Fracturing pump plunger fatigue life prediction method

Abstract

The invention relates to the technical field of equipment health management, and discloses a fracturing pump plunger fatigue life prediction method, which comprises the following steps that: an excitation unit performs multi-point scanning along a preset gridding path on a key bearing area on the surface of a fracturing pump plunger, a sensing unit synchronously collects sound-induced electromagnetic response signals generated by each scanning point due to sound wave excitation, calculates a local damage value of each scanning point according to the sound-induced electromagnetic response signals, and integrates and quantizes the local damage values into a global damage index representing the overall damage state of the plunger. And inputting the global damage index obtained in real time into a pre-established reference prediction model so as to predict the remaining service life of the current plunger. According to the method, in-situ and lossless accurate quantitative evaluation can be carried out on the fatigue damage state of the fracturing pump plunger, accurate prediction of the fatigue life of a core quick-wear part is achieved, a decision basis is provided for condition-based maintenance and preventive guarantee, and the safety and economical efficiency of equipment operation are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of health management, in particular to a fracturing pump plunger fatigue life prediction method. BACKGROUND

[0002] The fracturing pump is the core power equipment in the oil and gas field production operation, and the working condition is extremely harsh. It has been subjected to high pressure, strong impact and high frequency reciprocating load for a long time. The plunger as a key load-bearing and vulnerable part, its performance state directly determines the operation safety and efficiency. Under the continuous action of cyclic load, micro-damage initiation and accumulation will occur in the plunger material, and eventually evolve into fatigue cracks and lead to fracture failure.

[0003] Currently, in order to ensure the safety of the equipment, the regular replacement strategy based on fixed time length is mainly adopted. This method does not consider the actual damage evolution process of the plunger, often leading to the waste of the plunger with a long safe durability being replaced in advance, or failing to find the plunger with accelerated damage due to abnormal working conditions, thus leaving safety hazards.

[0004] In order to realize predictive maintenance based on the actual state, the industry attempts to introduce non-destructive testing technology. However, the existing technology has several defects that cannot be overcome, which is not consistent with the demand for accurate fatigue life prediction. Fatigue is the process from micro-damage accumulation to macro-crack formation. The traditional detection method (such as ultrasonic or magnetic powder) is not sensitive to the early micro-damage in the initiation stage of the material inside, which is distributed. It can only be effectively detected after the formation of macro-cracks. At this time, the plunger has been in a state of impending failure, losing the significance of early warning. The fatigue damage of the plunger is usually distributed in a planar manner in the key load-bearing area. The existing technology lacks an effective means to comprehensively evaluate the damage distribution of the entire area and quantify it as a single overall damage index that can be used for life prediction. Therefore, the existing technology cannot establish a reliable mapping relationship between the detected local and qualitative defect information and the overall macroscopic residual life of the component. The existing technology generally lacks an effective method to accurately quantify the early and micro fatigue damage of the fracturing pump plunger in situ and non-destructively, and to comprehensively evaluate the overall damage state of the key area, so as to realize accurate residual fatigue life prediction. SUMMARY

[0005] In view of the shortcomings of the prior art, the present application provides a fracturing pump plunger fatigue life prediction method, which solves the problems of discontinuous control trajectory, static and unadjustable emotion mapping, lack of user feedback closed loop, weak self-adaptability and insufficient fusion of multi-dimensional control information in the prior art.

[0006] To achieve the above purpose, the present application is implemented by the following technical scheme: a fracturing pump plunger fatigue life prediction method, comprising the following steps:

[0007] S1, using a central control processing unit deployed outside the fracturing pump to control an excitation unit also deployed outside the fracturing pump to emit a wide-band probe acoustic wave signal to a plunger of the fracturing pump, and using a sensing unit arranged around a key load-bearing area of the plunger to synchronously collect a plurality of groups of original acoustic-electromagnetic response signals excited by the wide-band probe acoustic wave signal in the material of the plunger;

[0008] S2, performing a matched filtering process on the plurality of groups of original acoustic-electromagnetic response signals collected in step S1 by the central control processing unit to obtain a group of acoustic-electromagnetic response signals with enhanced signal-to-noise ratio;

[0009] S3, performing a spectrum analysis process on the acoustic-electromagnetic response signals with enhanced signal-to-noise ratio obtained in step S2 by the central control processing unit to determine an optimal excitation frequency, which is a frequency corresponding to a maximum peak of a power spectral density of the acoustic-electromagnetic response signals with enhanced signal-to-noise ratio;

[0010] S4, controlling the excitation unit by the central control processing unit to perform a multi-point scanning on the key load-bearing area of the plunger at the optimal excitation frequency determined in step S3, and collecting a group of acoustic-electromagnetic response signals at each scanning point in the multi-point scanning by the sensing unit;

[0011] S5, performing a quantitative calculation on the group of acoustic-electromagnetic response signals at each scanning point collected in step S4 by the central control processing unit to obtain a global damage index representing a current overall damage degree of the plunger;

[0012] S6, substituting the global damage index obtained in step S5 into a pre-established reference prediction model to obtain a remaining service life of the plunger.

