Drilling pump fluid end fault diagnosis method based on cyclostationary-Gaussian index

Abstract

The invention discloses a drilling pump fluid end fault diagnosis method based on a cyclostationary-Gaussian index, and the method comprises the steps: firstly collecting original single-channel vibration signals of a fluid end under the normal operation of a target multi-cylinder drilling pump at equal intervals, then carrying out the signal preprocessing, and calculating the index values of VIAI and VGWI for the multi-section vibration signals; counting a mean value and a standard deviation; and then, setting a health threshold range, establishing a health threshold range of two diagnosis indexes VIAI and VGWI, and finally, calculating whether the two index values of the to-be-detected data are in the health threshold range so as to judge the health state of the fluid end of the drilling pump.

Description

Technical Field

[0001] This invention belongs to the field of mechanical fault diagnosis technology, and more specifically, relates to a method for diagnosing hydraulic end faults of drilling pumps based on the cyclic stability-Gaussianity index. Background Technology

[0002] Drilling pumps are the core equipment of the drilling fluid circulation system in oil and gas exploration. Their main function is to convert mechanical energy into hydraulic energy to power the circulation of drilling fluid. With the advancement of oil and gas exploration, such as the "Deep Earth Tak 1 Well" and "Deep Earth Chuanke 1 Well" ultra-deep well projects, higher requirements have been placed on the strength, durability, and operational stability of drilling pumps. As a critical component of the drilling pump, the hydraulic end withstands high pressure, high temperature, and erosion from complex media. Its failure can lead to drilling operation interruption or even serious safety accidents such as well blowouts. Therefore, fault diagnosis of critical components of the hydraulic end is of significant engineering importance.

[0003] Currently, fault diagnosis of the hydraulic end of drilling pumps largely relies on periodic manual inspections or experience-based judgment. This method has blind spots and struggles to capture early fault characteristics in a timely manner. In recent years, fault diagnosis technology based on vibration signals has received widespread attention due to its non-invasiveness and cost-effectiveness. Vibration signals can reflect the dynamic characteristics of a mechanical system; when hydraulic end components (such as valves and pistons) malfunction, their time and frequency domain characteristics change significantly. However, the vibration signals of multi-cylinder drilling pumps are characterized by homogeneous multi-source aliasing, non-stationarity, and strong noise interference. Traditional statistical indicators such as root mean square, peak-to-peak value, and kurtosis are difficult to effectively extract fault characteristics, especially in multi-cylinder structures, where fault location faces even greater challenges.

[0004] To address the aforementioned issues, there is an urgent need for a method that can extract fault features from vibration signals and achieve accurate diagnosis. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a fault diagnosis method for the hydraulic end of drilling pumps based on cyclic stationarity-Gaussianity indices. Since the vibration signal of the hydraulic end of a drilling pump exhibits significant cyclic stationarity, meaning the statistical characteristics of the signal change periodically with valve opening and closing; under healthy conditions, valve impacts have stable periodicity and amplitude; when a fault occurs (such as valve leakage or piston wear), the periodic characteristics of the impact change, manifesting as abnormal amplitude or phase, accompanied by additional random impacts or turbulent vibrations. Therefore, this invention constructs a Valve Impulse Amplitude Index (VIAI) and a Vibration Grading Weighted Index (VGWI) based on the cyclic stationarity and Gaussianity characteristics of the vibration signal, to achieve accurate detection and location of faults in key components of the hydraulic end of the drilling pump.

[0006] To achieve the above-mentioned objective, this invention provides a method for diagnosing hydraulic end faults in drilling pumps based on a cyclic stability-Gaussianity index, characterized by comprising the following steps:

[0007] (1) Collect vibration signals from the hydraulic end of the drilling pump;

[0008] Vibration sensors are placed on the cylinder head or cylinder wall of a multi-cylinder drilling pump so that they can collect the vibration and impact of valve opening and closing.

