CN-121765693-B - Construction method of multi-seam combined production coal-seam gas well productivity split model

CN121765693BCN 121765693 BCN121765693 BCN 121765693BCN-121765693-B

Abstract

The invention relates to the technical field of coal mining, in particular to a method for constructing a split model of the productivity of a multi-seam coal-seam gas well. According to the method, historical combined production data and historical production capacity of each similar historical coal-bed gas well similar to the static data of the reservoir are screened, a capacity split preliminary model of the target coal-bed gas well is established, the predicted production capacity of the target coal-bed gas well in each unit period is output through a real-time combined production data sequence of each coal-bed in each unit period, the capacity split preliminary model is adjusted based on the deviation value of the predicted production capacity and the actual production capacity to obtain a capacity split optimization model, incremental training and iterative adjustment of the model are achieved, the predicted production capacity of each coal-bed in future set time period is predicted, the main coal-bed of the target coal-bed gas well in future set time period is identified, the capacity potential of the main coal-bed is helped to be exerted, the overall production efficiency of multi-coal laminated production is improved, and invalid production cost is reduced.

Inventors

YAN ZHIHUA
Yan Lichong
YANG LEI
HU HAIYANG
CHEN JIE
GAO WEI
LIANG LONGJUN
ZHU JIAWEI
TAO JINJIN
FENG YUNFEI
YI WANG

Assignees

贵州省煤层气页岩气工程技术研究中心
贵州优驰能源科技有限责任公司

Dates

Publication Date: 20260508
Application Date: 20260304

Claims (9)