[0013] In a specific embodiment, the wide-band probe acoustic wave signal in step S1 is a linear frequency modulation signal, and the linear frequency modulation signal waveform s(t) can be expressed by the following formula:

[0014] s(t) = A cos(2p(f start t + (k / 2)t 2 ));

[0015] wherein s(t) is an instantaneous waveform of the signal at time t; A is a signal amplitude; f start is a preset initial frequency; t is a time variable; k is a rate of frequency change; the rate is determined by k = (f stop -f start ) / T; wherein, fstop where T is the duration of the signal.

[0016] Correspondingly, the matched filtering process in step S2, in technical effect, realizes pulse compression of the linear frequency modulation signal, and concentrates the energy of the wideband probe acoustic wave signal distributed in a long duration T into a pulse with a time width much smaller than T in the time axis, so as to form a pulse peak with higher amplitude and higher time resolution in the signal-to-noise ratio enhanced acoustic electromagnetism response signal.

[0017] In one specific embodiment, the spectrum analysis process in step S3 includes:

[0018] Firstly, Fourier transform is performed on the signal-to-noise ratio enhanced acoustic electromagnetism response signal, so as to obtain a power spectrum density function;

[0019] Then, the maximum peak point of the power spectrum density function is searched in the frequency coverage range (i.e. the interval from f start to f stop ) of the wideband probe acoustic wave signal;

[0020] Finally, the frequency corresponding to the maximum peak point is determined as the optimal excitation frequency.

[0021] In one specific embodiment, the multi-point scanning performed on the key load-bearing area of the plunger in step S4 includes the following steps:

[0022] Firstly, the central control processing unit controls the excitation unit to emit a single-frequency focused acoustic wave signal, and the frequency of the single-frequency focused acoustic wave signal is equal to the optimal excitation frequency;

[0023] Then, the focal point of the single-frequency focused acoustic wave signal is controlled to move along a preset gridded path of the key load-bearing area, so as to complete multi-point scanning of the key load-bearing area.

[0024] In one specific embodiment, the quantitative calculation in step S5 includes:

[0025] Firstly, signal energy integration operation is performed on the group of acoustic electromagnetism response signals V(t) at each scanning point collected in step S4 in a preset time integration window [t1, t2] (t1 and t2 are the starting time and the ending time of integration, respectively), so as to obtain a local damage value E local at the scanning point, and the calculation formula is as follows:

[0026] E local =∫[t1,t2]·|V(t)| 2• dt;

[0027] Then, the local damage values of all scanning points are accumulated and summed to obtain the global damage index GDI, and the calculation formula is:

[0028] GDI =∑ · E local i;

[0029] Wherein, E local i is the local damage value corresponding to the scanning point with index i; ∑ is the summation operation on all scanning points (i.e. on all indexes i).

[0030] In one specific embodiment, the reference prediction model in step S6 is established by the following steps:

[0031] First, at least one brand new fracturing pump plunger sample is subjected to an accelerated fatigue experiment, and the accelerated fatigue experiment has a total fatigue cycle number N total ;

[0032] Secondly, in the process of the accelerated fatigue experiment, the steps S1 to S5 of the present application are repeatedly executed according to the preset fatigue cycle interval ΔN, so as to obtain the global damage index GDI(N j ) corresponding to different damage stages, wherein N j is the current fatigue cycle number applied;

[0033] At the same time, when the global damage index GDI(N j ) is obtained each time, an actual remaining useful life value RUL j (N actual ) corresponding to the global damage index GDI(N j ) is determined by the following formula:

[0034] RUL actual (N j ) = N totaj -N j ;

[0035] Wherein, RUL actual (N j ) is the actual remaining useful life value corresponding to the fatigue cycle N j ; N total is the total fatigue cycle number determined in advance; N j is the current fatigue cycle number applied when the global damage index GDI(N j ) is obtained, and j is the ordinal number of measurement.

[0036] Finally, a function fitting is performed on the obtained series of global damage indexes and the determined series of actual residual life values to establish a function relationship RUL=F(GDI) between the global damage index GDI and the residual useful life RUL, and the function F is the reference prediction model. j , RUL actual (N j ).

