Early comprehensive monitoring and dynamic evolution grading early warning method for gas invasion in drilling process

Abstract

The present application belongs to the field of oil exploration and development engineering, and particularly relates to a method for comprehensive monitoring and dynamic evolution grading early warning of gas invasion in the drilling process. The present application simultaneously collects multi-source information to jointly monitor whether a gas invasion event occurs, judges a risk coefficient according to a gas invasion state, sums the risk coefficient, and grades the early warning of the gas invasion event. The present application fully utilizes the low-frequency elastic wave characteristic response data near the drill bit, the gas-sensitive sensor and the logging data, establishes a gas invasion multi-source information early warning method, sets multiple levels of early warning, and determines the gas invasion risk level according to the slight fluctuation of different monitoring results at the well bottom, along the wellbore and at the wellhead, proposes the concept of grading early warning, realizes a continuous early warning system from "early identification and early intervention" to "timely control", and effectively enhances the accuracy and timeliness of the gas invasion early warning.

Description

Technical Field

[0001] This invention belongs to the field of petroleum exploration and development engineering, specifically relating to a method for early comprehensive monitoring and dynamic evolution classification and early warning of gas invasion during drilling. Background Technology

[0002] With the in-depth research on oil and gas resource development, the exploration and development of deep oil and gas resources, represented by deep water and deep earth, is becoming a research hotspot. Deep oil and gas drilling requires huge capital investment per well and faces challenges such as high temperature, high pressure, and narrow density windows. Complex drilling situations, represented by gas invasion, occur frequently, which are difficult and costly to handle, bringing many challenges to well control work.

[0003] Currently, most gas intrusion monitoring methods rely on surface monitoring or near-wellhead monitoring. Gas intrusion is only identified when parameters monitored at the surface or near-wellhead change. However, this method suffers from low accuracy and poor timeliness. For example, when a gas intrusion alarm occurs near the wellhead or on the surface, the warning level is usually already quite high, and subsequent response measures may not be possible in time. It is also difficult to detect weak gas intrusion phenomena at the bottom of deep-water or deep-earth wells, thus missing some early intervention opportunities. Summary of the Invention

[0004] The purpose of this invention is to provide a comprehensive early monitoring and dynamic evolution-based graded early warning method for gas invasion during drilling, addressing the problems of sensitivity, reliability, and timeliness in existing gas invasion monitoring methods. This invention establishes multi-level early warning systems, determining the gas invasion risk level step-by-step based on subtle fluctuations in monitoring results at the bottom of the well, along the wellbore, and at the wellhead, thus achieving a continuous early warning system from "early identification and early intervention" to "timely control."

[0005] To achieve the aforementioned objectives, the technical solution adopted by this invention is as follows: simultaneously collecting multi-source information to jointly monitor whether an air intrusion event has occurred, determining the risk coefficient based on the air intrusion status, summing the risk coefficients, and issuing graded early warnings for air intrusion events. The specific steps are as follows:

[0006] S1. During drilling, monitor the low-frequency elastic wave characteristic response near the drill bit to calculate the gas content of the cross section, monitor whether gas invasion events occur, and determine the risk factor R1;

[0007] S2. During drilling, gas-sensitive sensors are used to monitor the gas content of the wellbore cross-section, monitor whether gas intrusion events occur, and determine the risk factor R2.

[0008] S3. During the drilling process, obtain engineering logging parameters to calculate the probability of gas invasion, monitor whether gas invasion events occur, and determine the risk coefficient R3;

[0009] S4. During the drilling process, use wellhead mud logging data to monitor whether gas invasion events occur and determine the risk factor R4;

[0010] S5. Sum the risk coefficients from steps S1 to S4 to obtain R, and then issue a graded early warning.

[0011] In step S1, a low-frequency elastic wave monitoring sub is typically used. This sub includes a signal generator, a signal receiver, and a signal processing unit. When the monitoring sub detects gas intrusion near the drill bit, the signal receiver generates a characteristic response, i.e., a phase velocity change. The signal processing unit can calculate the gas content of the wellbore cross-section after gas intrusion based on this characteristic response, achieving quantitative monitoring of the gas content. This gas content is used to determine whether gas intrusion has occurred. The gas content monitoring threshold is related to the depth near the drill bit and the properties of the drilling fluid used. When the gas content exceeds the threshold but is less than 20% of it, a gas intrusion event occurs, with a risk factor R1 = 1. When the gas content exceeds the threshold by 20-50%, a gas intrusion event occurs, with a risk factor R1 = 2. When the gas content exceeds the threshold by 50%, a gas intrusion event occurs, with a risk factor R1 = 3.

