Risk assessment method and system for three-frequency fusion of slurry pump and electronic equipment

Abstract

The invention provides a mud pump three-frequency fusion risk assessment method and system and electronic equipment. The method comprises the steps that vibration signals of all measuring point positions in the healthy operation state of a mud pump are monitored; variational mode decomposition is carried out on the vibration signals, mode components are obtained, and permutation entropies of the mode components are extracted to form an evaluation feature vector set; constructing an SVDD hypersphere model based on the evaluation feature vector set; and monitoring operation vibration signals of each measuring point position of the to-be-tested slurry pump in a working state, analyzing the operation vibration signals by adopting the SVDD hypersphere model, and performing risk assessment on each measuring point position of the to-be-tested slurry pump. According to the three-frequency fusion risk assessment method for the slurry pump, the SVDD hypersphere model is constructed on the basis of the vibration signals of all the measuring point positions, the operation vibration signals of all the measuring point positions of the to-be-measured slurry pump are monitored, and real-time comprehensive monitoring of the running state of the slurry pump and rapid and accurate identification of potential risks are achieved.

Description

Technical Field

[0001] The present invention belongs to the technical field of mud pumps, and particularly relates to a risk assessment method, system and electronic device for three-frequency fusion of mud pumps. Background Art

[0002] The mud pump is a key equipment in the drilling fluid circulation system. The mud pump has a complex structure and is difficult to manufacture and process. During the drilling process, the failures of vulnerable parts in the mud pump include: damage of the bearing bush, wear of the cylinder block and crosshead, burnout of the motor, etc., which will affect the stable and safe operation of the entire mud pump system. Therefore, it is of great engineering significance to monitor and identify the risks of the mud pump.

[0003] Currently, in the research of mud pump monitoring systems, existing research mainly focuses on the risk diagnosis and life prediction of key components of mud pumps. For example, the on-site vibration signals of the drilling pump are collected, the characteristic signals are extracted through data preprocessing, and then the probability neural network is used to identify the faults of the drilling pump. However, these methods can often only identify risks for a specific part of the mud pump, and it is difficult to comprehensively grasp the operating state and risk characteristics of the mud pump from the system level.

[0004] As a result, the current mud pump monitoring system cannot real-time monitor the performance indicators of the mud pump through various sensors such as pressure sensors, flow sensors, temperature sensors and vibration sensors, and cannot perform risk assessment of the mud pump in a timely, rapid and accurate manner. Therefore, how to comprehensively monitor the drilling mud pump is a key technical problem that needs to be solved urgently in this field. Summary of the Invention

[0005] In view of the above problems, the present invention proposes a risk assessment method for three-frequency fusion of mud pumps, including the following steps:

[0006] Monitor the vibration signals at each measuring point during the healthy operation of the mud pump;

[0007] Perform variational mode decomposition on the vibration signals to obtain modal components, and extract the permutation entropy of the modal components to form an evaluation feature vector set;

[0008] Construct an SVDD hypersphere model based on the evaluation feature vector set;

[0009] Monitor the operation vibration signals at each measuring point during the operation of the mud pump to be measured, analyze the operation vibration signals using the SVDD hypersphere model, and perform risk assessment on each measuring point of the mud pump to be measured.

[0010] Furthermore, before performing variational mode decomposition on the vibration signals to obtain modal components, it further includes:

[0011] Preset the key parameters of the variational mode, and use the particle swarm optimization algorithm to complete the optimization of the key parameters of the variational mode;

[0012] Among them, the key parameters of the variational mode include: the number of modal decomposition layers and the penalty factor.

[0013] Furthermore, perform variational mode decomposition on the vibration signal and obtain modal components, and extract the permutation entropy of the modal components to form an evaluation feature vector set, including the following steps:

[0014] Use variational mode to decompose the vibration signal into the first intrinsic mode components, and screen out the sensitive intrinsic mode components;

[0015] Construct a denoised original signal based on the sensitive intrinsic mode components;

[0016] Use variational mode decomposition to decompose the denoised original signal into the second intrinsic mode components;

[0017] Calculate the permutation entropy of the second intrinsic mode components, and extract the feature vectors of the vibration signals at each measurement point;

[0018] Combine the feature vectors in the order of the measurement points to form an evaluation feature vector set.

[0019] Furthermore, screening out the sensitive intrinsic mode components includes the following steps:

[0020] Calculate the Pearson correlation coefficient between each first intrinsic mode component and the original signal;

[0021] Screen out the first intrinsic mode components with Pearson correlation coefficients greater than the threshold as sensitive intrinsic mode components.

[0022] Furthermore, before constructing the SVDD hypersphere model based on the evaluation feature vector set, it also includes:

[0023] Preset the key parameters of the SVDD hypersphere model, and use the particle swarm optimization algorithm to complete the optimization of the key parameters of the SVDD hypersphere model;

[0024] Among them, the key parameters of the SVDD hypersphere model include: the penalty factor and the kernel function parameter.

