Method and system for evaluating groutability of ground grouting drill hole based on fuzzy dynamic model

Abstract

The invention provides a method and system for evaluating groutability of a grouting drill hole based on a fuzzy dynamic model, and the method comprises the steps: obtaining measured values of a plurality of parameters, including but not limited to drilling fluid consumption, gamma logging value, rock debris logging data and water pressure test pressure injection amount; according to each preset parameter evaluation grade, calculating the membership degree of the measured value to the corresponding parameter evaluation grade by adopting a membership degree function; adopting the membership degree of at least two actually-measured parameters to evaluate the credibility level confidence coefficient; and evaluating the groutability evaluation grade of the grouting drill hole according to the groutability grade confidence coefficient. The problem that a traditional evaluation method excessively depends on subjective experience and threshold judgment of a single parameter can be solved, the fuzzy dynamic evaluation method capable of fully utilizing multi-source actual measurement parameters is provided, and the regional water disaster treatment effect and the engineering decision level can be improved.

Description

Technical Field

[0001] This invention relates to the field of geological engineering technology, and in particular to a method and system for evaluating the groutability of grouting boreholes based on a fuzzy dynamic model. Background Technology

[0002] As an emerging field in mine water control engineering, surface area water hazard management still has methods for determining the injectability of grouting, which are still under development. Currently, in engineering practice, the injectability of boreholes is mainly evaluated based on field parameters such as drilling fluid consumption, gamma logging values, cuttings content, and water injection volume in pressure tests. However, existing methods often rely on threshold judgments of single parameters, which has significant limitations. On the one hand, the failure to fully integrate and utilize multi-source construction data and the neglect of the interrelationships and synergistic effects among various indicators resulted in incomplete information utilization. On the other hand, under heterogeneous and significantly nonlinear geological conditions, the judgment method based on a single threshold is not accurate enough and has poor adaptability.

[0003] Furthermore, existing methods lack a systematic multi-index comprehensive evaluation model, quantitative analysis methods are insufficient, and the overall technical system has not yet achieved process-oriented, systematic, and quantitative results.

[0004] Therefore, it is necessary to develop a method for assessing the injectability of ground grouting holes that can improve the effectiveness of regional water hazard control and engineering decision-making. Summary of the Invention

[0005] The purpose of this invention is to provide a method for evaluating the injectability of ground grouting boreholes based on a fuzzy dynamic model, in order to solve the problems of insufficient information utilization, poor adaptability, and lack of systematic quantitative models in the existing determination methods described in the background art.

[0006] The above-mentioned technical objectives of the present invention are mainly achieved through the following technical solutions.

[0007] This invention proposes a method for evaluating the groutability of grouting boreholes based on a fuzzy dynamic model, comprising: Obtain measured values ​​of multiple parameters, including but not limited to drilling fluid consumption, gamma logging value, cuttings logging data, and water injection volume in the pressure test; calculate the membership degree of the measured values ​​to the corresponding parameter evaluation level using a membership function based on the preset evaluation levels of each parameter; assess the confidence level of the injectability level using the membership degrees of at least two of the measured parameters; and evaluate the injectability of the grouting borehole based on the confidence level of the injectability level.

[0008] The parameter evaluation levels are first fuzzy levels, second fuzzy levels, and third fuzzy levels constructed using a first threshold, a second threshold, and a third threshold.

[0009] The membership function is: In the formula: r 1 （x） The first fuzzy level membership function, r 2 （x） For the second fuzzy level membership function, r 3 （x） For the third fuzzy level membership function, y These are the measured values ​​of the parameters. x 1 is the first threshold. x 2 The second threshold, x 3 This is the third threshold.

