A co2 storage area intelligent monitoring method based on multi-source data joint

Abstract

The application discloses a kind of CO2 storage area intelligent monitoring methods based on multi-source data joint, utilize hyperspectral image macroscopically monitored to obtain the hyperspectral image of CO2 storage area, extract the spectral index with CO2 sensitivity and high correlation;Through high-density electrical method, the underground resistivity distribution of CO2 storage area is inverted, then the carbonate content in the soil of different positions is determined, the relationship between carbonate content and spectral index is determined respectively to establish spectral-CO2 storage correlation model, and the relationship between carbonate content and resistivity is determined to establish resistivity-CO2 storage correlation model;Finally, a variety of parameter data are fused to construct random forest joint monitoring model, realize CO2 storage area leakage prediction.The method breaks through the limitation of single data dimension by spectral-electricity characteristic coupling, realizes the monitoring of CO2 storage area from surface to deep whole area storage state, guarantees the long-term stability of CO2 storage area, and provides technical support for the safe implementation of coal mine carbon waste storage project.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of underground carbon sequestration monitoring, and particularly relates to a CO2 sequestration region intelligent monitoring method based on multi-source data jointing. BACKGROUND

[0002] During the whole period from CO2 injection to long-term stable sequestration in the underground space of a coal mine, there are many uncertainties and potential risks. If the sealing is insufficient, acidic CO2 fluid may penetrate and pollute groundwater or damage surface vegetation, causing irreversible harm to the environment. Therefore, it is particularly important to carry out accurate monitoring of the carbon sequestration site and master its sequestration state and migration trend underground. However, traditional single means cannot achieve large-scale and efficient monitoring, and it is difficult to timely warn of possible CO2 leakage risks.

[0003] Remote sensing image technology, as an important method for identifying the state of surface vegetation and soil, can achieve large-area synchronous observation and quickly obtain macro data based on the principle that CO2 leakage will change the spectral characteristics of soil and vegetation. However, this technology mainly focuses on the surface and near-surface, and cannot reach the deep underground. In addition, the vegetation, buildings, and water bodies on the surface seriously interfere with the remote sensing signal, and their spectral characteristics are confused with each other, greatly reducing the accuracy of identification and inversion.

[0004] High-density electrical method in geophysical exploration can directly obtain underground electrical characteristics by inputting electric current into the underground and measuring resistivity information at different depths. Since the change of soil CO2 content will reduce the regional resistivity, the high-density electrical method can be used to monitor whether CO2 is leaked. Compared with remote sensing images, high-density electrical method can clearly show the vertical distribution state of CO2 sequestration region in depth detection, but the electrical exploration signal is easily disturbed by surface environmental factors such as terrain undulation, soil moisture change, and surface metal objects. These interference factors will cause noise and error in the collected data, seriously affecting the accuracy and resolution of the electrical exploration results. Therefore, it is difficult to accurately define the CO2 sequestration state and range by relying on this method alone.

[0005] Therefore, how to provide a new monitoring method that can monitor the sequestration state of CO2 sequestration region from the surface to the deep region, accurately judge whether CO2 leakage occurs, and ensure the long-term stability of CO2 sequestration region, thereby providing technical support for the safe implementation of coal mine carbon waste sequestration project, is the research direction of the present application. SUMMARY

[0006] In view of the problems in the prior art, the application provides a CO2 storage area intelligent monitoring method based on multi-source data jointing, which can realize monitoring of the whole storage state of a CO2 storage area from the ground to the deep part, can accurately determine whether CO2 leakage occurs, ensures long-term stability of the CO2 storage area, and provides technical support for safe implementation of a coal mine carbon waste storage project.