[0025] Furthermore, use the SVDD hypersphere model to analyze the operation vibration signal and complete the risk assessment of each measurement point position of the mud pump to be measured, including the following steps:

[0026] Perform variational mode decomposition on the operation vibration signals at each measurement point position of the mud pump to be measured and obtain modal components, and form a risk feature vector set;

[0027] Input the risk feature vector set into the SVDD hypersphere model to form sample points, and calculate the distances from the sample points to the center of the hypersphere;

[0028] Refer to the 3σ principle to convert the distances into health degrees for the risk assessment of the measured points of the mud pump to be tested.

[0029] The present invention also proposes a risk assessment system for three-frequency fusion of mud pumps, including:

[0030] A data collection unit for monitoring the vibration signals at each measured point during the healthy operation of the mud pump;

[0031] A data sorting unit for performing variational mode decomposition on the vibration signals to obtain modal components, and extracting the permutation entropy of the modal components to form an evaluation feature vector set;

[0032] A model construction unit for constructing an SVDD hypersphere model based on the evaluation feature vector set;

[0033] A risk assessment unit for monitoring the operation vibration signals at each measured point during the operation of the mud pump to be tested, analyzing the operation vibration signals using the SVDD hypersphere model, and completing the risk assessment of each measured point of the mud pump to be tested.

[0034] Furthermore, it further includes:

[0035] A parameter optimization unit for presetting the key parameters of the SVDD hypersphere model and using the particle swarm optimization algorithm to complete the optimization of the key parameters of the SVDD hypersphere model;

[0036] The parameter optimization unit is also used for presetting the key parameters of the SVDD hypersphere model and using the particle swarm optimization algorithm to complete the optimization of the key parameters of the SVDD hypersphere model.

[0037] Furthermore, the data sorting unit is specifically used for:

[0038] Using variational mode to decompose the vibration signal into the first intrinsic mode function and screening out the sensitive intrinsic mode function components;

[0039] Constructing a denoised original signal based on the sensitive intrinsic mode function components;

[0040] Using variational mode decomposition to decompose the denoised original signal into the second intrinsic mode function;

[0041] Calculating the permutation entropy of the second intrinsic mode function components and extracting the feature vectors of the vibration signals at each measured point;

[0042] Combining the feature vectors in the order of the measured points to form an evaluation feature vector set.

[0043] The present invention also provides an electronic device, which includes at least one processor and at least one memory. The memory is data-connected to the processor. Among them,

[0044] the memory stores instructions executable by at least one processor. The instructions are executed by at least one processor so that at least one processor can execute the above-mentioned method.

[0045] Compared with the prior art, the embodiments of the present invention have at least the following advantages:

[0046] The risk assessment method for triple-frequency fusion of the mud pump in the present invention is based on the vibration signals at each measuring point position under the healthy operating state of the mud pump, and an evaluation feature vector set is formed for use in constructing the SVDD hypersphere model. During the specific use process, the operating vibration signals at each measuring point position under the working state of the mud pump to be measured are monitored, and the analysis and monitoring of the internal environment vibration signals of the mud pump to be measured are completed through the SVDD hypersphere model, realizing the risk assessment of each measuring point position of the mud pump to be measured, avoiding the problems such as incomplete single-frequency monitoring and inaccurate risk identification existing in the traditional mud pump monitoring, and further clarifying the operating states of each component of the mud pump to be measured. It realizes the real-time and comprehensive monitoring of the operating state of the mud pump and the rapid and accurate identification of potential risks, can continuously provide guarantee for the safe and stable operation of the mud pump, greatly improves the monitoring accuracy while enhancing the risk identification ability.

[0047] Other features and advantages of the present invention will be described in the following specification, and part of them will become obvious from the specification or be understood by implementing the present invention. The objectives and other advantages of the present invention can be achieved and obtained through the structures pointed out in the specification and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0049] Figure 1 shows a schematic flow chart of the risk assessment method for triple-frequency fusion of the mud pump in the embodiments of the present invention;

[0050] Figure 2 shows a schematic flow chart of optimizing the variational mode in the embodiments of the present invention;

[0051] Figure 3 shows a schematic flow chart of decomposing vibration signals by variational mode decomposition in the embodiments of the present invention;

[0052] Figure 4 Shows the schematic flow chart of optimizing the SVDD hypersphere model in the embodiment of the present invention;

[0053] Figure 5 Shows the schematic diagram of the operation and maintenance structure of the risk assessment system for the three-frequency fusion of the mud pump in the embodiment of the present invention;

[0054] Figure 6 Shows the schematic diagram of the structural layout of the mud pump in the embodiment of the present invention;

[0055] Figure 7 Shows the schematic diagram of the interface of the upper computer control system in the embodiment of the present invention;