[0037] In one embodiment, the method further comprises a pre-warning step:

[0038] The pre-warning step is performed after the determination of the residual useful life RUL of the plunger in step S6, and comprises:

[0039] comparing the residual useful life RUL with a preset maintenance threshold RUL threshold ; and, under the condition of RUL threshold , automatically generating a maintenance pre-warning signal.

[0040] Preferably, the excitation unit is configured to emit a wideband probe acoustic wave signal and a single-frequency acoustic wave signal to the plunger of the fracturing pump;

[0041] The sensing unit is configured to be arranged outside the key load-bearing area of the plunger in a wrap-around manner and is used to synchronously collect the acoustic-electromagnetic response signals;

[0042] The central control processing unit is connected with the excitation unit and the sensing unit and is configured to:

[0043] control the excitation unit to emit the wideband probe acoustic wave signal and control the sensing unit to collect the original acoustic-electromagnetic response signals;

[0044] perform matching filtering processing and spectrum analysis processing on the original acoustic-electromagnetic response signals to determine an optimal excitation frequency;

[0045] control the excitation unit to perform a multi-point scanning at the optimal excitation frequency and control the sensing unit to collect the acoustic-electromagnetic response signals in the multi-point scanning;

[0046] perform quantitative calculation on the collected signals in the multi-point scanning to obtain a global damage index;

[0047] substitute the global damage index into a pre-established reference prediction model to perform calculation, thereby determining the residual useful life of the plunger.

[0048] Preferably, the excitation unit is a wideband piezoelectric ultrasonic transducer, the operating frequency range of which covers the frequency range of the wideband detection acoustic wave signal in step S1, and the wideband piezoelectric ultrasonic transducer is also used to emit the acoustic wave signal with the frequency equal to the optimal excitation frequency in step S4.

[0049] Preferably, the sensing unit is a sensor array composed of a plurality of tunnel magnetoresistance sensors. The plurality of tunnel magnetoresistance sensors are arranged in the sensor array according to a preset geometric configuration, thereby providing spatial resolution for the calculation of the damage distribution, and each tunnel magnetoresistance sensor has a sensitivity of detecting magnetic field fluctuations on the order of nanotesla (nT).

[0050] Preferably, the excitation unit and the sensing unit are deployed in a non-contact manner, and in the deployed state, there is a physical gap between the excitation unit, the sensing unit and the surface of the fracturing pump.

[0051] The present application provides a fracturing pump plunger fatigue life prediction method. It has the following advantages:

[0052] 1. The present application dynamically determines the optimal excitation frequency under the current damage state by first performing wideband detection and spectrum analysis, overcoming the signal distortion or attenuation problem caused by the change of response characteristics due to material damage accumulation in the traditional fixed frequency detection method, ensuring that the subsequent multi-point scanning is always carried out under the best signal-to-noise ratio condition, thereby being able to more sensitively capture early micro-damage, and significantly improving the accuracy of damage assessment throughout the fatigue life cycle.

[0053] 2. The present application does not rely on single-point or local area detection, but performs multi-point scanning on the key load-bearing area of the plunger, and integrates and accumulates the local damage value of each scanning point, finally forming a global damage index, avoiding misjudgment caused by uneven damage distribution, so that the quantitative result can more comprehensively and stably reflect the overall health condition of the plunger, thereby providing a more reliable input for the baseline prediction model, and making the final remaining life prediction result more reliable.

[0054] 3. The excitation unit and the sensing unit of the present application are deployed outside the fracturing pump, and maintain a physical gap with the surface of the plunger. The detection can be completed without any mechanical disassembly. This non-contact, in-situ detection mode greatly simplifies the operation process, shortens the downtime required for detection, avoids the secondary damage risk caused by disassembly of the equipment, thereby greatly reducing the labor cost and time cost of maintenance, and improving the overall operation efficiency of the fracturing equipment. BRIEF DESCRIPTION OF DRAWINGS

[0055] Figure 1 The flowchart of the present application;

[0056] Figure 2 Deployment diagram for hardware units of the present invention. DETAILED DESCRIPTION

[0057] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the specification of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0058] Please refer to the drawings in the specification of the present application Figure 1 The embodiments of the present application provide a method for predicting fatigue life of a fracturing pump plunger, comprising the following steps:

[0059] The execution of the method for predicting fatigue life of the fracturing pump plunger is realized by the central control processing unit 10, the excitation unit 20 and the sensing unit 30 deployed outside the fracturing pump:

[0060] In step S1, the central control processing unit 10 controls the excitation unit 20 to emit a wideband probe acoustic wave signal to the plunger of the fracturing pump.