[0010] Construct a fuzzy relation matrix R consisting of the membership degrees of the measured parameters, wherein the dimension of R is n×3; Preferably, the confidence level of the registrability level is assessed based on the over-weighted method, wherein the calculation method for the parameter weights specifically includes: First, calculate the exceedance ratio of each parameter separately, using the following formula:

[0011] In the formula: e i This represents the percentage of parameters that exceed the limit for the i-th parameter. Let y be the optimal critical value of the i-th parameter in the corresponding evaluation level. i Let be the measured value of the i-th parameter; Preferably, the optimal critical value It equals the first threshold; Then, calculate the weight of each parameter separately, using the following formula:

[0012] In the formula: W i Let be the weight of the i-th parameter, and n be the number of parameters; Finally, a parameter weight set, W = [W1, W2, ..., Wn], is constructed, consisting of the weights of each of the aforementioned parameters.

[0013] The parameter weight set With the fuzzy relation matrix Perform a composite operation to obtain the confidence level of the inscribeability level, i.e., the membership degree of the inscribeability level U = [U1, U2, U3], calculated using the following formula: , ,

[0014] In the formula: r jk Let r be a matrix element representing the membership degree of the j-th factor to the k-th level. jk ∈[0,1]; The evaluation level corresponding to the maximum value among [U1, U2, U3] is taken as the first inscribeability evaluation level.

[0015] Preferably, the confidence level of the inscribeability level can also be predicted based on a deep neural network. The deep neural network includes a fuzzy attention layer, a wavelet transform layer, a residual convolutional layer, and an output layer. The prediction method specifically includes: converting the membership degree of the measured parameters into an n×3 fuzzy feature vector, where n is the number of parameters; inputting the fuzzy feature vector into the trained deep neural network for prediction; obtaining the confidence level of the inscribeability level, i.e., the probability distribution of the three inscribeability levels; and outputting the inscribeability level corresponding to the highest probability as the second inscribeability evaluation level.

[0016] The training method of the deep neural network includes: First, acquiring historical measured data as training samples, the historical measured data including measured values ​​of multiple parameters, and labeling each sample with an annotability level; then, performing fuzzy feature extraction, calculating the membership degree of each measured value to different evaluation levels based on preset evaluation levels and membership functions of each parameter, and combining the membership degrees of all parameters into an n×3 fuzzy feature vector, where n is the number of parameters; finally, inputting the n×3 fuzzy feature vector into the deep neural network for training to obtain the trained deep neural network.

[0017] Preferably, the confidence level of the inscribeability level can also be evaluated based on the overweighting method and a deep neural network, specifically including: calculating a first inscribeability evaluation level based on the overweighting method; predicting a second inscribeability evaluation level based on the deep neural network; comparing the first inscribeability evaluation level and the second inscribeability evaluation level, and if they are consistent, outputting the inscribeability evaluation level; if they are inconsistent, outputting the membership degree of the inscribeability level, and / or the parameter weight value, and / or the probability distribution of the inscribeability level, and / or the first inscribeability evaluation level, and / or the second inscribeability evaluation level.

[0018] This invention also proposes a system for evaluating the groutability of grouting boreholes based on a fuzzy dynamic model, comprising: The data acquisition module is used to acquire measured values ​​of multiple field parameters, including but not limited to well fluid consumption, gamma logging values, cuttings logging data, and water injection volume in the pressure test; the fuzzy preprocessing module is used to calculate the membership degree of the measured values ​​to the corresponding parameter evaluation level using a membership function based on the preset evaluation level of each parameter; the injectability assessment module is used to assess the confidence level of the injectability level using the membership degrees of at least two measured parameters; and the result output module is used to output the injectability assessment result based on the confidence level of the injectability level.

[0019] The inscribeability assessment module further includes: a fuzzy comprehensive evaluation module, used to calculate and obtain a first inscribeability assessment level based on the over-weighted method; and / or a deep neural network module, used to predict and obtain a second inscribeability assessment level based on the deep neural network.