[0056] Figure 8 Shows the schematic diagram of the convergence curve of the fitness value of the pump valve inlet sample in the embodiment of the present invention;

[0057] Figure 9 Shows the schematic diagram of the convergence curve of the fitness value of the crosshead sample in the embodiment of the present invention;

[0058] Figure 10 Shows the schematic diagram of the convergence curve of the fitness value of the crankshaft sample in the embodiment of the present invention;

[0059] Figure 11 Shows the schematic diagram of the convergence curve of the fitness value of the drive motor output end sample in the embodiment of the present invention;

[0060] Figure 12 Shows the example diagram of the distance from the sample point of the evaluation feature vector set to the center of the hypersphere in the embodiment of the present invention;

[0061] Figure 13 Shows the example diagram of the distance from the sample point of the risk feature vector set to the center of the hypersphere in the embodiment of the present invention;

[0062] Figure 14 Shows the schematic diagram of the health degree in the embodiment of the present invention;

[0063] Figure 15 Shows the block diagram of the risk assessment system for the three-frequency fusion of the mud pump in the embodiment of the present invention.

[0064] In the figure, flow sensor 1, suction valve temperature and vibration sensor 2, discharge valve temperature and vibration sensor 3, left acoustic emission sensor 4, piston cylinder temperature and vibration sensor 5, crosshead temperature and vibration sensor 6, temperature sensor 7, transmission shaft temperature and vibration sensor 8, crankshaft temperature and vibration sensor 9, risk control cabinet 10, upper computer control system 11, right acoustic emission host 12, calculation gateway 13, hub 14, acoustic emission sensor 15, PLC controller 16, frequency conversion controller 17, pressure sensor 18, frequency conversion control cabinet 19, right crankshaft bearing 20, left crankshaft bearing 21, left transmission shaft bearing 22, right transmission shaft bearing 23, lubricating oil temperature 24, right crosshead 25, middle crosshead 26, left crosshead 27, right piston cylinder 28, middle piston cylinder 29, left piston cylinder 30, right discharge valve 31, middle discharge valve 32, left discharge valve 33, right suction valve 34, middle suction valve 35, left suction valve 36, suction pipe 37, discharge pipe 38, right acoustic emission point 39, left acoustic emission point 40, three-dimensional model area 41, acquisition configuration area 42, risk early warning area 43, downhole configuration area 44. Detailed implementation manners

[0065] The following description provides many different embodiments or examples for implementing different features of the present invention. The elements and arrangements described in the following specific examples are only used to concisely express the present invention, and are only examples, rather than limiting the present invention.

[0066] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0067] The present invention provides a risk assessment method for three-frequency fusion of a mud pump. Figure 1 The flow chart of the risk assessment method for three-frequency fusion of the mud pump in the embodiment of the present invention is shown, as Figure 1 shown, the risk assessment method includes:

[0068] S101. Monitor the vibration signals at each measuring point position under the healthy operation state of the mud pump;

[0069] S102. Perform variational mode decomposition on the vibration signals to obtain modal components, and extract the permutation entropy of the modal components to form an evaluation feature vector set;

[0070] S103. Construct an SVDD hypersphere model based on the evaluation feature vector set;

[0071] S104. Monitor the operation vibration signals at each measuring point position under the working state of the mud pump to be measured, and use the SVDD hypersphere model to analyze the operation vibration signals to complete the risk assessment of each measuring point position of the mud pump to be measured.

[0072] It should be added that in this embodiment, the SVDD (Support Vector Data Description) hypersphere model is a machine learning method for data anomaly detection and classification. It is a one-class classification method based on the idea of support vector machine (SVM). Its main goal is to find a smallest hypersphere that can enclose as many normal samples in the dataset as possible while excluding abnormal samples.

[0073] The risk assessment method for three-frequency fusion of the mud pump of the present invention is based on the vibration signals at each measuring point position under the healthy operation state of the mud pump, and forms an evaluation feature vector set for use in constructing the SVDD hypersphere model. In the specific use process, monitor the operation vibration signals at each measuring point position under the working state of the mud pump to be measured, and complete the analysis and monitoring of the internal environment vibration signals of the mud pump to be measured through the SVDD hypersphere model, realizing the risk assessment of each measuring point position of the mud pump to be measured, avoiding the problems of incomplete single-frequency monitoring and inaccurate risk identification existing in traditional mud pump monitoring, and then clarifying the operation states of each component of the mud pump to be measured, realizing the real-time and comprehensive monitoring of the operation state of the mud pump and the rapid and accurate identification of potential risks, being able to continuously provide guarantee for the safe and stable operation of the mud pump, greatly improving the monitoring accuracy while enhancing the risk identification ability.

[0074] In addition, Variational Mode Decomposition (VMD) is a non-stationary signal analysis method used to decompose complex signals into several finite Intrinsic Mode Functions (IMFs).