[0061] In one embodiment, the wideband probe acoustic wave signal is a linear frequency modulation signal, and the waveform s(t) of the linear frequency modulation signal can be expressed by the following formula:

[0062] s(t) = A cos(2p(f start ·t + (k / 2) t 2 ));

[0063] Wherein, s(t) is the instantaneous waveform of the signal at time t; A is the signal amplitude; f start is the preset starting frequency; t is the time variable; k is the rate of frequency change, which is determined by k = (f stop -f start ) / T; in the rate calculation formula, f stop is the preset ending frequency; T is the duration of the signal.

[0064] At the same time, the sensing unit 30 synchronously collects a plurality of groups of original acoustic electromagnetic response signals generated by the wideband probe acoustic wave signal exciting the acoustic electromagnetic effect inside the plunger material, and the sensing unit 30 is arranged in a wraparound manner outside the key load-bearing area of the plunger.

[0065] In step S2, the central control processing unit 10 performs a matched filtering process on the multiple sets of raw ACOUSTICALLY-GENERATED ELECTROMAGNETIC RESPONSE SIGNALS collected in step S1, which compresses the energy of the linear frequency modulation signal in the time axis, forms a pulse peak with higher amplitude and time resolution in the processed signal, and the result is a set of ACOUSTICALLY-GENERATED ELECTROMAGNETIC RESPONSE SIGNALS with enhanced signal-to-noise ratio.

[0066] In step S3, the central control processing unit 10 performs a spectral analysis process on the ACOUSTICALLY-GENERATED ELECTROMAGNETIC RESPONSE SIGNALS with enhanced signal-to-noise ratio obtained in step S2, which calculates a power spectral density function by performing Fourier transform on the ACOUSTICALLY-GENERATED ELECTROMAGNETIC RESPONSE SIGNALS with enhanced signal-to-noise ratio, and searches for the maximum peak point of the power spectral density function within the frequency coverage of the wideband detection acoustic signal. The frequency corresponding to the maximum peak point is determined as the optimal excitation frequency.

[0067] Then, a comprehensive damage scanning and quantification process based on the optimal frequency is performed. In step S4, the central control processing unit 10 controls the excitation unit 20 to perform a multi-point scanning on the key load-bearing area of the plunger using the optimal excitation frequency determined in step S3, and simultaneously collects a set of ACOUSTICALLY-GENERATED ELECTROMAGNETIC RESPONSE SIGNALS at each scanning point in the multi-point scanning using the sensing unit 30.

[0068] In step S5, the central control processing unit 10 performs a quantification calculation on the set of ACOUSTICALLY-GENERATED ELECTROMAGNETIC RESPONSE SIGNALS collected at each scanning point in step S4, to obtain a global damage index representing the current overall damage degree of the plunger. The quantification calculation first performs a signal energy integration operation on the signal of each scanning point within a preset time integration window, to obtain a local damage value E local for the scanning point.

[0069] E local =∫[t1,t2]·|V(t)| 2 ·dt;

[0070] where E local is the local damage value, V(t) is the ACOUSTICALLY-GENERATED ELECTROMAGNETIC RESPONSE SIGNAL collected at the scanning point, t is the time variable, and [t1,t2] is the preset time integration window.

[0071] Subsequently, the local damage values of all scanning points are added up to obtain the global damage index GDI. The calculation formula of the global damage index is:

[0072] GDI=Σ·E local i;

[0073] In the formula, GDI is the global damage index, and E locali is the local damage value corresponding to the scanning point with index i.

[0074] Finally, a model-based residual life prediction process is performed, in step S6, the global damage index obtained in step S5 is substituted into a pre-established reference prediction model for calculation, the reference prediction model is obtained by performing an accelerated fatigue experiment on the plunger sample and establishing a functional relationship between the global damage index and the actual residual life value by using a function fitting method, through calculation, the residual service life of the plunger is finally obtained

[0075] Referring to the drawings Figure 1 In step S1, the central control processing unit 10 controls the broadband probe acoustic wave signal emitted by the excitation unit 20, which is specifically implemented as a linear frequency modulation signal, the frequency of the linear frequency modulation signal changes linearly from a preset starting frequency f start to a preset ending frequency f stop .

[0076] In a specific embodiment, the starting frequency f start may be set (for example: 2 megahertz (MHz)), the ending frequency f stop may be set (for example: 8 megahertz (MHz)), and the signal duration T can be set (for example: 20 microseconds (μs)).

[0077] By using a linear frequency modulation signal with a large time bandwidth, a higher total energy acoustic wave can be injected into the plunger without increasing the peak power load of the excitation unit 20.