[0007] In order to achieve the above-mentioned purpose, the technical scheme adopted by the application is as follows: a CO2 storage area intelligent monitoring method based on multi-source data jointing, comprising the following steps:

[0008] Step one, obtaining CO2 storage area remote sensing data: determining a CO2 storage area and obtaining hyperspectral images and digital elevation model data of the ground surface of the area;

[0009] Step two, obtaining real spectrum information: after preprocessing the hyperspectral images obtained in step one, obtaining real spectrum information of the CO2 storage area;

[0010] Step three, remote sensing data processing: after processing the real spectrum information processed in step two by using a remote sensing image processing software, extracting vegetation index and texture features of the ground surface of the CO2 storage area, and then using a band operation tool to screen and calculate spectral indices with high CO2 sensitivity and correlation;

[0011] Step four, determining the resistivity distribution of the CO2 storage area: determining the layout position of the high-density electrical method side line of the detection area according to the real spectrum information and the digital elevation model data, then collecting data of different depths of the CO2 storage area, the resistivity of the CO2 accumulation area will be reduced, and the resistivity distribution of different positions of the CO2 storage area is determined by regularized inversion of the measurement data;

[0012] Step five, measuring the carbonate content of the CO2 storage area: according to the real spectrum information and the resistivity distribution determined in step four, collecting soil samples at different positions of the CO2 storage area, and measuring the carbonate content in the soil at different positions;

[0013] Step six, constructing a spectrum-CO2 storage correlation model: the carbonate content at different positions in step five is one-to-one corresponding to the spectral index at the corresponding position, the relationship between the carbonate content in the soil and the spectral index is established by using a partial least squares method regression, so as to construct a spectrum-CO2 storage correlation model;

[0014] Step seven, constructing a resistivity-CO2 storage correlation model: the carbonate content at different positions in step five is one-to-one corresponding to the resistivity at the corresponding position, the change relationship between the carbonate content and the resistivity in the CO2 storage area is determined, so as to construct a resistivity-CO2 storage correlation model;

[0015] Step eight, CO2 storage area continuous monitoring: obtaining the carbonate content and spectral index distribution of all positions in the CO2 storage area according to the model constructed in step six, and obtaining the carbonate content and resistivity distribution of all positions in the CO2 storage area according to the model constructed in step seven, so as to obtain the corresponding carbonate content, spectral index and resistivity of each position; constructing a random forest joint inversion model, taking part of the data of all positions as a training set and the other part as a validation set, training the model, and then using the model to realize the monitoring of the CO2 migration in the CO2 storage area.

[0016] Further, the preprocessing of step two includes geometric correction, radiation correction and atmospheric correction.

[0017] Further, the spectral index of step three with high CO2 sensitivity and correlation includes normalized vegetation index (NDVI), vegetation stress index.

[0018] Further, in step four, the regularization inversion determines the resistivity distribution of different positions in the CO2 storage area, and the specific formula is:

[0019] Φ=‖W d (d obs -d pred )| 2 +α||W m m| 2

[0020] Where d obs is the observation data, d pred is the prediction data, W d is the weight matrix, W m is the model weight matrix, m is the model parameter vector, and α is the regularization parameter.

[0021] Further, in step five, the carbonate content W in the soil at different positions is determined, and the specific formula is:

[0022]

[0023] Where C x is the sample determination concentration, V is the constant volume after digestion, and m is the mass of the soil sample.

[0024] Further, the spectral-CO2 storage correlation model constructed in step six is specifically:

[0025] The partial least squares regression (PLSR) is used to establish the relationship between the carbonate content y in the soil and the spectral index x1, x2, x3, x4, … x p , and the soil sample is n. The independent variable matrix X=(x ij) nxp and the independent variable y, which is standardized to obtain X * and y * Then, the principal components t1, t2…, t are extracted by iterative calculation m , and a regression equation is established:

[0026]

[0027] The predicted value y* of the soil carbonate content is denormalized to obtain the predicted value Finally, the spectrum-CO2 storage correlation model is obtained.

[0028] Further, the resistivity-CO2 storage correlation model constructed in step seven is specifically:

[0029] The linear regression method is used to determine the change relationship between the carbonate content and the resistivity in the CO2 storage area, wherein the soil carbonate content is y, and the corresponding resistivity obtained by the regularization inversion is Z. The linear regression equation is obtained by analyzing a plurality of sampling points:

[0030] y=AZ+B

[0031] Wherein A and B are regression coefficients, the regression coefficients A and B are solved by taking the derivative of the difference between the predicted value y* and the observed value y to be zero, and finally the resistivity-CO2 storage correlation model is obtained.

