Search

CN-121994729-A - Oilfield methane leakage detection method using satellite remote sensing data

CN121994729ACN 121994729 ACN121994729 ACN 121994729ACN-121994729-A

Abstract

The invention discloses an oilfield methane leakage detection method utilizing satellite remote sensing data, which belongs to the field of atmospheric remote sensing and energy production, and adopts the technical scheme that the oilfield methane leakage detection method utilizing the satellite remote sensing data comprises the steps of screening out a methane absorption spectrum data range as a simulated spectrum; the method comprises the steps of obtaining hyperspectral image data in an analog spectrum range, preprocessing the hyperspectral image data to be used as an observation spectrum, simulating a spectrum absorption curve of gas in a real-time state of an oil field area to find an ideal reference spectrum, carrying out logarithmic difference processing on the screened analog spectrum and the reference spectrum to obtain priori spectrum data, inputting the observation spectrum and the priori spectrum into a matched filter, inverting to obtain an XCH4 enhanced image, and identifying a plume area in the XCH4 enhanced image by using a convolutional neural network. The invention has the beneficial effect of providing an oilfield methane leakage detection method by utilizing satellite remote sensing data.

Inventors

  • HE HU
  • SUN JIANCHENG
  • ZHAO JINGANG
  • SUN ENCHENG
  • LI HAORAN
  • ZHANG QIONG
  • WANG QIYUN
  • LIU XIAO
  • LI YUAN

Assignees

  • 中国石油化工股份有限公司
  • 中国石油化工股份有限公司胜利油田分公司

Dates

Publication Date
20260508
Application Date
20241108

Claims (10)

  1. 1. A method for detecting methane leakage in oil field by using satellite remote sensing data is characterized in that, Screening out the data range of the methane absorption spectrum as an analog spectrum; acquiring hyperspectral image data in an analog spectrum range, and preprocessing the hyperspectral image data to be used as an observation spectrum; simulating a spectrum absorption curve of gas in a real-time state in an oilfield area, and finding an ideal reference spectral line; Carrying out logarithmic difference processing on the screened analog spectrum and the reference spectrum to obtain priori spectrum data; inputting the observation spectrum and the prior spectrum into a matched filter, and inverting to obtain an XCH4 enhanced image; and identifying a plume region in the XCH4 enhanced image by using a convolutional neural network.
  2. 2. The method for detecting methane leakage in oil field according to claim 1, wherein the data range of methane absorption spectrum is screened out as an analog spectrum, specifically: and selecting an absorption spectrum range with the smallest error from the HITRAN database by utilizing the Monte Carlo simulation principle as a search window, and screening out the optimal combination of spectrum bands to be used as a simulation spectrum.
  3. 3. The oilfield methane leak detection method of claim 2, wherein the preprocessing of the hyperspectral image data comprises: Acquiring hyperspectral image data in an optimal combination range; According to the information obtained from the satellite header file, carrying out radiation calibration and atmosphere correction on the hyperspectral image data; And carrying out cloud layer removal processing on the hyperspectral image data by utilizing a cloud mask, and giving a value of 0 to the data containing cloud layers.
  4. 4. The oilfield methane leak detection method of claim 2, wherein finding an ideal reference spectral line comprises: Calculating the center wavelength and the full width at half maximum of the optimal combination range; Obtaining a unit gas absorption spectrum under the resolution of 0.1cm < -1 > by using an HITRAN molecular absorption library; and simulating a spectral response function of the hyperspectral remote sensor wave band by using a Gaussian function according to the full width of the half maximum, and obtaining a reference spectral line for stably recording the light intensity in the oilfield state.
  5. 5. The oilfield methane leak detection method according to claim 1, wherein the matched filter is an iterative lognormal matched filter, and specifically comprises:
  6. 6. the oilfield methane leak detection method of claim 1, wherein identifying plume regions in XCH4 enhanced images using convolutional neural networks comprises: constructing a data set of a convolutional neural network for identifying a plume region in an XCH4 enhanced image; constructing a convolutional neural network for identifying a plume region in the XCH4 enhanced image; Training a convolutional neural network; performing PCA processing on the data set by using the XCH4 enhanced image obtained by inversion; and inputting the data processed by the PCA into a convolutional neural network for recognition.
  7. 7. The oilfield methane leak detection method of claim 1, wherein PCA processing the inverted XCH4 enhanced image on the dataset comprises: selecting a principal component singular vector which can describe the scene spectrum change most; Determining the number of the first three optimal singular vectors; These vectors are connected to methane jacobian to construct a matrix J of dimensions 4 x PRISMA bands, which is used with the logarithm y of the measured radiation intensity to find a vector W in a linear least squares fit that minimizes the cost function for each pixel.
  8. 8. The oilfield methane leak detection method of claim 1, wherein training the convolutional neural network comprises: An encoder that captures context in the hyperspectral image, consisting of a convolutional layer and a max-pooling layer; A decoder, which locates the features captured by the encoder, is composed of a convolutional layer and an upsampling layer.
  9. 9. A computer device comprising a processor and a memory, the processor configured to execute a program for oilfield methane leak detection stored in the memory, to implement the oilfield methane leak detection method of any one of claims 1-7.
  10. 10. A storage medium storing one or more programs executable by one or more processors to implement the oilfield methane leak detection method of any one of claims 1-7.