[0020] Compared with existing technologies, this invention constructs a dynamically adjustable fuzzy comprehensive evaluation system that integrates multi-source construction parameters and uses real sample data to train and adaptively optimize the neural network weights. This effectively overcomes the problems of traditional methods, such as over-reliance on subjective experience, poor interpretability, and insufficient adaptability of single thresholds, achieving a scientific, systematic, and quantitative assessment of borehole grouting injectability. This scheme significantly improves the objectivity, accuracy, and engineering applicability of injectability judgment, providing highly reliable and scalable technical support for surface water hazard control, thereby enhancing the accuracy of engineering decisions and the effectiveness of control measures. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments or prior art of this specification, the drawings used in the description of the embodiments or prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a schematic diagram of the method for evaluating the grouting borehole injectability based on a fuzzy dynamic model provided by the present invention. Figure 2 A flowchart illustrating a method for evaluating the injectability of grouting boreholes, provided as an embodiment of this specification; Figure 3 A schematic diagram of a deep neural network structure provided in the embodiments of this specification; Figure 4 This is a flowchart of a deep neural network training process provided in the embodiments of this specification; Figure 5 A flowchart of deep neural network prediction provided for embodiments of this specification; Figure 6A flowchart illustrating another method for evaluating the injectability of grouting boreholes, provided in the embodiments of this specification; Figure 7 A schematic diagram of the system structure for evaluating the injectability of grouting boreholes based on a fuzzy dynamic model, provided by the present invention. Detailed Implementation

[0023] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this application.

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0025] In this embodiment of the invention, drilling fluid consumption (Q, unit: m³) is selected. 3 / h), natural gamma value (γ, unit: API), non-limestone fragment content (C, unit: %), and water pressure test injection volume (unit: m). 3 (h) are the four core measured parameters for evaluating the injectability of ground grouting boreholes.

[0026] For the above four parameters, a first threshold, a second threshold, and a third threshold are set according to their engineering response characteristics, and a first fuzzy level, a second fuzzy level, and a third fuzzy level are constructed in sequence.

[0027] To clearly illustrate the technical solution of the present invention, the following embodiments exemplarily provide a method for classifying evaluation levels and several specific numerical values. It should be understood that the evaluation level examples and other specific numerical values ​​involved in the embodiments are for illustrative purposes only and do not constitute a limitation on the scope of protection of the present invention.

[0028] Examples of evaluation levels are shown in Table 1:

[0029] Table 1 refer to Figure 1 This is a schematic flowchart of a method for evaluating the injectability of ground grouting boreholes based on a fuzzy dynamic model, provided by the present invention. The method includes: Step S1: Obtain the measured values ​​of multiple parameters.

[0030] For example, the measured parameters obtained on-site are: drilling fluid consumption Q = 12m 3 / h, natural gamma value γ=22API, non-limestone fragment content C=7%, water pressure test injection volume W=18m 3 / h.

[0031] Step S2: Based on the preset evaluation levels of various parameters, use the membership function to calculate the membership degree of the measured value to the corresponding parameter evaluation level.

[0032] The membership function is: In the formula: r 1 （x） The first fuzzy level membership function, r 2 （x） For the second fuzzy level membership function, r 3 （x） For the third fuzzy level membership function, y These are the measured values ​​of the parameters. x 1 is the first threshold. x 2 The second threshold, x 3 This is the third threshold.

[0033] Using the trapezoidal membership function, the actual measured value of the parameter y is transformed into membership degrees to three fuzzy levels D1, D2, and D3.

[0034] Based on the membership formula, and combining the evaluation levels in Table 1 of this embodiment and the measured values ​​of the field parameters in step S1, the membership degrees of each measured value are calculated: Membership of drilling fluid consumption: [0.60, 0.40, 0]; Membership of natural gamma values: [0.30, 0.70, 0]; Membership degree of non-limestone fragment content: [0, 0.60, 0.40]; Membership degree of water pressure test injection volume: [0, 0.4, 0.60].

[0035] Step S3: Evaluate the confidence level of the registrability grade using the membership degrees of at least two of the measured parameters.

[0036] Step S4: Assess the injectability of the grouting borehole based on the injectability level confidence level.

[0037] First embodiment: In the first embodiment of this specification, step S3 assesses the confidence level of the registrability grade based on the over-weighted method. The closer the actual value of the parameter is to or exceeds its optimal critical value, the greater its weight, indicating that the indicator is more important in the overall evaluation. The method flow is as follows: Figure 2 As shown.