[0078] In step S2, the matched filter processing performed by the central control processing unit 10 is a digital signal processing operation, and the matched filter processing is implemented in the following manner:

[0079] Each set of original acousto-electromagnetic response signals collected in step S1 is convoluted with a preset reference signal, and the waveform of the reference signal is the time-reversed conjugate form of the linear frequency modulation signal s(t) emitted by the excitation unit 20.

[0080] The output signal y(t) of the matched filter processing can be calculated by the following formula:

[0081] y(t) = ∫r(τ)·s*(τ-t)dτ;

[0082] In the formula: y(t) is the matched filter processing output signal at time t; r(t) is the original acousto-electromagnetic response signal collected; s*(τ-t) is the time-reversed conjugate form of the linear frequency modulation signal waveform s(t); τ is the time integral variable.

[0083] The matched filter processing realizes pulse compression of the linear frequency modulation signal in technical effect.

[0084] The energy distribution of the original acoustic electromagnetic response signal 40 is in a long time width, and the signal-to-noise ratio is low. After the matched filter processing, the signal energy is concentrated on a pulse peak with extremely narrow time width, so that the acoustic electromagnetic response signal 50 with enhanced signal-to-noise ratio is obtained. The amplitude gain of the pulse compressed signal peak is proportional to the square root of the time bandwidth product of the linear frequency modulation signal. The acoustic electromagnetic response signal 50 with enhanced signal-to-noise ratio obtained by processing provides a basis for subsequent accurate spectrum analysis and determination of the optimal excitation frequency in step S3.

[0085] Referring to the accompanying drawings Figure 1 In step S3, the spectrum analysis processing performed by the central control processing unit 10 is implemented in the following specific manner:

[0086] First, the central control processing unit 10 applies a fast Fourier transform (FFT) algorithm to the one or more groups of acoustic electromagnetic response signals y(t) with enhanced signal-to-noise ratio obtained in step S2. The fast Fourier transform algorithm converts the signal from time domain data to frequency domain data to obtain a complex spectrum.

[0087] The transformation process can be expressed by the following formula:

[0088] Y(f) = ∫y(t)·e -j2πft dt;

[0089] In the formula: Y(f) is the complex spectrum of the signal y(t) at frequency f; f is the frequency variable; j is the imaginary unit; e -j2πft represents the complex exponential or transformation kernel; and dt represents an infinitesimal time interval.

[0090] Subsequently, the central control processing unit 10 calculates the power spectral density function P(f) of the signal based on the complex spectrum Y(f). The specific calculation method is to square the amplitude of each frequency component in the complex spectrum. The calculation can be expressed by the following formula:

[0091] P(f) = |Y(f)| 2 ;

[0092] In the formula: P(f) is the power spectral density at frequency f; and |Y(f)| is the amplitude of the complex spectrum Y(f) at frequency f.

[0093] The obtained power spectral density function is a real function, which describes the distribution of signal power at different frequencies.

[0094] Finally, the central control processing unit 10 performs a peak search on the power spectral density function, the frequency range of the peak search being limited to the frequency coverage of the broadband probe acoustic signal emitted at step S1 (e.g. from 2 to 8 megahertz).

[0095] The central control processing unit 10 identifies the maximum peak point of the power spectral density function within the limited frequency range by a numerical comparison algorithm, the frequency f opt corresponding to the maximum peak point being determined as the optimal excitation frequency, the determination process being expressible by the following formula:

[0096] f opt = argmax {f start ≤ f ≤ f stop} P(f);

[0097] In the formula: f opt is the determined optimal excitation frequency; argmax is a function that seeks the independent variable corresponding to the maximum value; f start and f stop are the start frequency and end frequency of the broadband probe acoustic signal respectively.

[0098] This frequency represents the frequency point at which the conversion efficiency from acoustic wave energy to electromagnetic response signal energy reaches a maximum value under the current material damage state of the plunger.

[0099] Referring to the accompanying drawings, Figure 1 In step S4, the central control processing unit 10 performs the multi-point scanning in the following specific manner:

[0100] Firstly, the central control processing unit 10 sends a set of control instructions to the excitation unit 20 according to the optimal excitation frequency f opt determined at step S3, and the excitation unit 20 generates a single-frequency focused acoustic signal according to the control instructions, the frequency of the single-frequency focused acoustic signal being equal to the optimal excitation frequency f opt .

[0101] The multi-point scanning is performed on a pre-defined critical load-bearing area 70 of the plunger, the critical load-bearing area 70 being determined according to finite element stress analysis results or empirical data of the plunger (e.g. the area on the surface of the plunger that bears the maximum cyclic bending stress).