[0075] Correspondingly, referring to Figure 2 , before performing variational mode decomposition VMD on the vibration signal and obtaining the modal components, it further includes:

[0076] Preset the key parameters of variational mode decomposition VMD, and use the Particle Swarm Optimization (PSO) algorithm to complete the optimization of the key parameters of variational mode decomposition VMD;

[0077] Among them, the key parameters of variational mode decomposition VMD include: the number of modal decomposition layers K and the penalty factor α;

[0078] Specifically, it includes the following steps:

[0079] Set a number of modal decomposition layers \(K\) and penalty factor \(\alpha\) for variational mode decomposition (VMD), and combine the modal decomposition layer \(K\) and the penalty factor \(\alpha\) to form a number of parameter combinations \([K, \alpha]\) to be optimized.

[0080] Initialize the particle swarm and assign the number of parameter combinations \([K, \alpha]\) to a number of particles one by one.

[0081] Perform variational mode decomposition (VMD) on a number of particles, and calculate the permutation entropy (PE) of different decomposed particles.

[0082] Select the minimum value of the permutation entropy (PE) as the fitness value \(MPE\).

[0083] Iterate the above steps and update the fitness value \(MPE\).

[0084] Until the maximum number of iterations reaches the threshold, output the fitness value \(MPE\) and the parameter combination \([K, \alpha]\) of the corresponding particle, and complete the optimization of the key parameters of variational mode decomposition (VMD).

[0085] Among them, the threshold corresponding to the maximum number of iterations is 100 times.

[0086] Among them, refer to Figure 3 , perform variational mode decomposition (VMD) on the vibration signal and obtain the modal components, and extract the permutation entropy of the modal components to form an evaluation feature vector set, which specifically includes the following steps:

[0087] Use variational mode decomposition (VMD) to decompose the vibration signal into the first intrinsic mode function (IMF1), and screen out the sensitive intrinsic mode functions.

[0088] Construct a noise-reduced original signal based on the sensitive intrinsic mode functions.

[0089] Use variational mode decomposition (VMD) to denoise the original signal into the second intrinsic mode function (IMF2).

[0090] Calculate the permutation entropy (PE) of the second intrinsic mode function (IMF2), and extract the feature vectors of the vibration signals at each measuring point.

[0091] Combine the feature vectors in the order of the measuring points to form an evaluation feature vector set.

[0092] Among them, screening out the sensitive intrinsic mode functions specifically includes the following steps:

[0093] Calculate the Pearson correlation coefficient (PCC) between each first intrinsic mode function (IMF1) and the original signal.

[0094] Select the first intrinsic mode function (IMF1) with a Pearson correlation coefficient (PCC) greater than the threshold as the sensitive intrinsic mode function.

[0095] The linear relationship between each first intrinsic mode function IMF1 and the original signal is constructed through the Pearson correlation coefficient, improving the relationship recognition between the first intrinsic mode function IMF1 and the original signal.

[0096] It should be supplemented that the Pearson correlation coefficient is an index to measure the linear correlation between two variables. In this embodiment, the threshold is set to 0.3. When the Pearson correlation coefficient is greater than 0.3, it indicates that the corresponding first intrinsic mode function IMF1 is a sensitive intrinsic mode function, and at the same time indicates that the linear correlation between the first intrinsic mode function IMF1 and the original signal is stronger.

[0097] Among them, referring to Figure 4 the example shown, before constructing the SVDD hypersphere model based on the evaluation feature vector set, it further includes:

[0098] Preset the key parameters of the SVDD hypersphere model, and use the particle swarm optimization (PSO) algorithm to complete the optimization of the key parameters of the SVDD hypersphere model;

[0099] Among them, the key parameters of the SVDD hypersphere model include: the penalty factor C and the kernel function parameter s.

[0100] The optimization of the key parameters of the SVDD hypersphere model specifically includes the following steps:

[0101] Preset the penalty factor C and the kernel function parameter s of several SVDD hypersphere models, and combine the penalty factor C and the kernel function parameter s to form several parameter combinations [C, s] to be optimized;

[0102] Assign the parameter combinations [C, s] to be optimized to the particles and use them as samples to train the SVDD hypersphere model;

[0103] Calculate the fitness value F of all particles and screen the maximum value as the best model fitness value;

[0104] Perform iterative calculations and update the best model fitness value and the corresponding key parameters; specifically, after each iterative calculation, compare the fitness values and screen the maximum value as the best model fitness value;

[0105] Until the number of iterations reaches the threshold, output the parameter combination corresponding to the best model fitness value as the key parameters of the SVDD hypersphere model, thereby improving the classification performance of the SVDD model.

[0106] Among them, the iteration threshold is 100 times.

[0107] Among them, the calculation formula of the fitness value F of the particle is as follows:

[0108]

[0109] Among them, y represents the correct classification number of the test samples; N represents the total number of test samples, and F is the fitness value, representing the classification accuracy rate, with a range of [0, 1].