[0038] Step S11: Obtain measured values ​​of parameters such as drilling fluid consumption, γ logging value, cuttings logging data, and water injection volume in the pressure test.

[0039] Step S21: Calculate the parameter membership degree using the trapezoidal membership function based on the preset evaluation level.

[0040] Step S311: Calculate the exceedance ratio of each parameter separately. The calculation formula is as follows:

[0041] In the formula: e i This represents the percentage of parameters that exceed the limit for the i-th parameter. Let y be the optimal critical value of the i-th parameter in the corresponding evaluation level; i is the measured value of the i-th parameter.

[0042] The optimal critical value in this embodiment It can be the first threshold.

[0043] When the measured value y i Approaching or exceeding the optimal critical value s i Then e i A value close to or greater than 1 indicates that the parameter performs well; when the measured value y i Below the optimal critical value s i Then e i A value less than 1 indicates that the parameter performs poorly.

[0044] The membership calculation method based on the degree of parameter exceeding the standard can adaptively adjust the boundary threshold, solving the problem of poor adaptability of the traditional fixed membership function.

[0045] Step S312: Calculate the weight of each parameter separately, using the following formula:

[0046] In the formula: The weight of the i-th parameter is given by n, which is the number of parameters used. Step S313: Construct a parameter weight set, W=[W1, W2, ..., Wn], consisting of the weights of the aforementioned parameters. This weight can reflect the degree of deviation of the actual measured value from the optimal value, thus achieving adaptive weighting.

[0047] Using the specific example of this embodiment, the measured values ​​of the field parameters from step S1 are adopted, and the first threshold of each parameter in Table 1 is taken as the optimal critical value. Based on the formulas provided in steps S311 and S312, the exceedance ratio of each parameter in this embodiment is calculated: The excess rate of drilling fluid consumption e1 = 2.40; The percentage of natural gamma values ​​exceeding the standard was e2 = 1.47; The exceedance rate of non-limestone rock fragments was e3=1.40; The rate of exceeding the pressure injection volume limit in the pressure test was e4 = 1.80; For e i Normalization yields the weights W = [0.34, 0.21, 0.20, 0.25].

[0048] Step S314: Construct a fuzzy relation matrix composed of the membership degrees of the measured parameters. Based on the membership degrees of each measured parameter calculated in the example of step S2, the fuzzy relation matrix R is:

[0049] The execution order of step S314 can be flexibly arranged according to actual needs. It can be executed before step S311, simultaneously with step S311, or before step S313.

[0050] Step S315: Perform a composite operation on the parameter weight set and the fuzzy relation matrix to obtain the additability level confidence, i.e., the additability level membership degree U=[U1, U2, U3], calculated using the following formula: , ,

[0051] In the formula: r jk Let rjk be a matrix element representing the membership degree of the j-th factor to the k-th level, where rjk∈[0,1]; According to the formula, the inscribeability membership degree U of the aforementioned specific example is calculated to be [0.28, 0.46, 0.26].

[0052] Step S41: Output the first registrability evaluation result. Take the evaluation level D2 corresponding to the maximum value among [U1, U2, U3] as the first registrability evaluation level. The registrability evaluation level of the aforementioned specific example is medium registrability.

[0053] Second embodiment: In the second embodiment of this specification, step S3 is to obtain the confidence level of the inscribeability level based on deep neural network prediction.

[0054] One structure of the deep neural network is as follows: Figure 3 As shown, it includes a fuzzy attention layer, a wavelet transform layer, a residual convolutional layer, and an output layer.

[0055] For example, the input to the neural network is a 12-dimensional fuzzy feature vector extracted from four original parameters. This 12-dimensional fuzzy feature vector is first weighted by a fuzzy attention layer to obtain 8-dimensional features; the weighted features are then input to a wavelet transform layer for wavelet decomposition to obtain 16-dimensional features; the resulting features are fed into a residual convolution module, sequentially undergoing one-dimensional convolution for core feature extraction, batch normalization (BN) to stabilize the training distribution, and nonlinear corrected linear unit (ReLU) activation. The original input is then fused through residual connections to improve gradient flow and feature reuse, resulting in 32-dimensional features; finally, the features are integrated by a fully connected layer in the output layer and a Softmax activation function is used to generate 3-dimensional features corresponding to three levels of injectability probability distributions.