1. The method for constructing the split model for the production capacity of the multi-seam combined production coal-seam gas well is characterized by comprising the following steps of: Acquiring reservoir static data of each coal bed in a target coal bed gas well, screening historical combined production data and historical production capacity of each similar historical coal bed gas well similar to the reservoir static data, and establishing a capacity split preliminary model of the target coal bed gas well; collecting real-time co-production data of each coal bed in a target coal bed gas well, and generating a real-time co-production data sequence of each coal bed in each unit time period, wherein the co-production data comprises fluid flow pressure, fluid flow rate, fluid density and temperature; The method comprises the steps of inputting a real-time combined production data sequence into an established capacity split preliminary model, outputting the predicted production capacity of a target coal-bed gas well in each unit period, and calculating the deviation value of the predicted production capacity and the actual production capacity; Analyzing whether the productivity split preliminary model needs to be optimized or not based on the deviation value, and adjusting the productivity split preliminary model when the productivity split preliminary model needs to be optimized to obtain an productivity split optimization model; Analyzing whether the productivity split preliminary model needs to be optimized or not comprises setting a deviation value threshold value based on the historical deviation value of each unit time period of each similar historical coal-bed gas well in the corresponding drainage phase; When the productivity split preliminary model is optimized, screening real-time combined mining data sequences and predicting production productivity of each coal seam in each unit time period with the deviation value smaller than the set deviation value threshold value, supplementing the real-time combined mining data sequences and predicting production productivity to a data set corresponding to each coal seam in the current mining stage, performing incremental training on the productivity split preliminary model of each coal seam in the current mining stage, and updating the productivity split preliminary model; And predicting the predicted production capacity of the future set time length of each coal bed by combining the capacity split optimization model according to the real-time combined production data sequence of each coal bed in each unit time period, and identifying and displaying the main coal bed of the target coal bed gas well in the future set time length.
2. The method for constructing the split model for the production capacity of the multi-seam co-production coal-bed gas well, which is characterized in that the historical co-production data and the historical production capacity of each similar historical coal-bed gas well which are similar to the static data of the reservoir are screened, wherein the content is as follows: carrying out standardized processing on reservoir static data of each coal bed in the target coal bed gas well to form reservoir static data vectors of each coal bed; The method comprises the steps of calling historical coal-bed gas wells with the same number as that of coal beds from a coal-bed exploitation database, and obtaining historical reservoir static data vectors of the coal beds in the historical coal-bed gas wells; combining the static data vector of each coal bed in the target coal bed gas well, analyzing the comprehensive similarity of the target coal bed gas well and each historical coal bed gas well through the similarity, and taking the historical coal bed gas well with the comprehensive similarity being greater than a preset similarity threshold as a similar historical coal bed gas well; And screening historical co-production data and historical production capacity of each similar historical coal-bed gas well from the coal-bed production database.
3. The method for constructing the split capacity model of the multi-seam gas well is characterized by comprising the following steps of: Screening corresponding associated coal beds of each coal bed in the target coal bed gas well in each similar historical coal bed gas well according to the reservoir static data of each coal bed in the target coal bed gas well; Acquiring historical combined production data and historical production capacity of each coal bed in each similar historical coal bed gas well corresponding to the associated coal bed, and extracting historical combined production data time sequence and historical production capacity time sequence of the associated coal bed in each drainage stage; Performing time dimension matching on the historical combined mining data time sequence and the historical production energy time sequence to form a data set of corresponding combined mining data and production energy of each coal seam in each mining stage; Dividing the data set into a training set and a testing set according to a set proportion, and carrying out model training by taking the data collected in the training set as input characteristics and the production capacity as output characteristics to obtain an initial capacity split preliminary model; Predicting the production capacity of the test set through the initial capacity split preliminary model, optimizing the initial capacity split preliminary model by adopting accuracy, and outputting the optimized capacity split preliminary model; and counting the productivity split preliminary models of each coal bed in each drainage stage to form the productivity split preliminary model of the target coal bed gas well.
4. The method for constructing the split model for the capacity of the multi-seam combined production coal-bed gas well, which is characterized in that the real-time combined production data sequence of each coal seam in each unit time period is generated by the following steps: Collecting real-time combined production data of each coal bed through a monitoring device deployed in each coal bed in a target coal bed gas well; Performing time stamp alignment on the real-time combined mining data to obtain the real-time combined mining data of each unit time period in the current set time length; And carrying out outlier rejection and interpolation processing on the real-time combined mining data of each unit time period to obtain processed real-time combined mining data, and generating a real-time combined mining data sequence of each coal seam in each unit time period.
5. The method for constructing the split model for the productivity of the multi-seam gas production well is characterized in that the output target coal-seam gas well predicts the production productivity in each unit period, and the method comprises the following steps: screening the productivity split preliminary model of each coal bed in the current drainage and production stage from the productivity split preliminary model of the target coal bed gas well according to the current drainage and production stage corresponding to the target coal bed gas well; inputting real-time combined mining data sequences of all coal beds in each unit time period into a corresponding productivity split preliminary model, and outputting predicted production capacity of all coal beds in each unit time period; And carrying out accumulated calculation on the predicted production capacity of each coal bed in each unit time period to obtain the predicted production capacity of the target coal bed gas well in each unit time period.
6. The method for constructing the capacity split model of the multi-seam gas production well according to claim 3, wherein the method for analyzing whether the capacity split preliminary model needs to be optimized further comprises the following steps: According to the current drainage and production stage corresponding to the target coal-bed gas well, extracting the historical predicted production capacity and the historical actual production capacity of each unit period of each similar historical coal-bed gas well in the corresponding drainage and production stage from a coal-bed mining database, and comparing to obtain corresponding historical deviation values; And removing abnormal values from the historical deviation values of each unit period in the corresponding drainage stage of each similar historical coal-bed gas well, forming a historical deviation value set from the rest historical deviation values, screening the maximum historical deviation value, and taking the maximum historical deviation value as a set deviation value threshold.
7. The method for constructing the capacity split model of the multi-seam coal-seam gas well is characterized in that the method for acquiring the capacity split optimization model further comprises the following steps: based on the updated productivity split preliminary model, predicting the predicted production productivity of the target coal-bed gas well in each unit period again, and analyzing whether the updated productivity split preliminary model needs to be optimized or not; And if the optimization is not needed, taking the updated capacity split preliminary model as a capacity split optimization model, and if the optimization is needed, performing iterative adjustment on the updated capacity split preliminary model until the model subjected to iterative adjustment is not needed to be optimized.
8. The method for constructing the split model for the capacity of the multi-seam coal-seam gas well is characterized by comprising the following steps of: analyzing the change trend and the change rate of the corresponding combined mining data of each coal seam according to the real-time combined mining data sequence of each coal seam in each unit time period; based on the change trend and the change rate of the combined mining data, determining the combined mining data of each unit time period in the future set time length, and forming a combined mining data sequence of each unit time period in the future set time length of each coal seam; substituting the combined mining data sequence of each unit time period in the future set time period of each coal seam into a corresponding productivity split optimization model, outputting the predicted production capacity of each unit time period, and determining the predicted production capacity of the future set time period of each coal seam; And analyzing the productivity contribution rate of each coal bed according to the predicted production productivity of the future set duration of each coal bed, and identifying the dominant coal bed of the target coal bed gas well in the future set duration based on the productivity contribution rate.
9. The method for constructing the split model for the productivity of the multi-seam coal-seam gas well is characterized in that the productivity contribution rate of each coal seam is analyzed in the following way: And accumulating the predicted production capacity of the future set time length of each coal bed to obtain the total predicted production capacity of the future set time length, calculating the ratio of the predicted production capacity of the future set time length of each coal bed to the total predicted production capacity, and taking the ratio as the capacity contribution rate of each coal bed.