[0110] It should be noted that the larger the fitness value, the better the optimization effect of the current C and s on the performance of the SVDD model.

[0111] Correspondingly, the SVDD hypersphere model is used to analyze the operation vibration signal to complete the risk assessment of each measuring point position of the mud pump to be tested, specifically including the following steps:

[0112] Perform variational mode decomposition (VMD) on the operation vibration signals at each measuring point position under the working state of the mud pump to be tested and obtain the modal components, and form a risk feature vector set;

[0113] Input the risk feature vector set into the SVDD hypersphere model to form sample points, and calculate the distance D from the sample points to the center of the hypersphere;

[0114] Refer to the 3-sigma rule to convert the distance D into the health degree H to complete the risk assessment of the measuring point position of the mud pump to be tested, and map the complex distance or features into an intuitive health degree for easy understanding and monitoring.

[0115] In this embodiment, it is set that the value range of the health degree H is 1 - 100, representing the degree of the health state of the equipment; correspondingly, the closer H is to 100, the healthier the equipment is; on the contrary, the closer H is to 0, the more abnormal the equipment is.

[0116] To more clearly explain the present disclosure, the following will be described in conjunction with the operation and maintenance architecture of the risk assessment system for the three-frequency fusion of the mud pump. As Figure 5 shows the schematic diagram of the operation and maintenance structure of the risk assessment system for the three-frequency fusion of the mud pump according to the embodiment of the present disclosure.

[0117] The operation and maintenance structure of this transponder transmission system may include: flow sensor 1, suction valve temperature and vibration sensor 2, discharge valve temperature and vibration sensor 3, left acoustic emission sensor 4, piston cylinder temperature and vibration sensor 5, crosshead temperature and vibration sensor 6, temperature sensor 7, transmission shaft temperature and vibration sensor 8, crankshaft temperature and vibration sensor 9, risk control cabinet 10, upper computer control system 11, right acoustic emission host 12, calculation gateway 13, hub 14, acoustic emission sensor 15, PLC controller 16, frequency conversion controller 17, pressure sensor 18, frequency conversion control cabinet 19.

[0118] In addition, referring to Figure 6 , Figure 6A schematic diagram of the mud pump structure for installing a risk assessment system for triple-frequency fusion of a mud pump is shown. The mud pump includes: a right crankshaft bearing 20, a left crankshaft bearing 21, a left transmission shaft bearing 22, a right transmission shaft bearing 23, a lubricating oil temperature 24, a right crosshead 25, a middle crosshead 26, a left crosshead 27, a right piston cylinder 28, a middle piston cylinder 29, a left piston cylinder 30, a right discharge valve 31, a middle discharge valve 32, a left discharge valve 33, a right suction valve 34, a middle suction valve 35, a left suction valve 36, a suction pipe 37, a discharge pipe 38, a right acoustic emission monitoring 39, and a left acoustic emission point 40.

[0119] However, it should be noted that the disclosed embodiments of the present invention are not limited to such an operation and maintenance structure, nor do they necessarily include bearings and crossheads, etc. As long as the sensor (such as the suction valve temperature vibration sensor 2) obtains the required parameter data (such as the vibration data of the right suction valve 34, the middle suction valve 35, and the left suction valve 36, etc.), the analysis of the parameter data can be realized. Figure 5 and Figure 6 The structure is only for exemplary explanation of the present disclosure and is not limited to such a system.

[0120] Among them, two high-sensitivity acoustic emission sensors are arranged on the mud pump valve cover, namely the left acoustic emission sensor 4 and the right acoustic emission sensor 15, which are used to capture the elastic waves in the mud pump environment in real time. The piezoelectric ceramics inside the sensor convert the mechanical vibration of the elastic wave into an electrical signal, and then the signal is amplified by the acoustic emission amplifier to enhance the signal strength and clarity. The signal is transmitted to the acoustic emission host 12 in the risk control cabinet 10 through the 4G network. The acoustic emission host 12 is responsible for receiving and processing the amplified acoustic emission signal to improve the signal quality and facilitate subsequent processing. It is connected to the data acquisition card through a dedicated signal transmission cable. The data acquisition card transmits the processed acoustic emission signal to the host computer control system 11 for in-depth acoustic emission signal feature analysis and storage.