[0056] Another structure of the deep neural network may include a fuzzy attention layer, a wavelet transform layer, a residual connection layer, a depthwise separable convolutional layer, a global average pooling layer, and a softmax output layer. The fuzzy attention layer generates an attention weight vector 'a' through a fully connected layer with sigmoid activation and multiplies it element-wise with the input fuzzy feature vector to achieve dynamic weighting of key features. The wavelet transform layer uses a fully connected layer with a sinusoidal activation function sin(2πx) to simulate the feature extraction capability of wavelet basis functions in the frequency domain, thereby enhancing the model's ability to express non-stationary features. The residual connection layer adds the output of the wavelet transform layer to the linear mapping result of another fully connected layer to alleviate the gradient vanishing problem in deep networks and improve the model's training stability. The depthwise separable convolutional layer reconstructs the residual features into a dimension (64, 1) and extracts local temporal dependencies through one-dimensional separable convolution (SeparableConv1D), maintaining feature extraction efficiency while reducing the number of parameters. The global average pooling layer compresses the features along the temporal dimension to obtain a fixed-length global feature representation. The output layer uses three types of Softmax activation functions, corresponding to three levels of indicability, and outputs the probability distribution of each level.

[0057] The neural network uses the Adam optimizer for parameter optimization, with an initial learning rate of 0.001 and a weight decay rate of 0.0001 to enhance generalization performance. The loss function chosen is Categorical Crosssentropy to supervise multi-class classification tasks with injectable levels. Early stopping and model checkpointing mechanisms are employed during training to ensure that training stops and optimal model weights are preserved when optimal performance is achieved.

[0058] The training process of the second embodiment is as follows: Figure 4 As shown.

[0059] Step S321: Obtain historical measured data as training samples, and label the annotability level of each sample. The historical measured data includes measured values ​​of multiple parameters such as drilling fluid consumption, natural gamma value, non-limestone cuttings content, and water pressure test injection volume, with each sample manually labeled with its injectability level.

[0060] Step S322: Fuzzy Feature Extraction Based on the preset evaluation levels and membership functions of each parameter, the membership degree of each measured value to different evaluation levels is calculated. Then, MinMaxScaler is used to normalize the features, and the membership degree of the parameters is converted into an n×3 fuzzy feature vector, where n is the number of parameters.

[0061] Step S323: Iterative training of the deep neural network to obtain the trained deep neural network. The processed samples are divided into training and validation sets in a ratio of, for example, 8:2. The input is as follows: Figure 3 The deep neural network shown uses a training set to iteratively train the model, and monitors the changes in accuracy and loss function on a validation set to evaluate the model's convergence and generalization ability.

[0062] The prediction process in the second embodiment is as follows: Figure 5 As shown.

[0063] Step S12: Input the measured parameter values ​​obtained on site; Step S22: Calculate the membership degree of the measured parameters in the field; Step S32: Convert the membership degrees into fuzzy feature vectors and input them into the trained deep neural network model to predict the probability distribution of the inscribeability level; Step S42: Output the registrability level corresponding to the highest probability as the second registrability evaluation level.

[0064] For example, the measured parameters obtained on-site are: drilling fluid consumption Q = 12m 3 / h, natural gamma value γ=22API, non-limestone fragment content C=7%, water pressure test injection volume W=18m 3 / h.

[0065] First, based on the membership calculation method and fuzzy feature extraction method described above, these four parameters are converted into a set of 12-dimensional fuzzy feature vectors [0.6,0.4,0,0.3,0.7,0,0,0.6,0.4,0,0.4,0.6]. Then, the fuzzy feature vectors are input into the trained neural network model to obtain the probability distribution of the illustrative level, P(D1)=0.12, P(D2)=0.71, P(D3)=0.17. The illustrative level D2 corresponding to the highest probability result is taken as the second illustrative evaluation level, that is, the illustrative evaluation level is medium illustrative.