Description

Construction method of multi-seam combined production coal-seam gas well productivity split model Technical Field The invention relates to the technical field of coal mining, in particular to a method for constructing a split model of the productivity of a multi-seam coal-seam gas well. Background Along with the continuous deep development of coalbed methane resources, multi-coal-bed combined production has become an important technical means for improving single-well productivity and recovery ratio. In the multi-layer combined production process, the physical properties, gas content and permeability of reservoirs of different coal layers are obviously different, so that the contribution degree of each coal layer to the total yield is dynamically changed. The method is important to optimize the drainage system, formulate the accurate development strategy and realize the long-term stable production of the coal-bed gas well. The prior art mainly adopts an empirical model with fixed parameters or a static prediction model with one-time training to split the productivity, and mainly has the following problems that firstly, the prior art generally establishes the static prediction model based on initial geological conditions, a dynamic optimization mechanism is lacked in the production process, the problem of insufficient timeliness of the model exists, the model cannot adapt to the dynamic change characteristics of the productivity of each coal seam along with the advance of the drainage and mining stage, the prediction precision is gradually reduced, the dominant coal seams in different periods are difficult to accurately identify, and the timely adjustment of the mining scheme is influenced. Secondly, in the prior art, similarity matching of coal seam reservoir characteristics is not fully considered when historical data is utilized, only all historical well data or random sampling is simply adopted, the problem of poor model applicability exists, the method is difficult to adapt to personalized characteristics of a coal seam gas well, and further accuracy of later mining strategy formulation is affected, so that mining resource configuration is unreasonable. Disclosure of Invention The invention aims to overcome the defects of the prior art, provides a method for constructing a split model of the productivity of a multi-seam coal-seam gas well, and provides technical support for efficient development of multi-seam coal-seam gas production. The method for constructing the capacity split model of the multi-seam combined production coal-bed gas well comprises the steps of obtaining reservoir static data of all coal beds in a target coal-bed gas well, screening historical combined production data and historical production capacity of all similar historical coal-bed gas wells similar to the reservoir static data, and establishing a capacity split preliminary model of the target coal-bed gas well. And collecting real-time combined production data of each coal bed in the target coal bed gas well, and generating a real-time combined production data sequence of each coal bed in each unit time period. And inputting the real-time combined production data sequence into the established capacity split preliminary model, outputting the predicted production capacity of the target coal-bed gas well in each unit period, and calculating the deviation value of the predicted production capacity and the actual production capacity. And analyzing whether the productivity split preliminary model needs to be optimized or not based on the deviation value, and adjusting the productivity split preliminary model when the productivity split preliminary model needs to be optimized to obtain the productivity split optimization model. And predicting the predicted production capacity of the future set time length of each coal bed by combining the capacity split optimization model according to the real-time combined production data sequence of each coal bed in each unit time period, and identifying and displaying the main coal bed of the target coal bed gas well in the future set time length. Compared with the prior art, the method has the advantages that (1) the historical combined production data and the historical production capacity of each similar historical coal-bed gas well similar to the static data of the reservoir are screened, and the capacity split preliminary model of the target coal-bed gas well is established, so that the adaptability and the initial prediction precision of the construction model are improved, effective data is provided for formulating accuracy of a later mining strategy, and the mining resource allocation rationality is improved. (2) According to the method, the real-time combined production data sequence of each coal bed in each unit time period is used for outputting the predicted production capacity of the target coal bed gas well in each unit time period, the capacity split preliminary model is adjusted bas