[0121] The suction valve temperature and vibration sensor 2 is deployed in the mud pump monitoring area to monitor the temperature and vibration of the suction valve part of the mud pump, so as to timely detect temperature and vibration abnormalities caused by seal wear, valve core movement obstruction, etc., and avoid insufficient suction affecting work efficiency;

[0122] Deploy a discharge valve temperature and vibration sensor 3. By monitoring the temperature, it can prevent overheating caused by excessive discharge pressure, abnormal friction between the valve core and the valve seat, etc. Vibration monitoring can promptly handle abnormal vibration caused by abnormal opening and closing movements and internal parts problems, ensuring stable discharge and discharge pressure;

[0123] Deploy the piston cylinder temperature and vibration sensor 5 to detect temperature changes caused by friction between the piston and the cylinder wall, wear of the seal, etc. Vibration monitoring can reflect problems such as the clearance between the piston and the cylinder wall and the state of the piston ring, ensuring the normal reciprocating motion of the piston;

[0124] Deploy the crosshead temperature and vibration sensor 6 to monitor the mud pump crosshead, and can promptly detect vibration and temperature change problems caused by loose connecting parts of the crosshead, uneven force, etc., ensuring normal power transmission and motion conversion;

[0125] Deploy the drive shaft temperature and vibration sensor 8 and the crankshaft temperature and vibration sensor 9, aiming at the drive shaft bearing and the crankshaft bearing respectively, and simultaneously monitor the temperature and vibration of the environment, and can promptly detect problems such as poor bearing lubrication and component wear,

[0126] The above sensors transmit data through the NB-IOT network technology, and send the collected temperature and vibration data to the computing gateway 13 in the risk control cabinet 10 for preliminary processing and analysis, reducing the data transmission volume and improving the processing efficiency, and then uploading it to the host computer control system 11 for further analysis and storage.

[0127] In addition, a flow sensor 1, a temperature sensor 7 and a pressure sensor 18 are also configured to comprehensively monitor environmental parameters and flow information, transmit data to the PLC controller 16 through a 4-20mA signal line, the PLC controller 16 exchanges data through an intelligent gateway, and the intelligent gateway transmits the data to the host computer control system 11 for storage, processing and visualization operations.

[0128] Among them, in Figure 7 the example shown, the interface of the host computer control system 11 includes: a three-dimensional model area 41, a collection configuration area 42, a risk warning area 43 and a downhole configuration area 44;

[0129] The three-dimensional model area 41 is displayed in the center of the host computer interface, the collection configuration area 42 is on the left, and the risk warning area 43 and the downhole configuration area 44 are on the right.

[0130] Among them, the three-dimensional model area 41 realizes the all-round and high-precision visualization of the mud pump. This system also incorporates an intelligent risk warning mechanism: when a potential risk or abnormal condition occurs in a certain module or its subordinate components in the mud pump, the module will be immediately automatically identified by the system and immediately prompted by means of highlighting.

[0131] The acquisition configuration area 42 consists of high-frequency temperature and vibration acquisition configuration, low-frequency data acquisition, acoustic emission ultra-high-frequency acquisition configuration, and mud pump configuration. The mud pump configuration focuses on the setting of key parameters of the mud pump. Users can precisely adjust the core parameters such as the working area of the mud pump, pump power, outlet pressure, and inlet flow rate here. The high-frequency temperature and vibration acquisition configuration can monitor the high-frequency temperature and vibration data in the system in real time. The low-frequency data acquisition can view the real-time data of temperature sensors and pressure sensors, including key indicators such as temperature and pressure. The acoustic emission ultra-high-frequency acquisition configuration can monitor the acoustic emission ultra-high-frequency data in the system in real time. It meets the acquisition and identification of data information of various different sensors in the mud pump.

[0132] The risk warning area 43 consists of the risk identification result of the mud pump, predictive maintenance suggestions, and intelligent algorithm correction. The main function of the risk identification result of the mud pump is to display the risk scores of key components of the mud pump, including the scores of the power end, hydraulic end, lubrication system, and pressure maintaining system, as well as the overall score, facilitating users to understand the overall risk status of the mud pump and providing data support for subsequent risk management. The predictive maintenance suggestions are based on the risk identification result of the mud pump and comprehensively give warnings through two methods: a pre-set threshold and algorithm analysis of trends. For risks that the mud pump may face, such as severe wear of the suction valve body and seat, severe wear of the discharge valve body and seat, severe wear of the liner or piston, and severe eccentric wear of the piston against the liner, etc., it provides targeted predictive maintenance suggestions for users.

[0133] The downhole configuration area 44 consists of downhole risk correlation analysis and mud pump matching suggestions, and is used to display key parameters such as well depth, well temperature and pressure, wellbore structure, as well as the power and number of mud pumps.

[0134] At the same time, in order to more clearly explain the risk assessment method of the three-frequency fusion of the mud pump disclosed in the present invention, a healthy operating mud pump A that has been installed for one month, a healthy operating mud pump B that has been operating for half a year and has not had any failures, and a mud pump C with a poor health status whose vibration is so severe that it has been replaced with a temporary pipeline are taken as examples for supplementary explanation.