[0032] Further, in step eight, 90% of the data is taken as the training set, and 10% of the sample data is taken as the validation set, and the model is trained with the spectral index and the resistivity as the input and the carbonate content as the output. After training, the carbonate content threshold is set, and in the subsequent CO2 storage area monitoring process, the spectral index and the resistivity of different positions are collected in real time, the model is input, the model outputs the predicted value of the carbonate content at the corresponding position, the predicted value is compared with the set threshold, all positions with predicted values exceeding the threshold are marked, thereby the CO2 migration range is circled, and the continuous monitoring of the CO2 storage area is realized.

[0033] Compared with the prior art, the present application proposes a combination of multi-source data fusion and deep learning, utilizes the advantages of macroscopic monitoring of hyperspectral images and underground detection of high-density electrical methods, realizes stereoscopic monitoring from the ground to a certain depth underground, respectively obtains hyperspectral images of the CO2 storage area, extracts spectral indices with high sensitivity and correlation to CO2, including surface vegetation index, soil spectral characteristics and other information; through high-density electrical method inversion, the distribution of CO2 storage area underground resistivity and CO2 plume migration path are obtained, then the carbonate content in the soil at different positions is determined, the relationship between carbonate content and spectral index is determined respectively to establish a spectral-CO2 storage correlation model, and the relationship between carbonate content and resistivity is determined to establish a resistivity-CO2 storage correlation model; finally, remote sensing spectral characteristics, resistivity parameters and terrain data are fused to construct a random forest joint monitoring model, realizing CO2 storage space leakage prediction. This method breaks through the limitation of single data dimension by coupling spectral-electricity characteristics, establishes a model by combining machine learning algorithm to improve the identification accuracy of storage boundary, and preliminarily screens with the help of remote sensing to reduce the amount of field sampling, and combines high-density electrical method for targeted detection, which can greatly reduce the cost and shorten the detection period, and quickly provide basis for decision-making. Therefore, the present application can realize monitoring of the whole region of the CO2 storage area from the ground to the deep region, and can accurately judge whether CO2 leakage occurs, ensuring the long-term stability of the CO2 storage area, and providing technical support for the safe implementation of the coal mine carbon waste storage project. BRIEF DESCRIPTION OF DRAWINGS

[0034] Figure 1 is the overall flowchart of the present application;

[0035] Figure 2 is the resistivity distribution map of the CO2 storage area in the embodiment of the present application. DETAILED DESCRIPTION

[0036] The present application will be further described below.

[0037] As shown in Figure 1 , the present application comprises the following steps:

[0038] Step one, obtaining remote sensing data of the CO2 storage area: determining the CO2 storage area and obtaining the hyperspectral of the high-resolution 5 satellite (GF-5) of the ground surface of the area and the digital elevation model data (DEM data) of the area in the China Resource Satellite Application Center;

[0039] Step two, obtaining real spectral information: after geometric correction, radiation correction and atmospheric correction of the hyperspectral image obtained in step one, obtaining the real spectral information of the CO2 storage area;

[0040] Step three, remote sensing data processing: after processing the real spectral information processed in step two by remote sensing image processing software, the vegetation index and texture features of the surface of the CO2 storage area are extracted, and then the spectral index with high sensitivity and correlation to CO2 is calculated by using the band operation tool, including normalized difference vegetation index (NDVI) and vegetation stress index;

[0041] Step four, determining the resistivity distribution of the CO2 storage area: according to the real spectral information and digital elevation model data, the layout position of the high-density electrical method side line of the detection area is determined, and then the data of different depth resistivity of the CO2 storage area is collected. The resistivity of the CO2 accumulation area will be reduced. The resistivity distribution of the CO2 storage area at different positions is determined by regularized inversion of the measured data, as shown in Figure 2 The specific formula is:

[0042] Φ=||W d (d obs -d pred )| 2 +α||W m m| 2

[0043] Where d obs is the observation data, d pred is the prediction data, W d is the weighting matrix, W m is the model weighting matrix, m is the model parameter vector, and a is the regularization parameter.

[0044] Step five, determining the carbonate content of the CO2 storage area: according to the real spectral information and the resistivity distribution determined in step four, soil samples are collected at different positions of the CO2 storage area, and the carbonate content W in the soil at different positions is measured in the laboratory by atomic absorption spectrometry (AAS). The specific formula is:

[0045]

[0046] Where C x is the measured concentration of the sample, V is the constant volume after digestion, and m is the mass of the soil sample.

