Valve impact phase diagnosis index and method for fluid end of multi-cylinder drilling pump

Abstract

The invention discloses a drilling pump fluid end anomaly detection method based on a valve impact phase index. The method comprises the steps that firstly, a signal decomposition method is adopted to extract valve impact components in vibration signals; then, extracting an impact peak value in each period, and taking the impact peak value as a valve impact phase of the corresponding period; the relative phase is calculated according to the phase data between the adjacent cylinder bodies; further calculating a variance index and a mean value index of the relative phase to represent a phase distribution characteristic; and finally, an anomaly correlation matrix is constructed based on the indexes, and anomaly detection and fault cylinder positioning identification are realized in combination with a preset judgment rule.

Description

Technical Field

[0001] The present invention belongs to the technical field of mechanical fault diagnosis, and more specifically, relates to a valve impact phase diagnosis index and method for a hydraulic end of a multi-cylinder drilling pump. Background Art

[0002] The drilling pump is a core component of the drilling fluid circulation system. Its primary function is to convert mechanical energy into hydraulic energy, powering the bottomhole drill bit and maintaining continuous circulation of the drilling fluid. Widely used in oil and gas drilling, water injection, and polymer flooding operations, this equipment is an essential component of the oil and gas production process. To meet the demands of deep-layer oil and gas exploration and development, new five-cylinder drilling pumps are gradually replacing traditional three-cylinder pumps and becoming the mainstream equipment. Compared to traditional models, these five-cylinder drilling pumps offer higher power, larger displacement, and improved operational stability, significantly improving the efficiency and safety of deep-layer operations.

[0003] The hydraulic end of a drilling pump consists of several key components. The piston and valve components (such as the valve body and valve seat) are subject to long-term, harsh operating conditions of high pressure and high frequency impact, making them susceptible to wear, erosion, and fatigue damage, leading to frequent failures. These failures not only severely impact drilling efficiency but can also cause safety incidents such as equipment damage or personal injury. Therefore, abnormality detection and predictive maintenance of drilling pumps have important engineering applications.

[0004] Currently, maintenance of drilling pumps at oil and gas drilling sites primarily relies on regular inspections and post-event repairs, lacking an efficient and intelligent fault warning mechanism. Due to the complex structure of five-cylinder drilling pumps, any anomaly typically requires disassembly and inspection of each cylinder, which is not only time-consuming and labor-intensive but also hinders operational progress. Therefore, a technical solution is urgently needed that can enable online monitoring of drilling pump operating status, detect anomalies, and quickly locate faulty cylinders.

[0005] Common hydraulic end failures in drilling pumps can be identified using the principles of equipment operation. Damage to valve or piston components can degrade the cylinder seal, leading to puncture leaks and abnormal pressure fluctuations within the cylinder. These pressure fluctuations directly impact the impact behavior during valve opening and closing, causing phase shifts in the vibration signal. Based on this principle, by analyzing the phase characteristics of the impact generated when the discharge valve is closed and constructing corresponding diagnostic indicators, it is possible to automatically identify abnormal drilling pump conditions and locate faulty cylinders, providing a reliable basis for equipment maintenance. Summary of the Invention

[0006] The purpose of the present application is to overcome the shortcomings of the prior art, provide a valve impact phase diagnostic index and method for a hydraulic end of a multi-cylinder drilling pump, which quantifies the phase difference and fluctuation amplitude between cylinders by the distribution characteristics of the discharge valve closing impact phase, so as to realize abnormal detection of the fault cylinder.

[0007] In order to achieve the above-mentioned purpose of the application, the valve impact phase diagnostic index and method for the hydraulic end of the multi-cylinder drilling pump comprises the following steps:

[0008] (1) Collecting valve vibration signals of each cylinder of the hydraulic end of the multi-cylinder drilling pump;

[0009] (2) Extracting the discharge valve closing impact signal in the valve vibration signal of each cylinder;

[0010] (3) Extracting the valve impact phase of each cylinder;

[0011] (4) Calculating the relative phase of the valve impact between cylinders;

[0012] (5) Calculating the diagnostic index based on the relative phase;

[0013] (6) Constructing an abnormal correlation matrix based on the diagnostic index of the relative phase;

[0014] (7) Determining the abnormality of each cylinder through the abnormal correlation matrix.

[0015] The purpose of the present application is achieved as follows:

[0016] The drilling pump hydraulic end abnormality detection method based on the valve impact phase index of the present application first extracts the valve impact component in the vibration signal by using a signal decomposition method; then, the impact peak value in each cycle is extracted and used as the valve impact phase of the corresponding cycle; the relative phase between adjacent cylinder bodies is calculated; the variance index and mean index of the relative phase are further calculated to represent the phase distribution characteristics; finally, the abnormal correlation matrix is constructed based on the above-mentioned index, and the abnormal detection and positioning identification of the fault cylinder are realized by combining the preset determination rule.