[0066] Third embodiment: The third embodiment of this specification provides a method for simultaneously employing integrated fuzzy evaluation and deep neural network prediction to assess borehole injectability. The method flow is as follows: Figure 6 As shown.

[0067] Step S13: Input the measured values ​​of the parameters; Step S23: Based on the preset evaluation levels of each parameter, use the membership function to calculate the membership degree of the measured value to the corresponding parameter evaluation level; Step S331: Perform fuzzy comprehensive evaluation and output the first inscribeability evaluation level in step S331'; The first step is to calculate the weight vector of each parameter using the over-weighted method; then, the comprehensive membership vector is synthesized by weighted average; finally, the first inscribeability assessment level is determined based on the largest component.

[0068] Step S332: Predict using a deep neural network and output the second inscribeability evaluation level in step S332'; Steps S331 and S332 are performed in parallel, and their specific implementation can adopt the methods described in the first and second embodiments.

[0069] Step S43: Compare the first injectability assessment level and the second injectability assessment level, and output the injectability result.

[0070] If the injectability assessment levels are consistent, the injectability assessment level will be output; if they are inconsistent, the injectability level membership degree, and / or parameter weight value, and / or probability distribution of the injectability level, and / or first injectability assessment level, and / or second injectability assessment level will be output for reference by on-site engineering personnel.

[0071] The third embodiment constructs a dual-path evaluation model of fuzzy feature enhancement and neural network learning, which can output two results in parallel: fuzzy comprehensive evaluation and neural network prediction, supporting engineers to cross-validate and make comprehensive decisions.

[0072] Fourth embodiment: This embodiment provides a system for evaluating the grouting capability of boreholes based on a fuzzy dynamic model, used to implement the methods of the first, second, or third embodiments. Figure 7 As shown, the system includes: a data acquisition module, a fuzzy preprocessing module, a hybrid neural network module, and a result output module.

[0073] The data acquisition module is used to acquire measured values ​​of multiple field parameters, including but not limited to well fluid consumption, gamma logging values, cuttings logging data, and water injection volume in the pressure test. The fuzzy preprocessing module is used to calculate the membership degree of the measured value to the corresponding parameter evaluation level using a membership function based on the preset evaluation level of each parameter. The injectability assessment module is used to assess the confidence level of injectability using the membership of at least two measured parameters; The result output module is used to output the inscribeability assessment result based on the confidence level of the inscribeability level.

[0074] The inscribeability assessment module further includes: a fuzzy comprehensive evaluation module, used to calculate and obtain a first inscribeability assessment level based on the over-weighted method; and / or a deep neural network module, used to predict and obtain a second inscribeability assessment level based on the deep neural network.

[0075] The embodiments in this specification can be implemented entirely using the Python language.

[0076] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for evaluating grouting drillability based on a fuzzy dynamic model, characterized in that, The method comprises: obtaining measured values of multiple parameters, including but not limited to drilling fluid consumption, gamma logging values, cutting logging data, and water injection test injection volume; calculating the membership degree of the measured values to the corresponding parameter evaluation grade according to the preset evaluation grade of each parameter using a membership function; evaluating the injectability grade confidence level using the membership degrees of at least two measured parameters; evaluating the injectability of the grouting borehole according to the injectability grade confidence level.

2. The method of claim 1, wherein, The parameter evaluation grade is a first fuzzy grade, a second fuzzy grade, and a third fuzzy grade constructed by a first threshold value, a second threshold value, and a third threshold value.

3. The method of claim 2, wherein the step of determining the location of the mobile device is performed by the mobile device. The membership function is: ； ； ； wherein: r 1 （x） is a first fuzzy grade membership function, r 2 （x） is a second fuzzy grade membership function, r 3 （x） is a third fuzzy grade membership function, y is a measured value of a parameter, x 1 is a first threshold value, x 2 is a second threshold value, x 3 is a third threshold value.