[0047] Step six, building a spectrum-CO2 storage correlation model: through the one-to-one correspondence between the carbonate content at different positions in step five and the spectral index at the corresponding position, the relationship between the carbonate content in the soil and the spectral index is established by using partial least squares regression, thereby building a spectrum-CO2 storage correlation model, which is specifically:

[0048] The partial least squares regression (PLSR) is used to establish the relationship between the carbonate content y in the soil and the spectral index x1, x2, x3, x4, … x pThe relationship between the soil samples is n, and the independent variable matrix X=(x ij ) nxp and the dependent variable y is obtained by standardizing X * and y * Then, the principal components t1, t2…, t m are extracted by iterative calculation, and the regression equation is established.

[0049]

[0050] The predicted value of soil carbonate content y* is obtained by denormalizing the predicted value Finally, the spectrum-CO2 storage correlation model is obtained.

[0051] Step seven, build the resistivity-CO2 storage correlation model: through the carbonate content of step five in different positions and the corresponding resistivity one by one, the change relationship between the carbonate content and the resistivity in the CO2 storage area is determined, and the resistivity-CO2 storage correlation model is built, which is:

[0052] The linear regression method is used to determine the change relationship between the carbonate content and the resistivity in the CO2 storage area, wherein the soil carbonate content is y, and the corresponding resistivity obtained by regularization inversion is Z. The linear regression equation is obtained by analyzing a plurality of sampling points:

[0053] y=AZ+B

[0054] Wherein A, B is the regression coefficient, the difference between the predicted value y* and the observed value y is zero, and the regression coefficients A and B are solved, and finally the resistivity-CO2 storage correlation model is obtained.

[0055] Step eight, CO2 storage area continuous monitoring: obtain the carbonate content and spectral index distribution of all positions in the CO2 storage area according to the model constructed in step six, and obtain the carbonate content and resistivity distribution of all positions in the CO2 storage area according to the model constructed in step seven, so as to obtain the corresponding carbonate content, spectral index and resistivity of each position; construct a random forest joint inversion model, take 90% of the data as the training set and 10% of the sample data as the validation set, and take the spectral index and resistivity as the input and the carbonate content as the output to train the model, set the carbonate content threshold after training, and in the subsequent monitoring process of the CO2 storage area, input the spectral index and resistivity of different positions into the model in real time, the model outputs the predicted value of the carbonate content of the corresponding position, compares the predicted value with the set threshold, and marks all positions whose predicted value exceeds the threshold, so as to delineate the CO2 migration range and realize the continuous monitoring of the CO2 storage area; in addition, in the CO2 migration range, it is further determined whether there is a region with resistivity <100Ω·m, if there is, it is judged as a CO2 high pollution area, so that corresponding measures can be taken in time.

[0056] The above only describes the preferred embodiments of the present application, and it should be pointed out that for ordinary skilled persons in the art, several improvements and refinements can be made without departing from the principles of the present application, and these improvements and refinements should also be considered within the protection scope of the present application.