[0017] Meanwhile, the drilling pump hydraulic end abnormality detection method based on the valve impact phase index of the present application also has the following beneficial effects:

[0018] (1) The present application researches based on the phase characteristics of valve vibration impact. When a leakage fault occurs, it will cause abnormal fluctuation of the cylinder pressure, and then cause the phase of the vibration signal to change. By extracting and analyzing the above-mentioned phase change characteristics, the abnormal state can be identified and diagnosed.

[0019] (2) The present invention proposes a relative phase distribution index to characterize the distribution characteristics of the phase in the system. This index can effectively reflect the rationality of the phase distribution, thereby assisting in determining whether the equipment is operating normally.

[0020] (3) The present invention constructs an abnormal correlation matrix. By analyzing the relative phase index values ​​between two adjacent cylinders, it is determined whether their relative phases are within the normal range. Furthermore, by combining the abnormal correlation index between a cylinder and its two adjacent cylinders, it is possible to comprehensively determine whether the cylinder has a fault, providing a basis for system fault location. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a flow chart of a method for detecting abnormalities at the hydraulic end of a drilling pump based on a valve impact phase index according to the present invention;

[0022] Figure 2 It is a schematic diagram of extracting the valve impact sequence.

[0023] Figure 3 Schematic diagram of the phase obtained by peak extraction from the valve impact sequence.

[0024] Figure 4 It is a schematic diagram of constructing a relative phase indicator matrix and calculating an abnormal correlation matrix based on a threshold. DETAILED DESCRIPTION

[0025] The following describes the specific embodiments of the present invention in conjunction with the accompanying drawings so that those skilled in the art can better understand the present invention. It should be noted that in the following description, when detailed descriptions of known functions and designs may dilute the main content of the present invention, such descriptions will be omitted here.

[0026] Example

[0027] In this embodiment, if Figure 1 As shown, the present invention provides a valve impact phase diagnostic indicator and method for the hydraulic end of a multi-cylinder drilling pump, comprising the following steps:

[0028] (1) Collect the vibration signals of valves of each cylinder at the hydraulic end of a multi-cylinder drilling pump;

[0029] (1.1) Assume that a multi-cylinder drilling pump contains H cylinders. Number each cylinder according to the movement sequence between the cylinders of the drilling pump, and record it as i = 1, 2, ..., H;

[0030] In this embodiment, a five-cylinder drilling pump is taken as an example, that is, H=5;

[0031] (1.2) A vibration sensor is arranged in the middle of the bottom of each cylinder to collect the valve vibration signal of each cylinder. The valve vibration signal collected by the i-th cylinder is denoted as Xi ={x i,1 ,x i,2 ,…,x i,k ,…,x i,N}, x i,k represents the sampling value of k sampling points in the i-th valve vibration signal, the sampling frequency is 25000 Hz, and each signal is collected for 250 seconds;

[0032] (2) Extract the impact signal from each valve vibration signal, and record the impact signal sequence extracted from the valve vibration signal of the i-th valve as {h i,1 ,h i,2 ,…,h i,l ,…,h i,n}, h i,l represents the lth impact signal extracted from the i-th valve vibration signal, and n is the total number of extracted impact signals;

[0033] In this embodiment, the key phase signal or strain signal can be combined to locate each impact in the valve vibration signal to achieve the extraction of the impact signal; the valve vibration signal can also be decomposed by using the same type of multi-source aliasing signal decoupling method to extract the impact signal generated by the valve closing.

[0034] In this embodiment, the total number of impact signals extracted from the valve vibration signal of each cylinder is 499, 496, and 500, which we unified to 496 impact signals;

[0035] Let's take a 5-second vibration signal of a valve as an example. The extracted impact signal is as follows: Figure 2 As shown, a total of 8 impact signals were extracted;

[0036] (3) Extract the valve impact phase of each cylinder;

[0037] (3.1), traverse the impulse signal sequence {h i,1 ,h i,2 ,…,h i,l ,…,h i,n}, search for each impact signal h i,l The peak value of , and record the timestamp corresponding to each peak value;

[0038] In this embodiment, an extreme value search algorithm can be used to search for each impact signal h i,l Search for the extreme value of and obtain the peak value. Figure 3 In the figure, the peak value of the shock signal is extracted from a ten-second vibration signal, and 18 shock peak values ​​are obtained;

[0039] (3.2) The timestamp corresponding to each peak is used as the valve impact phase of the periodic valve closing impact of the corresponding cylinder, where the valve impact phase of the i-th cylinder is recorded as Pi ={p i,1 ,p i,2 ,…,p i,l ,…,p i,n}, p i,l represents the lth valve impact phase of the i-th cylinder;

[0040] (4) Calculate the relative phase of valve impact between cylinders;