[0102] The multi-point scanning is performed along a pre-defined gridded path 71, the gridded path 71 being composed of a two-dimensional array of scanning points 72 covering the critical load-bearing area 70, each scanning point 72 in the array having a unique coordinate.

[0103] In a specific embodiment, the spacing between the scanning points (i.e. the scanning step) can be set (e.g. 5 millimetres (mm)).

[0104] The central control processing unit 10 controls the focal point of the single-frequency focused acoustic wave signal to move from one scanning point 72 to the next scanning point 72 in a predetermined order until all scanning points on the gridded path 71 are scanned.

[0105] The movement can be achieved by controlling a two-dimensional precision displacement platform connected to the excitation unit 20, or in the case of the excitation unit 20 being a phased array transducer, by electronically regulating the focal point of the acoustic beam.

[0106] At each scanning point 72, the central control processing unit 10 performs a signal emission and collection operation:

[0107] The excitation unit 20 is controlled to emit a single-frequency focused acoustic wave signal pulse with a fixed duration, and the sensing unit 30 is instructed to collect a corresponding set of acoustic electromagnetism response signals within a preset time window;

[0108] The collected signal data is stored and associated with the coordinates of the current scanning point 72, and the process is repeated until the key load-bearing area 70 is fully scanned.

[0109] Referring to the accompanying drawings Figure 1 In step S5, the quantitative calculation performed by the central control processing unit 10 includes two consecutive calculation operations:

[0110] First, the central control processing unit 10 performs a calculation of the local damage value for the acoustic electromagnetism response signal collected at each scanning point in step S4.

[0111] The specific operation of this calculation is as follows:

[0112] A numerical integration operation is performed on the square of the amplitude of the signal waveform V(t) corresponding to each scanning point, and the integration operation is performed within a preset time integration window [t1, t2], with the start time t1 and the end time t2 of the time integration window being set to completely cover the duration of each valid acoustic electromagnetism response signal pulse. Through the integration operation, a local damage value E local for the scanning point is obtained, and the calculation formula for the local damage value is:

[0113] E local =∫[t1,t2]|V(t)| 2 dt;

[0114] In the formula: E local is the local damage value; V(t) is the acoustic electromagnetism response signal waveform collected at the scanning point; t is the time variable; and [t1, t2] is the preset time integration window.

[0115] Local damage value E local The physical dimension is energy. After calculating the local damage values ​​of all scanning points, the central control processing unit 10 will output the local damage values ​​E corresponding to all scanning points within the critical load-bearing area 70. local The values ​​of i are summed, and the summation result is defined as the Global Damage Index (GDI). The formula for calculating the Global Damage Index is as follows:

[0116] GDI=ΣE local i;

[0117] In the formula: GDI is the global damage index; E local i represents the local damage value corresponding to the scan point with index i.

[0118] The Global Damage Index (GDI) is a single scalar value used as an input parameter in step S6 to calculate the remaining useful life of the baseline prediction model.

[0119] See attached document Figure 1 In step S6, the central control processing unit 10 first uses a pre-established benchmark prediction model to calculate the remaining service life of the plunger. The establishment of the benchmark prediction model is an offline operation.

[0120] The operation begins by selecting at least one brand-new fracturing pump plunger sample and performing an accelerated fatigue test on the sample on a fatigue testing machine.

[0121] The experiment was conducted under a preset cyclic stress amplitude until the sample fractured, thereby determining the total fatigue life cycle number N of the sample. total .

[0122] During the accelerated fatigue test, steps S1 to S5 of the method of the present invention are repeated according to a preset fatigue cycle interval ΔN (e.g., every 50,000 cycles). Each execution yields a value corresponding to the current fatigue cycle number N. j The corresponding global damage index GDI(N) j At the same time, calculate the relationship between each GDI(N) j The corresponding actual remaining useful life value RUL actual (N j The formula for calculating the actual remaining life value is:

[0123] RUL actual (N j ) = N totaj -N j ;

[0124] Accelerated fatigue experiments were conducted to obtain a series of data pairs consisting of the Global Damage Index (GDI) and the actual remaining life value (N).j ),RUL actual (N j Using the acquired data pairs, the central control processing unit 10 employs a function fitting algorithm (e.g., least squares method) to establish the Global Damage Index (GDI) and the Actual Remaining Life Value (RUL). actual The functional relationship between them can be a polynomial function or an exponential function. The established function RUL = F(GDI) is the benchmark prediction model 90 and is stored in the memory of the central control processing unit 10.