[0135] Vibration signal monitoring is carried out at four measuring points: the pump valve inlet, crosshead position, crankshaft position, and drive motor output end of mud pump A, mud pump B, and mud pump C. The sampling frequency is 1652 Hz, the single sampling time is 10 seconds, and sampling is carried out once every 5 minutes for 10 times. Then 40 samples are collected for each mud pump, and a total of 120 samples are collected for the three mud pumps. Each sample contains approximately 17,000 data points.

[0136] According to the risk assessment method proposed in the present invention, referring to Figures 8 - 11, from the data samples of the four measuring points of the mud pump A, a sample is randomly selected from each measuring point. 2000 data points are intercepted from each of the 4 selected samples as the calculation samples. The particle swarm optimization algorithm PSO is used to automatically optimize the key parameters of the variational mode decomposition VMD for each of them, and the convergence curves of the corresponding fitness values are plotted to obtain the optimal key parameters of the variational mode decomposition VMD for the data of the four measuring points, as shown in Table 1;

[0137] Table 1 Optimization results of the key parameters of the variational mode decomposition VMD for the four measuring points

[0138]

[0139] From all the data samples of the 4 measuring points of the mud pump A, 80000 data points are intercepted respectively, and they are evenly divided into 40 groups with 2000 data points in each group, obtaining a total of 160 groups of data as the data basis for the mud pump health evaluation model. The variational mode decomposition VMD is performed on each group of data according to the key parameters determined by their respective positions, that is, the sample data at each pump valve is decomposed into 5 first intrinsic mode functions IMF1, the sample data at each crosshead is decomposed into 4 first intrinsic mode functions IMF1, the sample data at each crankshaft part is decomposed into 5 first intrinsic mode functions IMF1, and the sample data at each motor output end is decomposed into 3 first intrinsic mode functions IMF1. The sensitive intrinsic mode functions are screened out to construct 4 groups of noise-reduced original signals and perform the second variational mode decomposition VMD.

[0140] After noise reduction, the permutation entropy PE of the second intrinsic mode function IMF2 after all the second variational mode decompositions VMD is calculated. The permutation entropy PE of each group of the second intrinsic mode function IMF2 is also combined into a 17-dimensional evaluation feature vector set that can reflect the operating state of the healthy mud pump A in the order of pump valve, crosshead, crankshaft, and motor, as shown in Table 2, where the permutation entropy value of each second intrinsic mode function IMF2 is represented by P n denotes, and n denotes the permutation entropy serial number.

[0141] Table 2 Evaluation feature vector set

[0142]

[0143]

[0144] Using the above evaluation feature vector set as the training samples, input them into the PSO-SVDD algorithm program built with Matlab to construct the SVDD hypersphere model.

[0145] Correspondingly, the key parameters of the hypersphere obtained after training the key parameters of the SVDD hypersphere model by the PSO method are shown in Table 3.

[0146] Table 3 Key Parameters of the SVDD Model

[0147]

[0148] Reference Figure 12 , Figure 12 shows an example diagram of the distance from the sample points corresponding to the evaluation feature vector set to the center of the hypersphere. The maximum distance among the 40 data samples is 0.0305, and there are 7 data points at a distance of 0.0305.

[0149] Based on the vibration signals at four measurement points: the inlet of the pump valve of Mud Pump A, the crosshead part, the crankshaft part, and the output end of the drive motor, the construction of the SVDD hypersphere model is completed.

[0150] Taking Mud Pump B and Mud Pump C as the mud pumps to be measured for illustration.

[0151] 80,000 data points are intercepted from the data samples of each measurement point of Mud Pump B and Mud Pump C respectively. According to the construction method of the evaluation feature vector set of Mud Pump A, a risk feature vector set is constructed and input into the established SVDD hypersphere model to obtain the distance D from each sample point to the center of the hypersphere.

[0152] As Figure 13 shown, taking 0.0305 as the radius, a total of 140 sample points are described, and the distance from the sample points to the center of the hypersphere is calculated;

[0153] As Figure 14 shown, by referring to the 3σ principle to convert the distance into health degree, the risk assessment of the measurement point positions of the mud pumps to be measured is completed. In this embodiment, it is set that the value range of the health degree H is 1 - 100, indicating the degree of the health state of the equipment; correspondingly, the closer H is to 100, the healthier the equipment; on the contrary, the closer H is to 0, the more abnormal the equipment. Then, 90% of the sample points in Mud Pump A are located at the position of the health degree score of 100, 72.5% of the sample points in Mud Pump B are located at the position of the health degree score of 100, all the sample points of Mud Pump C are below the position of the health degree score of 90, and 12.5% of the sample points of Mud Pump C are below the position of the health degree score of 70.

[0154] In summary, Mud Pump A is in a healthy state, Mud Pump B is in a good state, and Mud Pump C is in an unhealthy state. At the same time, according to the measurement point positions of Mud Pump C corresponding to the sample points, the faulty parts are further confirmed, the risk level and potential risk points are determined, and predictive maintenance suggestions are given. For example: there is a risk at the crosshead on the alarm interface, and the predictive maintenance suggestion is that the roughness of the crosshead guide plate is relatively high and needs to be replaced in time.