Description

Oilfield methane leakage detection method using satellite remote sensing data Technical Field The invention belongs to the field of atmospheric remote sensing and energy production, and particularly relates to an oilfield methane leakage detection method utilizing satellite remote sensing data. Background Since the industrial age, the concentration of methane (CH 4) in the atmosphere has increased by more than 150%, and by the beginning of 2023, the global average concentration has reached 1920ppb. CH4 is a key factor for artificial greenhouse effect, and has a non-negligible effect on global warming. It is estimated that nearly one third of the global warming effect has been contributed to so far, with emissions of one quarter of the total greenhouse gases. The warming effect of a single CH4 molecule on the 20 year scale is equivalent to 80 times that of carbon dioxide. Thus, development of methane abatement is critical for rapid mitigation of climate change. Oil and gas development is one of the important sources of global energy supply, with its associated CH4 leakage being the most important source of methane emissions. The relevant statistics indicate that the relevant CH4 emissions for oil and gas exploitation, processing and transportation account for more than 30% of the total CH4 emissions. Most of the oil gas related CH4 emission belongs to leakage, and the emission reduction cost is lower than the economic value generated after collection. Therefore, the emission reduction of the oil gas methane has urgent nature in science and high feasibility in the practical operation level. The key point of emission reduction of the oil gas CH4 is to determine leakage points, so the application provides an oilfield methane leakage detection method utilizing satellite remote sensing data. Disclosure of Invention The invention aims to provide an oilfield methane leakage detection method utilizing satellite remote sensing data. In a first aspect of an embodiment of the present application, there is provided an oilfield methane leak detection method using satellite remote sensing data, characterized in that, Screening out the data range of the methane absorption spectrum as an analog spectrum; acquiring hyperspectral image data in an analog spectrum range, and preprocessing the hyperspectral image data to be used as an observation spectrum; simulating a spectrum absorption curve of gas in a real-time state in an oilfield area, and finding an ideal reference spectral line; Carrying out logarithmic difference processing on the screened analog spectrum and the reference spectrum to obtain priori spectrum data; inputting the observation spectrum and the prior spectrum into a matched filter, and inverting to obtain an XCH4 enhanced image; and identifying a plume region in the XCH4 enhanced image by using a convolutional neural network. Further, the data range of the methane absorption spectrum is screened out and is taken as a simulated spectrum, specifically: and selecting an absorption spectrum range with the smallest error from the HITRAN database by utilizing the Monte Carlo simulation principle as a search window, and screening out the optimal combination of spectrum bands to be used as a simulation spectrum. Further, preprocessing the hyperspectral image data includes: Acquiring hyperspectral image data in an optimal combination range; According to the information obtained from the satellite header file, carrying out radiation calibration and atmosphere correction on the hyperspectral image data; And carrying out cloud layer removal processing on the hyperspectral image data by utilizing a cloud mask, and giving a value of 0 to the data containing cloud layers. Further, finding an ideal reference spectral line specifically comprises: Calculating the center wavelength and the full width at half maximum of the optimal combination range; Obtaining a unit gas absorption spectrum under the resolution of 0.1cm < -1 > by using an HITRAN molecular absorption library; and simulating a spectral response function of the hyperspectral remote sensor wave band by using a Gaussian function according to the full width of the half maximum, and obtaining a reference spectral line for stably recording the light intensity in the oilfield state. Further, the matched filter is an iterative lognormal matched filter, which specifically comprises: Further, identifying a plume region in the XCH4 enhanced image using a convolutional neural network, comprising: constructing a data set of a convolutional neural network for identifying a plume region in an XCH4 enhanced image; constructing a convolutional neural network for identifying a plume region in the XCH4 enhanced image; Training a convolutional neural network; performing PCA processing on the data set by using the XCH4 enhanced image obtained by inversion; and inputting the data processed by the PCA into a convolutional neural network for recognition. Further, performing PCA processi