[0012] Since this invention uses low-frequency elastic waves, the frequency has a relatively small impact on the propagation of low-frequency sound waves. Propagation in bubble-containing liquids is mainly affected by the gas content. In practical field applications of acoustic gas intrusion monitoring technology, the sound wave frequency is usually fixed. To simplify calculations, the relationship between the cross-sectional gas content and the phase velocity is as follows:

[0013] ;

[0014] In the formula, v is the phase velocity, m / s; β is the gas content, %; a and b are fitting coefficients, which need to be obtained by fitting based on the actual working conditions and historical data of nearby wells, m. 0.5 / s 0.5 .

[0015] Based on this equation, the gas content β can be inversely calculated using the phase velocity. Compared with the prior art, this invention can save a lot of complex number operations, has a good fitting effect under low gas content conditions, and greatly reduces the calculation difficulty while ensuring accuracy, thus enabling quantitative monitoring of gas content.

[0016] In step S2, a gas sensor is typically used for monitoring. After each drilling cycle, the gas sensor is lowered to the designed position via a cable that provides power and transmits monitoring data. The gas sensor needs to be manufactured into a curved shell shape with a curvature matching the curvature of the casing's inner wall. It is then attracted to the casing's inner wall using a strong magnet. Simultaneously, the shell thickness should be less than the gap between the casing's inner diameter and the drill bit diameter for the next drilling cycle to prevent stuck pipe during tripping. This gas sensor monitors whether the gas content in the wellbore reaches the gas intrusion alarm threshold; if it does, gas intrusion is considered to have occurred. The cross-sectional gas content monitoring threshold is related to the depth of the gas sensor's monitoring location and the properties of the drilling fluid used. When the cross-sectional gas content exceeds the monitoring threshold, a gas intrusion event is considered to have a risk factor R2 = 2.

[0017] In step S3, the engineering logging parameters include drilling pressure, hook load, torque, standpipe pressure, displacement, and mechanical drilling speed. These engineering logging parameters show a significant response after gas invasion occurs, meeting the timeliness requirements for gas invasion judgment. This invention utilizes drilled well data from the same block, combined with the characteristic responses of logging parameters to gas invasion events, and calculates the predicted probability based on an intelligent classification model of logging parameters-gas invasion status using an artificial neural network to determine whether a gas invasion event has occurred. The specific process is as follows:

[0018] The input sample contains six feature parameters: drill pressure, hook load, torque, riser pressure, displacement, and mechanical drilling speed, denoted as x.

[0019] ;

[0020] The corresponding output label y takes the value of 0 (no air intrusion) or 1 (air intrusion occurs); when an air intrusion event is determined to have occurred, the risk coefficient R3 = 3;

[0021] The intelligent classification model for logging parameters and gas invasion status uses a feedforward neural network, which includes an input layer, an output layer, and three hidden layers.

[0022] The formula for calculating the hidden layer is:

[0023] ;

[0024] The formula for calculating the output layer is:

[0025] ;

[0026] In the formula, W is the weight matrix, and b is the bias vector. is a non-linear activation function, z is an intermediate variable, and p represents the probability of predicting air intrusion.

[0027] In step S4, the mud logging data refers to the increase in the mud pit volume; when the increase in the mud pit reaches 1m... 3An air intrusion event occurs, with a risk factor R4=2.

[0028] The risk coefficients in each step of this invention are not arbitrarily determined, but are proposed based on the hierarchy of an event and the effectiveness of the response to a gas intrusion event. For example, step S1, which monitors gas intrusion near the drill bit, has good timeliness and high accuracy; step S3, which comprehensively utilizes engineering logging data, has weaker timeliness but can be comprehensively judged and has high accuracy, so the highest risk coefficient for both is set to 3; the highest risk coefficient for steps S2 and S4 is set to 2. If no gas intrusion event occurs, the risk coefficient is 0.

[0029] In step S5, the warning level is assessed step by step according to the following criteria, and the highest warning level is taken as the final warning level.

[0030] (S5-1) If 1≤R≤2, Level 1 warning; if 3≤R≤5, Level 2 warning; if 6≤R≤8, Level 3 warning; if 9≤R≤10, Level 4 warning;

[0031] (S5-2) Considering the timeliness and accuracy requirements of air intrusion monitoring, if R1=3 or R3=3, a Level 3 warning is issued; if R1=3 and R3=3, a Level 4 warning is issued.

[0032] (S5-3) Consider the situation where there is no free gas near the drill bit or in a deeper part of the wellbore due to excessive pressure or gas dissolving in drilling fluid. If R3=3 and R4=2, a level 4 warning is issued.