[0041] Taking the cylinder with the least number of impact signals among the two adjacent cylinders in the motion sequence as the benchmark, the valve closing impact phases of the two adjacent cylinders in the same cycle are subtracted to obtain the relative phase sequence X i+1,i ;

[0042] X i+1,i ={(p i+1,1 -p i,1 ),(p i+1,2 -p i,2 ),…,(p i+1,l -p i,l ),…,(p i+1,n -p i,n )}

[0043] The movement order between the cylinders of a certain type of five-cylinder drilling pump is cylinder 1, cylinder 4, cylinder 2, cylinder 5 and cylinder 3. The relative phase {X 4,1 ,X 2,4 ,X 5,2 ,X 3,5 ,X 3,1};

[0044] X 4,1 ={(p 4,1 -p 1,1 ),(p 4,2 -p 1,2 ),…,(p 4,l -p 1,l ),…,(p 4,n -p 1,n )}

[0045] X 2,4 ={(p 2,1 -p 4,1 ),(p 2,2 -p 4,2 ),…,(p 2,i -p 4,i ),…,(p 2,n -p 4,n )}

[0046]

[0047] X 3,1 ={(p3,1 -p 1,1 ),(p 3,2 -p 1,2 ),…,(p 3,i -p 1,i ),…,(p 3,n -p 1,n )}

[0048] In this embodiment, assuming that the number of impact signals extracted from two adjacent cylinders is 8 and 10 respectively, the last two impact signals extracted from the other cylinder are discarded based on the 8 impact signals.

[0049] (5) Calculate diagnostic indicators based on relative phase;

[0050] (5.1) In this embodiment, the sliding step is set to 8, and the relative phase sequence X is adjusted by using a sliding window according to a fixed step. i,i+1 After cutting, 61 relative phase samples are obtained, and the length of each relative phase sample is 16;

[0051] In this embodiment, the total number of impact signals extracted from the valve vibration signals of each cylinder is unified to 496 impact signals, so the relative phase sequence X i+1,i contains 496 phases, and then the relative phase sequence X is converted to i+1,i Cut into 61 relative phase samples, each relative phase sample has a length of 16;

[0052] (5.2), calculate the relative phase variance of each relative phase sample;

[0053]

[0054] Where k = 1, 2, ..., 61, represents the mth sample element in the kth relative phase sample, represents the element-wise mean of the kth relative phase sample; represents the relative phase variance of the kth relative phase sample;

[0055] (5.3), calculate the relative phase mean of each relative phase sample;

[0056]

[0057] in, represents the relative phase mean of the kth relative phase sample;

[0058] (6) Construct anomaly correlation matrix;

[0059] (6.1), construct the relative phase index matrix;

[0060] According to the calculated relative phase variance or mean index, construct an H×H symmetric matrix. The rows and columns of the matrix correspond to the cylinder numbers. The element in the i-th row and j-th column of the matrix is ​​the relative phase variance or mean index between cylinders i and j. The other elements are left blank, where i,j∈[1,H];

[0061] In this embodiment, if Figure 4 As shown, we construct the relative phase variance Figure 4 The relative phase indicator matrix shown on the left;

[0062] (6.2), calculate the anomaly correlation matrix;

[0063] Set the corresponding diagnostic threshold T. When the value of a certain indicator in the relative phase indicator matrix exceeds n times the threshold, it is considered abnormal, and the corresponding matrix element is set to 1, and the rest are set to 0. Finally, an abnormal correlation matrix consisting of 0 and 1 is constructed.

[0064] (7) Abnormality determination;

[0065] When the corresponding elements of a cylinder and its two adjacent cylinders in the abnormal correlation matrix are all 1, the cylinder is determined to be an abnormal cylinder and there is an abnormal phase fluctuation. Otherwise, the cylinder is determined to be a normal cylinder.

[0066] In this embodiment, taking three cylinders as an example, the two adjacent cylinders are cylinder one and cylinder five. Figure 4 As shown in the right figure, the corresponding elements in the abnormal correlation matrix are 13, 31, 35 and 53, and their elements are all 1, so cylinder 3 is determined to be an abnormal cylinder.

[0067] Example verification

[0068] In this example, an XZQ-2200 five-cylinder drilling pump test platform was used to obtain a dataset for validating the method presented in this paper. The experiment employed two pump stroke modes: normal, piston leakage, upper valve leakage, and lower valve leakage. The datasets are shown in Table 1.

[0069] Table 1 Experimental dataset

[0070] Failure type Failure location Pump strokes / (rpm) Sampling duration / (s) Sampling frequency / (Hz) Normal No 100、120 250 25000 Piston damage One cylinder, three cylinders 100、120 250 25000 Upper valve damage One cylinder, three cylinders 100、120 250 25000 Lower valve damage One cylinder 100、120 250 25000

[0071] This embodiment is tested on a device installed with the MATLAB R2022a environment, and anomaly detection is performed based on the experimental data set shown in Table 1. The accuracy is shown in Table 2. It can be seen from Table 2 that the method proposed in the present invention achieves a high anomaly detection accuracy.