[0009] Set the sampling interval Δt of the vibration sensor;

[0010] Under normal operation of a multi-cylinder drilling pump, the vibration sensor collects a set of original single-channel vibration signals of length L from the hydraulic end of the drilling pump at intervals Δt, for a total of m sets, denoted as {x1[n], x2[n], ..., x l [n],…x m [n]}, n=0,1,2,…,L-1;

[0011] (2) Obtain the opening and closing impact frequency f of the hydraulic end valve of the multi-cylinder drilling pump. v ;

[0012] (3) The original single-channel vibration signal x of each group of drilling pump hydraulic ends l [n] is preprocessed, including but not limited to mean removal, outlier removal, and decoupling of homogeneous multi-source aliasing;

[0013] (4) Calculate x for each group l The valve impact amplitude index VIAI of [n];

[0014] (5) Calculate x for each group l The vibration grading weighted index VGWI of [n];

[0015] (6) Calculate m groups of VIAI indices {VIAI1,VIAI2,…,VIAI} l ,…,VIAI m} and VGWI indices {VGWI1, VGWI2, ..., VGWI} l ,…,VGWI m The mean and standard deviation of}: {μ VIAI ,σ VIAI}{μ VGWI ,σ VGWI};

[0016] Set the health threshold range as follows:

[0017]

[0018] Where k is a variable threshold coefficient;

[0019] (7) Fault diagnosis of multi-cylinder drilling pumps;

[0020] Under the unknown state of operation of a multi-cylinder drilling pump, a vibration sensor collects a set of original single-channel vibration signals x[n] of the hydraulic end of the drilling pump with a length of L, and then processes them according to steps (2) to (5) to calculate the valve impact amplitude index VIAI and the vibration classification weighted index VGWI.

[0021] Compare the two sets of indicators with the health threshold range set in step (6). If the average value of the two sets of indicators is within the health threshold range, the multi-cylinder drilling pump is determined to be in a healthy state; otherwise, the multi-cylinder drilling pump is determined to be in a fault state.

[0022] The objective of this invention is achieved as follows:

[0023] This invention relates to a drilling pump hydraulic end fault diagnosis method based on cyclic stability-Gaussianity indices. First, it collects raw single-channel vibration signals of the hydraulic end of a target multi-cylinder drilling pump under normal operation at equal intervals. Then, after signal preprocessing, it calculates the VIAI and VGWI indices for multiple vibration signals and statistically analyzes their mean and standard deviation. Subsequently, it sets health threshold ranges and establishes health threshold ranges for the two diagnostic indices, VIAI and VGWI. Finally, it calculates whether the values ​​of the two indices in the test data are within the health threshold range to determine the health status of the drilling pump hydraulic end.

[0024] Meanwhile, the drilling pump hydraulic end fault diagnosis method based on the cyclic stability-Gaussianity index of this invention also has the following beneficial effects:

[0025] (1) This invention proposes a two-dimensional vibration amplitude index for the cyclic stability and non-Gaussianity of the hydraulic end of a drilling pump, which can simultaneously reflect the change in impact amplitude and the thick tail of the distribution, overcoming the limitations of traditional statistical indicators.

[0026] (2) The index formula proposed in this invention is simple, does not rely on unknown statistical parameters, has physical interpretability, is suitable for online embedded real-time computing, and has a diagnostic accuracy rate that is significantly higher than traditional statistical indicators such as root mean square, peak-to-peak value and kurtosis.

[0027] (3) This invention proposes a fault diagnosis method for hydraulic end of drilling pump based on the cyclic stability-Gaussianity index, which is applicable to multi-cylinder reciprocating drilling pumps such as three-cylinder and five-cylinder pumps, fracturing pumps and other similar multi-cylinder reciprocating compression equipment, and has good promotion value. Attached Figure Description

[0028] Figure 1 This is a flowchart of the drilling pump hydraulic end fault diagnosis method based on the cyclic stability-Gaussianity index of the present invention.

[0029] Figure 2 This is a schematic diagram of VIAI indicator extraction.

[0030] Figure 3 This is a schematic diagram of VGWI indicator extraction.