[0033] Compared with existing technologies, this invention proposes a comprehensive early monitoring and dynamic evolution-based graded early warning method for gas intrusion at the wellhead, wellbore, and near-bit depths. It fully utilizes near-bit drilling data, mud logging data, and gas-sensitive sensors to establish a multi-source information early warning method for gas intrusion. By setting multiple warning levels, the risk level of gas intrusion can be determined step-by-step based on subtle fluctuations in monitoring results at the wellhead, along the wellbore, and at the wellbore. This invention introduces the concept of graded early warning, realizing a continuous early warning system from "early identification and early intervention" to "timely control," effectively enhancing the accuracy and timeliness of gas intrusion early warning. Furthermore, the graded early warning system allows for different processing strategies to be determined based on different warning levels, reducing unnecessary processing time. Attached Figure Description

[0034] Figure 1 This invention relates to a schematic diagram of the downhole distribution scheme;

[0035] Figure 2 This is a schematic diagram illustrating the application effect of the present invention. Detailed Implementation

[0036] To enable those skilled in the art to better understand the technical solutions of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0037] like Figure 1 As shown, a method for early comprehensive monitoring and dynamic evolution-based graded early warning of gas invasion during drilling includes the following steps:

[0038] (1) Collect existing field logging data. In real-world applications, such data can be obtained using logging tools. The logging data used in this invention includes depth, hook load, drilling pressure, torque, displacement, mechanical drilling speed, surface mud pit increment, as well as low-frequency elastic wave characteristic response and gas sensor monitoring data. The data comes from logging data of a certain actual drilled well from 17:00 to 17:45, with a collection frequency of 1HZ / s.

[0039] (2) During drilling, monitor the characteristic response near the drill bit to calculate the gas cut of the cross section, monitor whether a gas invasion event occurs and determine the risk coefficient R1. The gas cut monitoring threshold is set to 2% based on the drilling fluid properties. The specific calculation of the gas cut of the cross section is as follows:

[0040] ;

[0041] In the formula, v is the phase velocity, m / s; β is the gas content, %; a and b are fitting coefficients, obtained by fitting based on actual field conditions and historical data from nearby wells, m. 0.5 / s 0.5 R represents the goodness of fit.

[0042] (4) During the drilling process, the gas content of the cross section is monitored by the gas sensor located on the inner wall of the second casing of the wellbore, to monitor whether a gas invasion event occurs and to determine the risk coefficient R2. The monitoring threshold is set to 5% according to the properties of the drilling fluid.

[0043] (2) Use the collected engineering logging parameters to predict the probability of gas invasion, monitor whether gas invasion events occur and determine the risk coefficient R3.

[0044] The predicted probability is calculated based on an intelligent classification model of logging parameters and gas intrusion status using an artificial neural network; the input sample of the intelligent classification model of logging parameters and gas intrusion status is the engineering logging parameters, denoted as x:

[0045] ;

[0046] The corresponding output label y takes the value of 0 (no air intrusion) or 1 (air intrusion occurs);

[0047] The model uses a feedforward neural network, including an input layer, an output layer, and three hidden layers;

[0048] The formula for calculating the hidden layer is:

[0049] ;

[0050] The formula for calculating the output layer is:

[0051] ;

[0052] In the formula, W is the weight matrix, and b is the bias vector. is a non-linear activation function, z is an intermediate variable, and p represents the probability of predicting air intrusion.

[0053] (5) During drilling, use wellhead mud logging data to monitor whether gas invasion events occur and determine the risk factor R4. Mud logging data refers to the mud pit increment; when the mud pit increment reaches 1m 3 This can be considered as an air invasion.

[0054] (6) The collected field data was used to conduct simulation tests of the solution, and the results are as follows: Figure 2 As shown, at 17:02, the near-bit gas intrusion monitoring method at the bottom of the well indicates gas intrusion. The early warning method based on low-frequency elastic waves predicts that the gas content of the cross-section exceeds the threshold (2%) but is less than 20%, with a risk coefficient of 1, resulting in a Level 1 warning. At 17:07, the early warning method based on low-frequency elastic waves predicts that the gas content of the cross-section exceeds the threshold of 20% (2.4%), at which point the risk coefficient becomes 2, and the gas intrusion warning is still a Level 1 warning. At 17:12, the early warning method based on low-frequency elastic waves predicts that the gas content of the cross-section exceeds the threshold of 20% (2.4%), and the wellbore prediction method using gas-sensitive sensors predicts that the gas content exceeds 5%, with a risk coefficient of 2+2, resulting in a Level 2 warning. At 17:17, the early warning method based on low-frequency elastic waves predicts that the gas content of the cross-section exceeds the threshold of 50% (3%), and the risk coefficient of this method becomes 3. Although the sum of the risk coefficients is 5, it is still judged as a Level 3 warning. At 17:26, based on the monitoring method using engineering logging, gas intrusion was detected, and the gas intrusion warning level evolved to Level 4. Starting at 17:34, the mud pit monitoring method was used to detect gas intrusion, and the gas intrusion warning level evolved to Level 4. This case comparison shows that the monitoring scheme described in this invention can provide an earlier warning of gas intrusion events than conventional surface monitoring (using wellhead mud logging data), by at least 8 minutes, thus gaining valuable time for handling gas intrusion events during on-site construction.