4. The method of claim 3, wherein, The injectability grade confidence level is evaluated based on a comprehensive fuzzy evaluation method, which includes calculating parameter weights: First, calculate the exceeding ratio of each parameter, and the calculation formula is as follows: ； wherein: e i represents the over-standard ratio of the i-th parameter, is the optimal critical value of the i-th parameter in the corresponding evaluation grade; y i is the measured value of the i-th parameter; Then, calculate the weight of each parameter, and the calculation formula is as follows: ； wherein: W i is the weight of the i-th parameter, n is the number of parameters. Finally, construct a parameter weight set W = [W1, W2, …, Wn] composed of the weights of the parameters.

5. The method of claim 4, wherein, The optimal critical value is equal to the first threshold value.

6. The method of claim 4, wherein, The comprehensive fuzzy evaluation method further includes: Construct a fuzzy relation matrix R composed of the membership degrees of the measured parameters, and the dimension of R is n x 3; the parameter weight set with the fuzzy relation matrix a compound operation is performed to obtain the injectability level confidence, i.e. the injectability level membership U=[U1, U2, U3], and the calculation formula is: ; ； ； where: r jk is a matrix element, representing the membership of the jth factor to the kth class, r jk ∈ [0, 1]; Take the maximum value of [U1, U2, U3] as the first injectability evaluation grade.

7. The method of claim 3, wherein, The injectability grade confidence level is predicted based on a deep neural network, which includes a fuzzy attention layer, a wavelet transform layer, a residual convolution layer, and an output layer, and the prediction method specifically includes: Convert the membership degrees of the measured parameters into an n x 3-dimensional fuzzy feature vector, where n is the number of parameters; Input the fuzzy feature vector into the trained deep neural network for prediction; Obtain the injectability grade confidence level, i.e., the probability distribution of the three injectability grades; Output the injectability grade corresponding to the maximum probability as the second injectability evaluation grade.

8. The method of claim 7, wherein, The training method of the deep neural network includes: First, obtain historical measured data as training samples, and the historical measured data includes measured values of multiple parameters, and label the injectability grade of each sample; Then, perform fuzzy feature extraction, calculate the membership degrees of each measured value to different evaluation grades based on the preset evaluation grade of each parameter and the membership function, and combine the membership degrees of all parameters into an n x 3-dimensional fuzzy feature vector, where n is the number of parameters; Finally, input the n x 3-dimensional fuzzy feature vector into the deep neural network for training to obtain the trained deep neural network.

9. The method of any of claims 3 to 8, wherein, The injectability grade confidence level is evaluated based on comprehensive fuzzy evaluation and deep neural network, specifically including: Obtain the first injectability evaluation grade based on comprehensive fuzzy evaluation; Obtain the second injectability evaluation grade based on deep neural network prediction; Compare the first injectability evaluation grade and the second injectability evaluation grade, if they are consistent, output the injectability evaluation grade, if they are not consistent, output the injectability grade membership degree, and / or parameter weight value, and / or probability distribution of the injectability grade, and / or first injectability evaluation grade, and / or second injectability evaluation grade.

10. A system for evaluating grouting borehole injectability based on fuzzy dynamic model, characterized in that, ​ The data acquisition module is configured to acquire measured values of multiple field parameters, including but not limited to well fluid consumption, gamma logging values, cutting logging data, and water injection volume in a water injection test. The fuzzy preprocessing module is configured to calculate the membership degree of the measured values to the evaluation grade of the corresponding parameter according to the evaluation grade of each parameter. The injectability evaluation module is configured to evaluate the injectability grade confidence level by using the membership degrees of at least two measured parameters. The result output module is configured to output the injectability evaluation result according to the injectability grade confidence level. The injectability evaluation module further includes: The fuzzy comprehensive evaluation module is configured to obtain a first injectability evaluation grade based on fuzzy comprehensive evaluation; and / or The deep neural network module is configured to obtain a second injectability evaluation grade based on deep neural network prediction.

Images