[0125] In one embodiment, the functional relationship is a polynomial function, which can be expressed by the following formula:

[0126] RUL=F(GDI=a n ·GDI^n+a {n-1} ·GDI^{n-1}+...+a i ·GDI+a0;

[0127] In the formula: RUL is the calculated remaining useful life;

[0128] a i (where i = 0, 1, ..., n) are the polynomial coefficients determined by the function fitting algorithm;

[0129] n is the order of the polynomial.

[0130] In one embodiment, the functional relationship is an exponential function, which can be expressed by the following formula:

[0131] RUL=F(GDI)=α·e^(β·GDI);

[0132] In the formula: α and β are the coefficients of the exponential function determined by the function fitting algorithm; e is the base of the natural logarithm.

[0133] After determining the Remaining Usage Limit (RUL) using a baseline prediction model, the central control processing unit 10 executes an early warning step. This early warning step compares the RUL with a preset maintenance threshold RUL. threshold Perform numerical comparisons. Maintain the threshold RUL. threshold Set according to safety specifications or maintenance plans (e.g., set to total fatigue life cycles N). totaj 10% of the total.

[0134] When the calculated value of RUL is less than RUL thresholdWhen the preset value is reached, the central control processing unit 10 generates a maintenance warning signal. The specific form of the maintenance warning signal can be to display a warning message on a display device connected to the central control processing unit 10, or to output a high-level signal through a digital output port to drive an external audible and visual alarm.

[0135] See attached document Figure 2 The central control processing unit 10, the excitation unit 20 and the sensing unit 30 are deployed in a non-contact manner. The excitation unit 20 and the sensing unit 30 are mounted on the external structure of the fracturing pump through one or more fixing clamps 100, maintaining a preset physical gap d with the surface of the fracturing pump plunger 95.

[0136] In one embodiment, the distance of the physical gap d can be set (e.g., 10 mm), and the deployment method allows the entire detection process to be carried out in the gap during normal operation of the plunger without the need for mechanical disassembly of the fracturing pump, thereby achieving online, in-situ detection.

[0137] In one embodiment, the excitation unit 20 is a broadband piezoelectric ultrasonic transducer. When performing step S1, the central control processing unit 10 applies a broadband electrical excitation signal to the excitation unit 20, causing the excitation unit 20 to emit a broadband detection sound wave signal.

[0138] When performing step S4, the central control processing unit 10 applies a single-frequency electrical excitation signal with a frequency equal to the optimal excitation frequency to the excitation unit 20, causing the excitation unit 20 to emit a single-frequency focused acoustic wave signal.

[0139] The multi-point scanning is performed by mounting the excitation unit 20 on a two-dimensional precision displacement platform, and controlling the displacement platform by the central control processing unit 10, so that the excitation unit 20 moves according to a preset gridded path.

[0140] The specific implementation of the sensing unit 30 is a sensor array consisting of multiple tunnel magnetoresistive (TMR) sensors. The magnetic field detection sensitivity of the tunnel magnetoresistive sensors can reach the nanotesla (nT) level, and are used to detect weak magnetic field fluctuation signals generated by the acousto-electromagnetic effect.

[0141] Multiple tunnel magnetoresistive sensors are integrated on a circuit board substrate according to a preset two-dimensional rectangular array geometry. The array configuration enables the sensing unit 30 to simultaneously collect magnetic field signals at different spatial locations within the critical load-bearing area of ​​the plunger, providing spatially resolved data input for calculating local damage values ​​and generating a local damage value distribution map in the subsequent step S5.

[0142] The system in this embodiment can be used to execute the above algorithm embodiments, and its principle and technical effect are similar, so they will not be described again here.

Claims

1. A method for predicting the fatigue life of a fracturing pump plunger, characterized in that, Includes the following steps: S1. Using a central control processing unit deployed outside the fracturing pump, an excitation unit also deployed outside the fracturing pump is controlled to emit a broadband sound wave signal to the plunger of the fracturing pump. Using a sensing unit arranged in a surrounding manner outside the key load-bearing area of ​​the plunger, multiple sets of original acoustic electromagnetic response signals generated by the broadband sound wave signal excitation inside the plunger material are simultaneously acquired. S2. The central control processing unit performs a matched filtering process on the multiple sets of original acoustic electromagnetic response signals acquired in step S1 to obtain a set of acoustic electromagnetic response signals with enhanced signal-to-noise ratio. S3. The central control processing unit performs spectrum analysis on the signal-to-noise ratio enhanced acoustic electromagnetic response signal obtained in step S2 to determine an optimal excitation frequency. The optimal excitation frequency is the frequency at which the power spectral density of the signal-to-noise ratio enhanced acoustic electromagnetic response signal reaches its maximum peak value. S4. The central control processing unit controls the excitation unit to perform a multi-point scan of the critical load-bearing area of ​​the plunger through the optimal excitation frequency determined in step S3, and uses the sensing unit to collect a set of acoustic electromagnetic response signals at each scan point in the multi-point scan. S5. The central control processing unit performs quantization calculations on the set of acoustic electromagnetic response signals at each scanning point acquired in step S4 to obtain a global damage index characterizing the current overall damage level of the plunger. S6: Substitute the global damage index obtained in step S5 into a pre-established benchmark prediction model for calculation, thereby obtaining the remaining service life of the plunger.