Claims

1. A method for intelligent monitoring of CO2 sequestration areas based on multi-source data fusion, characterized in that, Includes the following steps: Step 1: Obtain remote sensing data of the CO2 sequestration area: Determine the CO2 sequestration area and obtain hyperspectral images and digital elevation model data of the surface of the area; Step 2: Obtain the true spectral information: After preprocessing the hyperspectral image obtained in Step 1, obtain the true spectral information of the CO2 sealing area; Step 3: Remote sensing data processing: After processing the real spectral information obtained in Step 2 using remote sensing image processing software, the vegetation index and texture features of the surface of the CO2 sequestration area are extracted. Then, the spectral index with high sensitivity and correlation to CO2 is selected and calculated using band calculation tools. Step 4: Determine the resistivity distribution of the CO2 storage area: Based on the actual spectral information and digital elevation model data, determine the layout location of the high-density electrical resistivity spectroscopy (EDS) transect in the detection area. Then, collect resistivity data at different depths in the CO2 storage area. Perform regularized inversion on the measured data to determine the resistivity distribution at different locations in the CO2 storage area. The specific formula is as follows: ； in For observation data, For predictive data, For weighted matrices, The weighting matrix for the model, For model parameter vectors, For regularization parameters; Step 5: Determine the carbonate content in the CO2 sequestration area: Based on the actual spectral information and the resistivity distribution determined in Step 4, soil samples were collected from different locations in the CO2 sequestration area, and the carbonate content in the soil at different locations was determined. Step Six: Constructing a Spectral-CO2 Sequestration Correlation Model: By establishing a one-to-one correspondence between the carbonate content at different locations and the corresponding spectral indices from Step Five, partial least squares regression is used to establish the relationship between the carbonate content and spectral indices in the soil, thereby constructing a spectral-CO2 sequestration correlation model. Specifically: Partial least squares regression was used to establish the relationship between soil carbonate content y and spectral indices x1, x2, x3, x4, ... x p Given n soil samples, generate an independent variable matrix X = (x... ij ) n×p And the dependent variable y, which is then standardized to obtain X. * and Then, principal components t1, t2, ..., t are extracted through iterative calculation. m Establish the regression equation: ； Predicted values ​​of soil carbonate content The predicted value is obtained after destandardization. Finally, a correlation model between spectral density and CO2 sequestration was obtained; Step 7: Construct a resistivity-CO2 storage correlation model: By matching the carbonate content at different locations in Step 5 with the resistivity at the corresponding locations, determine the relationship between the carbonate content and resistivity in the CO2 storage area, thereby constructing a resistivity-CO2 storage correlation model. Step 8: Continuous Monitoring of CO2 Sequestration Area: Based on the model constructed in Step 6, obtain the carbonate content and spectral index distribution of all locations in the CO2 sequestration area, and based on the model constructed in Step 7, obtain the carbonate content and resistivity distribution of all locations in the CO2 sequestration area, thereby obtaining the carbonate content, spectral index, and resistivity corresponding to each location; construct a random forest joint inversion model, using a portion of the data from all locations as the training set and another portion as the validation set, train the model, and then use the model to monitor CO2 migration within the CO2 sequestration area.

2. The intelligent monitoring method for CO2 storage areas based on multi-source data fusion according to claim 1, characterized in that, The preprocessing in step two includes geometric correction, radiometric correction, and atmospheric correction.

3. The intelligent monitoring method for CO2 storage areas based on multi-source data fusion according to claim 1, characterized in that, The spectral indices in step three that are highly sensitive to and correlated with CO2 include the Normalized Difference Vegetation Index (NDVI) and the Vegetation Stress Index (WSI).

4. The intelligent monitoring method for CO2 storage areas based on multi-source data fusion according to claim 1, characterized in that, In step five, the carbonate content W in the soil at different locations is determined using the following formula: ； in To determine the concentration of the sample, To determine the constant volume after digestion, For soil sample quality.

5. The intelligent monitoring method for CO2 sequestration areas based on multi-source data fusion according to claim 1, characterized in that, The specific steps for constructing the resistivity-CO2 storage correlation model in step seven are as follows: The relationship between carbonate content and resistivity in CO2 sequestration areas was determined using linear regression, where soil carbonate content is y and the corresponding resistivity obtained through regularized inversion is Z. The linear regression equation was obtained through analysis of multiple sampling points. ； Where A and B are regression coefficients, obtained through predicted values. The regression coefficients A and B are solved by taking the derivative of the difference square with the observed value y until it is zero, and finally the resistivity-CO2 storage correlation model is obtained.

6. The intelligent monitoring method for CO2 sequestration areas based on multi-source data fusion according to claim 1, characterized in that, In step eight, 90% of the data is used as the training set and 10% of the sample data is used as the validation set. The model is trained using spectral index and resistivity as inputs and carbonate content as output. After training, a carbonate content threshold is set. During subsequent monitoring of the CO2 storage area, spectral index and resistivity at different locations are collected in real time and input into the model. The model outputs the predicted carbonate content at the corresponding location. The predicted value is compared with the set threshold, and all locations where the predicted value exceeds the threshold are calibrated, thereby delineating the CO2 migration range and realizing continuous monitoring of the CO2 storage area.

Images