[0072] Table 2 Anomaly detection results of our method

[0073] Failure type Accuracy / % Precision / % Recall / % F1-Score / % Upper valve leak failure 95.0 95.0 95.0 95.0 Lower valve leak failure 95.0 100.0 95.0 97.4 Piston failure 98.7 100.0 97.5 98.7 Average value 96.2 98.3 95.8 97.1

[0074] Although the above describes the illustrative specific embodiments of the present invention to facilitate understanding of the present invention by those skilled in the art, it should be clear that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the appended claims, these changes are obvious, and all inventions and creations using the concepts of the present invention are protected.

Claims

1. A valve impact phase diagnostic indicator and method for the hydraulic end of a multi-cylinder drilling pump, characterized in that: The following steps are involved: (1) Collect the vibration signals of valves of each cylinder at the hydraulic end of a multi-cylinder drilling pump; (1.1) Assume that a multi-cylinder drilling pump contains H cylinders. Number each cylinder according to the movement sequence between the cylinders of the drilling pump, and record it as i = 1, 2, ..., H; (1.2) A vibration sensor is arranged at the bottom of each cylinder to collect the valve vibration signal of each cylinder. The valve vibration signal collected by the i-th cylinder is denoted as X i ={x i,1 ,x i,2 ,…,x i,k ,…,x i,N }, x i,k represents the sampling value of k sampling points in the vibration signal of the i-th valve, and N represents the number of sampling points; (2) Extract the impact signal from each valve vibration signal, and record the impact signal sequence extracted from the valve vibration signal of the i-th valve as {h i,1 ,h i,2 ,…,h i,l ,…,h i,n }, h i,l represents the lth impact signal extracted from the i-th valve vibration signal, and n is the total number of extracted impact signals; (3) Extract the valve impact phase of each cylinder; (3.1), traverse the impulse signal sequence {h i,1 ,h i,2 ,…,h i,l ,…,h i,n }, search for each impact signal h i,l The peak value of , and record the timestamp corresponding to each peak value; (3.2) The timestamp corresponding to each peak is used as the valve impact phase of the periodic valve closing impact of the corresponding cylinder, where the valve impact phase of the i-th cylinder is recorded as P i ={p i,1 ,p i,2 ,…,p i,l ,…,p i,n }, p i,l represents the lth valve impact phase of the i-th cylinder; (4) Calculate the relative phase of valve impact between cylinders; Taking the cylinder with the least number of impact signals among the two adjacent cylinders in the motion sequence as the benchmark, the valve closing impact phases of the two adjacent cylinders in the same cycle are subtracted to obtain the relative phase sequence X i,i+1 ; X i,i+1 ={(p i,1 -p i+1,1 ),(p i,2 -p i+1,2 ),…,(p i,l -p i+1,l ),…,(p i,n -p i+1,n )} (5) Calculate the diagnostic index based on relative phase; (6) Construct anomaly correlation matrix; (6.1), construct the relative phase index matrix; Based on the calculated relative phase variance or mean index, construct an H×n symmetric matrix, where the rows and columns of the matrix correspond to the cylinder numbers. The element in the i-th row and j-th column of the matrix is ​​the relative phase variance or mean index between cylinders i and j, and the remaining elements are left blank. (6.2), calculate the anomaly correlation matrix; Set the corresponding diagnostic threshold T. When the value of a certain indicator in the relative phase indicator matrix exceeds n times the threshold, it is considered abnormal, and the corresponding matrix element is set to 1, and the rest are set to 0. Finally, an abnormal correlation matrix consisting of 0 and 1 is constructed. (7) Abnormality determination; When the corresponding elements of a cylinder and its two adjacent cylinders in the abnormal correlation matrix are all 1, the cylinder is determined to be an abnormal cylinder and there is an abnormal phase fluctuation. Otherwise, the cylinder is determined to be a normal cylinder.

2. The valve impact phase diagnostic index and method for the hydraulic end of a multi-cylinder drilling pump according to claim 1, characterized in that: The calculation method of the diagnostic index based on relative phase is: (2.1) Using a sliding window to adjust the relative phase sequence X according to a fixed step i,i+1 Perform cutting to obtain M relative phase samples, each of which has a length of L; (2.2), calculate the relative phase variance of each relative phase sample; Where k = 1, 2, ..., M, represents the mth sample element in the kth relative phase sample, represents the element-wise mean of the kth relative phase sample; represents the relative phase variance of the kth relative phase sample; (2.3), calculate the relative phase mean of each relative phase sample; in, represents the relative phase mean of the kth relative phase sample.

Images