[0055] It should be noted that, for the sake of simplicity, the aforementioned method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, because according to this application, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and units involved are not necessarily essential to this application.

Claims

1. A method for early comprehensive monitoring and dynamic evolution-based graded early warning of gas invasion during drilling, characterized in that, Simultaneously, information from multiple sources is collected to jointly monitor whether air intrusion events occur, the risk coefficient is determined based on the air intrusion status, the risk coefficients are summed, and air intrusion events are classified and warned. The specific steps are as follows: S1. During drilling, monitor the low-frequency elastic wave characteristic response near the drill bit to calculate the gas content of the cross section, monitor whether gas invasion events occur, and determine the risk factor R1; When the gas content of the cross section exceeds the threshold threshold but is less than 20% of the threshold threshold, an intrusion event occurs, with a risk factor R1=1; when the gas content of the cross section exceeds the threshold threshold threshold by 20-50%, an intrusion event occurs, with a risk factor R1=2; when the gas content of the cross section exceeds the threshold threshold threshold by 50%, an intrusion event occurs, with a risk factor R1=3. S2. During drilling, gas-sensitive sensors are used to monitor the gas content of the wellbore cross-section, monitor whether a gas invasion event occurs, and determine the risk factor R2. When the gas content of the cross-section exceeds the monitoring threshold, a gas invasion event occurs, and the risk factor R2 = 2. S3. During the drilling process, obtain engineering logging parameters to calculate the probability of gas invasion, monitor whether a gas invasion event occurs, and determine the risk coefficient R3; when a gas invasion event is determined to have occurred, the risk coefficient R3 = 3; S4. During drilling, use wellhead mud logging data to monitor for gas invasion events and determine the risk factor R4; mud logging data refers to the mud pit increment; when the mud pit increment reaches 1m... 3 The risk factor R4 = 2 in the event of an air intrusion. S5. Sum the risk coefficients from steps S1 to S4 to obtain R, and then issue graded early warnings. In step S5, the warning level is assessed step by step according to the following criteria, and the highest warning level is taken as the final warning level. (S5-1) If 1≤R≤2, Level 1 warning; if 3≤R≤5, Level 2 warning; if 6≤R≤8, Level 3 warning; if 9≤R≤10, Level 4 warning; (S5-2) If R1=3 or R3=3, a Level 3 warning is issued; if R1=3 and R3=3, a Level 4 warning is issued. (S5-3) If R3=3 and R4=2, a level 4 warning is issued.

2. The method for early comprehensive monitoring and dynamic evolution-based graded early warning of gas invasion during drilling as described in claim 1, characterized in that, In step S1, the low-frequency elastic wave characteristic response is the phase velocity; the relationship between the cross-sectional gas content and the phase velocity is as follows: ； In the formula, v is the phase velocity, m / s; β is the gas content, %; a and b are fitting coefficients, which need to be obtained by fitting based on the actual working conditions and historical data of nearby wells, m. 0.5 / s 0.5 .

3. The method for early comprehensive monitoring and dynamic evolution-based graded early warning of gas invasion during drilling as described in claim 1, characterized in that, In step S1, the gas content monitoring threshold is related to the depth of the near-bit monitoring location and the properties of the drilling fluid used.

4. The method for early comprehensive monitoring and dynamic evolution-based graded early warning of gas invasion during drilling as described in claim 1, characterized in that, In step S2, the gas content monitoring threshold of the cross section is related to the depth of the gas sensor monitoring location and the properties of the drilling fluid used.

5. The method for early comprehensive monitoring and dynamic evolution-based graded early warning of gas invasion during drilling as described in claim 1, characterized in that, In step S3, the engineering logging parameters include drilling pressure, hook load, torque, standpipe pressure, displacement, and mechanical drilling speed.

6. The method for early comprehensive monitoring and dynamic evolution-based graded early warning of gas invasion during drilling as described in claim 1, characterized in that, In step S3, the predicted probability is calculated based on the intelligent classification model of logging parameters-gas invasion state using an artificial neural network; the input sample of the intelligent classification model of logging parameters-gas invasion state is the engineering logging parameters, denoted as x: ； The corresponding output label y takes a value of 0, indicating no air intrusion, or 1, indicating air intrusion. The intelligent classification model for logging parameters and gas invasion status uses a feedforward neural network, which includes an input layer, an output layer, and three hidden layers. The formula for calculating the hidden layer is: ； The formula for calculating the output layer is: ； In the formula, W is the weight matrix, and b is the bias vector. is a non-linear activation function, z is an intermediate variable, and p represents the probability of predicting air intrusion.

Images