2. The method for predicting the fatigue life of a fracturing pump plunger according to claim 1, characterized in that, In step S3, the spectrum analysis process includes: A power spectral density function is obtained by performing a Fourier transform on the signal-to-noise ratio enhanced acoustic electromagnetic response signal. And search for the maximum value and peak point of the power spectral density function within the frequency coverage range of the broadband sound wave signal; The frequency corresponding to the peak value is determined as the optimal excitation frequency.

3. The method for predicting the fatigue life of a fracturing pump plunger according to claim 1, characterized in that, In step S5, the quantization calculation includes: The set of acoustic electromagnetic response signals at each scanning point acquired in step S4 is integrated within a preset time integration window to obtain a local damage value corresponding to the scanning point. The local damage values ​​of all scan points are summed to obtain the global damage index.

4. The method for predicting the fatigue life of a fracturing pump plunger according to claim 1, characterized in that, In step S6, the baseline prediction model is established through the following steps: An accelerated fatigue test is performed on at least one brand-new fracturing pump plunger sample, the accelerated fatigue test having a total fatigue life cycle number predetermined by independent experiment or theoretical calculation; During the accelerated fatigue experiment, steps S1 to S5 are repeated at preset fatigue cycle intervals to obtain the global damage index corresponding to different damage stages. Each time the global damage index is acquired, the predetermined total fatigue life cycle count is subtracted from the currently applied fatigue cycle count to determine an actual remaining life value corresponding to the global damage index. A function is fitted using multiple data pairs consisting of the global damage index corresponding to the different damage stages and the actual remaining life value corresponding to the different damage stages. This establishes a functional relationship between the global damage index and the remaining useful life, and enables the establishment of a benchmark prediction model.

5. The method for predicting the fatigue life of a fracturing pump plunger according to claim 1, characterized in that, In step S6, after determining the remaining service life of the plunger, a warning step is executed, which includes: The remaining service life is compared with a preset maintenance threshold; When the remaining service life is less than the preset maintenance threshold, a maintenance warning signal is automatically generated.

6. The method for predicting the fatigue life of a fracturing pump plunger according to claim 1, characterized in that, In step S4, a multi-point scan is performed on the critical load-bearing area of ​​the plunger, including the following steps: The central control processing unit controls the excitation unit to emit a single-frequency focused acoustic wave signal, the frequency of which is equal to the optimal excitation frequency. The system controls the focal point of the single-frequency focused acoustic signal to move along a preset gridded path in the critical load-bearing area, thereby achieving multi-point scanning of the critical load-bearing area.

7. The method for predicting the fatigue life of a fracturing pump plunger according to claim 1, characterized in that, The excitation unit is a broadband piezoelectric ultrasonic transducer, which is configured to transmit the broadband probe acoustic signal in step S1 and to transmit an acoustic signal with a frequency equal to the optimal excitation frequency required for the multi-point scanning in step S4.

8. The method for predicting the fatigue life of a fracturing pump plunger according to claim 1, characterized in that, The sensing unit is a sensor array consisting of multiple tunneling magnetoresistive sensors, which are arranged in a preset geometric configuration to provide the spatial resolution required for damage mapping. Furthermore, each tunnel magnetoresistive sensor possesses the sensitivity to detect magnetic field fluctuations on the order of nanotesla.

9. The method for predicting the fatigue life of a fracturing pump plunger according to claim 1, characterized in that, Both the excitation unit and the sensing unit are deployed in a non-contact manner, and a physical gap is maintained between the excitation unit and the sensing unit and the surface of the fracturing pump.

10. The method for predicting the fatigue life of a fracturing pump plunger according to claim 1, characterized in that, In step S1, the broadband detection acoustic signal is a linear frequency modulated signal; Furthermore, the matched filtering process in step S2 compresses the energy of the linear frequency modulated signal on the time axis, thereby forming a pulse peak with higher amplitude and time resolution in the acoustic electromagnetic response signal with enhanced signal-to-noise ratio.

Images