[0031] Figure 4 This is a graph showing the classification results of health, valve failure, and piston failure data based on the VIAI+VGWI fusion index. Detailed Implementation

[0032] The specific embodiments of the present invention will now be described with reference to the accompanying drawings to enable those skilled in the art to better understand the invention. It should be particularly noted that in the following description, detailed descriptions of known functions and designs that might obscure the main content of the invention will be omitted here.

[0033] Example

[0034] Figure 1 This is a flowchart of the drilling pump hydraulic end fault diagnosis method based on the cyclic stability-Gaussianity index of the present invention.

[0035] In this embodiment, as Figure 1 As shown, a drilling pump hydraulic end fault diagnosis method based on the cyclic stability-Gaussianity index includes the following steps:

[0036] S1. Collect vibration signals from the hydraulic end of the drilling pump;

[0037] Vibration sensors are placed at the bottom of the cylinder wall of the five-cylinder drilling pump to collect the vibration and impact of valve opening and closing.

[0038] Set the sampling interval Δt of the vibration sensor;

[0039] Under normal operation of a multi-cylinder drilling pump, the vibration sensor collects a set of original single-channel vibration signals of length L from the hydraulic end of the drilling pump at intervals Δt, for a total of m sets, denoted as {x1[n], x2[n], ..., x l [n],…x m [n]}, n=0,1,2,…,L-1; In this embodiment, a total of 195 groups were collected, and the length L of each signal segment was 125000 data points.

[0040] S2, Obtain the opening and closing impact frequency f of the hydraulic valve. v ;

[0041] f v It can be determined by the pump impulse under the current operating conditions of the drilling pump, and can be automatically obtained and calculated through vibration signal spectrum analysis, current signal analysis, and key phase signal analysis.

[0042] In this embodiment, the valve opening and closing impact frequency is obtained by analyzing and converting the drilling pump motor current signal. The operating condition of the five-cylinder pump is 100 sppm, and the theoretical valve opening and closing impact frequency f is obtained by conversion. v The frequency is 1.67 Hz. Other methods for obtaining this information include, but are not limited to, vibration signal spectrum analysis, current signal analysis, and bond phase signal analysis.

[0043] S3. Vibration signal preprocessing;

[0044] For each group of drilling pump hydraulic end raw single-channel vibration signals x l [n] is preprocessed, including mean removal, outlier removal, and decoupling of homogeneous multi-source aliasing;

[0045] In this embodiment, the original signal is subjected to a mean removal operation, and the original vibration signal is processed using a homogeneous multi-source aliasing decoupling algorithm.

[0046] S4. Calculate the valve impact amplitude index VIAI;

[0047] S4.1, regarding x l Perform Discrete Fourier Transform on [n]:

[0048]

[0049] Among them, X l (f) is x l The frequency domain representation of [n], where f represents the discrete frequency index and j is the complex unit;

[0050] S4.2, X l The negative frequency component of (f) is set to zero, the positive frequency component is multiplied by 2, and then an inverse discrete Fourier transform is performed to obtain the imaginary part of the signal.

[0051] S4.3, x l [n] and Superimpose the signals to obtain the analytic signal z. l [n]:

[0052]

[0053] S4.4 Calculate the magnitude of the analytic signal:

[0054]

[0055] S4.5, regarding env l [n] is subjected to a discrete Fourier transform to obtain the envelope spectrum ENV. l (f):

[0056]

[0057] S4.6 Extract the valve opening and closing frequency f from the envelope spectrum. v The amplitude at that point is used as the value of the VIAI index. l :

[0058] VIAI l =|ENV l (f v )|

[0059] In this embodiment, as Figure 2 As shown, the maximum peak value near the valve opening and closing frequency point is selected as the VIAI index value.

[0060] S5. Calculate the vibration grading weighted index VGWI;

[0061] S5.1, regarding x l [n] Apply a logarithmic transformation:

[0062]

[0063] Where, k β The scaling parameter is 2 in this embodiment; the extreme values ​​of the signal x[n] are compressed and the tail features are preserved through logarithmic transformation.

