intelligent construction of deep engineering projects.
[0106] The above description is merely a preferred embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
Claims
1. A method for calculating the Helmholtz free energy storage over the entire life cycle of brittle rocks, characterized in that, include: Acquire multimodal characteristic data of the brittle rock to be analyzed. The multimodal characteristic data includes the structural features of the brittle rock to be analyzed, mineral composition, mineral grain size, stress and strain values at different stages, and elastic energy, dissipated energy and total energy corresponding to different stress and strain values. The multimodal characteristic data of the brittle rock to be analyzed are input into the full life cycle model for classification and prediction to obtain the Helmholtz free energy evolution law of the brittle rock to be analyzed in the whole life cycle; wherein, the full life cycle model is constructed based on neural networks and physical methods, including an input layer, a physical constraint layer and a prediction layer connected in sequence.
2. The method for calculating the Helmholtz free energy storage of brittle rocks throughout their entire life cycle according to claim 1, characterized in that, The training process of the full lifecycle model specifically includes: Acquire training data, which includes multimodal feature training data and corresponding energy evolution law labels; Construct an initial full lifecycle model by inputting the training data into the initial lifecycle model; Physical constraints are introduced to guide the model in learning the laws of energy evolution; the physical constraints are Gaussian function constraints on energy evolution. Error analysis is used to identify the perturbations of multimodal data on the model output. The prediction model is corrected based on the error analysis results to obtain the trained lifecycle model.
3. The method for calculating the Helmholtz free energy storage of brittle rocks throughout their entire life cycle according to claim 2, characterized in that, The process of acquiring the training data specifically includes: Several rock samples were prepared, and the initial static information and multimodal information of each rock sample were obtained. The obtained initial static information and multimodal information were subjected to spatiotemporal alignment processing, feature extraction and cross-modal data fusion modeling to obtain a database including multimodal feature data.
4. The method for calculating the Helmholtz free energy storage of brittle rocks throughout their entire life cycle according to claim 2, characterized in that, The correction of the prediction model based on the error analysis results specifically includes: Determine whether there is oscillation in the output of the initial full life cycle model. If there is oscillation, correct the Gaussian function of the physical constraint layer and ensure that the denominator of the initial full life cycle model is not zero when calculating the Helmholtz free energy. Substitute the corrected Gaussian function into the initial full life cycle model for calculation and prediction to obtain the trained life cycle model.
5. The method for calculating the Helmholtz free energy storage of brittle rocks throughout their entire life cycle according to claim 1, characterized in that, The processing procedure of the full lifecycle model specifically includes: The multimodal characteristic data of the brittle rock to be analyzed are input into the full life cycle model, and the multimodal characteristic data are aligned and fused to obtain the feature vector; The eigenvectors are input into the physical constraint layer. Based on the mechanical mechanism of energy drop after peak in brittle rock, the Gaussian function is used as the initial a priori form of the recoverable evolution of the Helmholtz free energy to solve the preliminary energy evolution function with strain. By inputting the preliminary energy evolution function with strain into the prediction layer, constitutive equations that can describe the entire life cycle behavior of brittle rocks are derived.
6. A Helmholtz free energy storage calculation system for the entire life cycle of brittle rocks, characterized in that, include: The data acquisition module is used to acquire multimodal characteristic data of the brittle rock to be analyzed. The multimodal characteristic data includes the structural features of the brittle rock to be analyzed, mineral composition, mineral grain size, stress and strain values at different stages, and elastic energy, dissipated energy and total energy corresponding to different stress and strain values. The Helmholtz free energy storage calculation module is used to input the multimodal characteristic data of the brittle rock to be analyzed into the full life cycle model for classification and prediction, so as to obtain the evolution law of the Helmholtz free energy of the brittle rock to be analyzed in the whole life cycle; wherein, the full life cycle model is constructed based on neural networks and physical methods, including an input layer, a physical constraint layer and a prediction layer connected in sequence.
7. An electronic device, characterized in that, The device includes a memory and a processor, the memory being used to store a computer program, and the processor running the computer program to enable the electronic device to perform a method for calculating the Helmholtz free energy storage of brittle rocks throughout their entire life cycle according to any one of claims 1-5.
8. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed by a processor, implements a method for calculating the Helmholtz free energy storage of brittle rocks throughout their entire life cycle as described in any one of claims 1-5.