[0064] S5.2, y l [n] maps to the interval [-1, 1]:

[0065]

[0066] in, Sequence y l Minimum and maximum values ​​in [n];

[0067] S5.3. Divide the interval [-1,1] into M equal levels, with the boundaries as follows:

[0068]

[0069] In this embodiment, M is set to 10;

[0070] S5.4, Calculate the values ​​for each interval [edges] j'-1 ,edges j' The number of points n in ] j' And set the weight ω for each interval. j' Set the amplitude factor A and calculate VGWI. l :

[0071]

[0072] In this embodiment, the VGWI indicator extraction diagram is as follows: Figure 3 As shown, in this embodiment, the weight ω j' Set to [10,1,0.5,0.03,0.001,0.001,0.03,0.5,1,10], with amplitude factor A set to 1.

[0073] S6. Calculate the diagnostic index values ​​of multiple sets of health data;

[0074] Calculate m groups of VIAI indices {VIAI1,VIAI2,…,VIAI} l ,…,VIAI m} and VGWI indices {VGWI1, VGWI2, ..., VGWI} l ,…,VGWI m The mean and standard deviation of}: {μ VIAI ,σ VIAI}{μ VGWI ,σ VGWI};

[0075] Set the health threshold range as follows:

[0076]

[0077] Where k is a variable threshold coefficient, which is set to 2 in this embodiment.

[0078] S7. Calculate the diagnostic index values ​​of the data to be tested and perform fault diagnosis.

[0079] Under the unknown operating conditions of the five-cylinder drilling pump, the vibration sensor collects a set of original single-channel vibration signals x[n] of the hydraulic end of the drilling pump with a length of L. Then, it is processed according to steps S2 to S5 to calculate the valve impact amplitude index VIAI and the vibration classification weighted index VGWI.

[0080] Compare the two sets of indicators with the health threshold range set in step S6. If the average value of the two sets of indicators is within the health threshold range, the multi-cylinder drilling pump is determined to be in a healthy state; otherwise, the multi-cylinder drilling pump is determined to be in a fault state.

[0081] Instance verification

[0082] In this embodiment, a fault simulation experiment was conducted using an XZQ-2200 five-cylinder single-acting drilling pump test platform to obtain a dataset for verifying the method of the present invention. The dataset is shown in Table 1. This embodiment uses a portion of the healthy data to construct a diagnostic threshold and uses fault data and other healthy data for testing. This embodiment was tested on a device with the MATLAB R2022a environment installed. The collected data was calculated according to steps S1 to S7 in the embodiment, and the final diagnostic results are shown in Table 2.

[0083] Table 1. Fault simulation dataset for XZQ-2200 five-cylinder drilling pump.

[0084] Fault type Pump flushing operation (spm) Sampling frequency (Hz) Sampling time per group / (s) Quantity / (groups) Health data 100 25000 5 65 Valve malfunction 100 25000 5 65 Piston failure 100 25000 5 65

[0085] Table 2. Diagnostic results of the XZQ-2200 five-cylinder drilling pump fault simulation dataset.

[0086]

[0087] Table 2 shows the diagnostic results of the proposed index and three traditional statistical indices (root mean square, peak-to-peak value, and kurtosis) on five-cylinder pump fault data. The classification results of the proposed VIAI+VGWI fusion index on health and fault data are as follows: Figure 4 As shown, piston failure generally leads to changes in the VIAI index, while valve failure is more often manifested as changes in the VGWI index. This demonstrates the effectiveness of the circulation stability-Gaussianity index in diagnosing hydraulic end failures of drilling pumps.

[0088] The above results show that, compared with traditional statistical indicators such as root mean square, peak-to-peak value, and kurtosis, the fault detection method of the present invention performs better in terms of fault diagnosis accuracy, confirming its good performance in detecting faults in key components of the hydraulic end of drilling pumps.

[0089] Although the illustrative specific embodiments of the present invention have been described above to enable those skilled in the art to understand the invention, it should be understood that the invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the invention as defined and determined by the appended claims, and all inventions utilizing the concept of the present invention are protected.

Claims

1. A method for diagnosing hydraulic end faults in drilling pumps based on cyclic stability-Gaussianity indices, characterized in that, The following steps are involved: (1) Collect vibration signals from the hydraulic end of the drilling pump; Vibration sensors are placed on the cylinder head or cylinder wall of a multi-cylinder drilling pump so that they can collect the vibration and impact of valve opening and closing. Set the sampling interval Δt of the vibration sensor; Under normal operation of a multi-cylinder drilling pump, the vibration sensor collects a set of original single-channel vibration signals of length L from the hydraulic end of the drilling pump at intervals Δt, for a total of m sets, denoted as {x1[n], x2[n], ..., x l [n],…x m [n]}, n=0,1,2,…,L-1; (2) Obtain the opening and closing impact frequency f of the hydraulic end valve of the multi-cylinder drilling pump. v ; (3) The original single-channel vibration signal x of each group of drilling pump hydraulic ends l [n] is preprocessed, including but not limited to mean removal, outlier removal, and decoupling of homogeneous multi-source aliasing; (4) Calculate x for each group l The valve impact amplitude index VIAI of [n]; (5) Calculate x for each group l The vibration grading weighted index VGWI of [n]; (6) Calculate m groups of VIAI indices {VIAI1,VIAI2,…,VIAI} l ,…,VIAI m } and VGWI indices {VGWI1, VGWI2, ..., VGWI} l ,…,VGWI m The mean and standard deviation of}: {μ VIAI ,σ VIAI }{μ VGWI ,σ VGWI }; Set the health threshold range as follows: Where k is a variable threshold coefficient; (7) Fault diagnosis of multi-cylinder drilling pumps; Under the unknown state of operation of a multi-cylinder drilling pump, a vibration sensor collects a set of original single-channel vibration signals x[n] of the hydraulic end of the drilling pump with a length of L, and then processes them according to steps (2) to (5) to calculate the valve impact amplitude index VIAI and the vibration classification weighted index VGWI. Compare the two sets of indicators with the health threshold range set in step (6). If the average value of the two sets of indicators is within the health threshold range, the multi-cylinder drilling pump is determined to be in a healthy state; otherwise, the multi-cylinder drilling pump is determined to be in a fault state.

2. The drilling pump hydraulic end fault diagnosis method based on the cyclic stability-Gaussianity index according to claim 1, characterized in that, The method for calculating the valve impact amplitude index VIAI is as follows: (2.1) For x l Perform Discrete Fourier Transform on [n]: Among them, X l (f) is x l The frequency domain representation of [n], where f represents the discrete frequency index and j is the complex unit; (2.2) X l The negative frequency component of (f) is set to zero, the positive frequency component is multiplied by 2, and then an inverse discrete Fourier transform is performed to obtain the imaginary part of the signal. (2.3) x l [n] and Superimpose the signals to obtain the analytic signal z. l [n]: (2.4) Calculate the magnitude of the analytic signal: (2.5) Regarding env l [n] is subjected to a discrete Fourier transform to obtain the envelope spectrum ENV. l (f): (2.6) Extract the valve opening and closing frequency f from the envelope spectrum v The amplitude at that point is used as the value of the VIAI index. l : WHY l =|ENV l (f v )|。 3. The drilling pump hydraulic end fault diagnosis method based on the cyclic stability-Gaussianity index according to claim 1, characterized in that, The vibration grading weighted index VGWI is calculated as follows: (3.1) For x l [n] Apply a logarithmic transformation: Where, k β For scale parameters; (3.2) y l [n] maps to the interval [-1, 1]: in, Sequence y l Minimum and maximum values ​​in [n]; (3.3) Divide the interval [-1,1] into M equal levels, with the boundaries as follows: (3.4) Calculate the values ​​for each interval [edges] j'-1 ,edges j' The number of points n in ] j' And set the weight ω for each interval. j' Set the amplitude factor A and calculate VGWI. l :

Images