[0155] On the basis of the above disclosure, referring to Figure 15, the present disclosure also provides a risk assessment system for triple - frequency fusion of a mud pump, including:

[0156] A data collection unit for monitoring vibration signals at various measuring point positions under the healthy operating state of the mud pump;

[0157] A data sorting unit for performing variational mode decomposition on the vibration signals to obtain modal components, and extracting the permutation entropy of the modal components to form an evaluation feature vector set;

[0158] A model construction unit for constructing an SVDD hypersphere model based on the evaluation feature vector set;

[0159] A risk assessment unit for monitoring the operation vibration signals at various measuring point positions during the operation of the mud pump to be measured, analyzing the operation vibration signals using the SVDD hypersphere model, and completing the risk assessment of various measuring point positions of the mud pump to be measured.

[0160] It further includes:

[0161] A parameter optimization unit for presetting the key parameters of the SVDD hypersphere model and using the particle swarm optimization algorithm to complete the optimization of the key parameters of the SVDD hypersphere model;

[0162] The parameter optimization unit is also used for presetting the key parameters of the SVDD hypersphere model and using the particle swarm optimization algorithm to complete the optimization of the key parameters of the SVDD hypersphere model.

[0163] Among them, the data sorting unit is specifically used for:

[0164] Using variational mode decomposition (VMD) to decompose the vibration signal into the first intrinsic mode function and screening out the sensitive intrinsic mode function components;

[0165] Constructing a noise - reduced original signal based on the sensitive intrinsic mode function components;

[0166] Using variational mode decomposition (VMD) to denoise the original signal into the second intrinsic mode function;

[0167] Calculating the permutation entropy of the second intrinsic mode function components and extracting the feature vectors of the vibration signals at each measuring point;

[0168] Combining the feature vectors in the order of the measuring points to form an evaluation feature vector set.

[0169] Based on the above - disclosed content, correspondingly, the present disclosure also provides an electronic device. As Figure 6As shown, the electronic device according to an embodiment of the present disclosure includes at least one processor and at least one memory that are electrically connected. The memory is electrically connected to the processor. Among them, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method as described above.

[0170] In the description of the present invention, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of the said features. In the description of the present invention, the meaning of "a plurality" is two or more unless otherwise specifically defined.

[0171] In the description of the present invention, it should be noted that unless otherwise clearly specified and defined, the terms "mounted", "connected" and "coupled" should be understood in a broad sense. For example, it may be a fixed connection, a detachable connection or an integral connection; it may be a mechanical connection, an electrical connection or a communication connection; it may be directly connected or indirectly connected through an intermediate medium, and it may be the internal connection of multiple components or the interaction relationship of multiple components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0172] In the description of the present invention, it should be understood that all terms indicating orientation or positional relationship are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the indicated device or element must have a specific orientation, be constructed and operated in a specific orientation, and cannot be understood as a limitation of the present invention.

[0173] Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A risk assessment method for three-frequency fusion of a mud pump, characterized in that: The following steps are involved: Monitor the vibration signals of each measuring point under the healthy operation state of the mud pump; Perform variational modal decomposition on the vibration signal and obtain modal components, extract the permutation entropy of the modal components to form an evaluation feature vector set; Construct the SVDD hypersphere model based on the evaluation feature vector set; Monitor the operating vibration signals of each measuring point under the working state of the mud pump to be tested, use the SVDD hypersphere model to analyze the operating vibration signals, and conduct risk assessment on each measuring point of the mud pump to be tested.

2. The risk assessment method for three-frequency fusion of mud pumps according to claim 1 is characterized in that: Before performing variational modal decomposition on the vibration signal and obtaining modal components, the method further includes: Preset the key parameters of the variational mode, and use the particle swarm optimization algorithm to optimize the key parameters of the variational mode; Among them, the key parameters of variational modes include: the number of mode decomposition layers and the penalty factor.

3. The risk assessment method for three-frequency fusion of mud pumps according to claim 2 is characterized in that: The method of performing variational modal decomposition on the vibration signal and obtaining modal components, and extracting the permutation entropy of the modal components to form an evaluation feature vector set, comprises the following steps: The vibration signal is decomposed into the first eigenmode component by using variational mode, and the sensitive eigenmode component is screened out; constructing a denoised original signal based on sensitive eigenmode components; The original signal is de-noised using variational mode decomposition to the second eigenmode component; Calculate the permutation entropy of the second eigenmode component and extract the eigenvector of the vibration signal at each measuring point; The feature vectors are combined in the order of the measuring points to form an evaluation feature vector set.

4. The risk assessment method for three-frequency fusion of mud pumps according to claim 3 is characterized in that: The method of screening out sensitive eigenmode components comprises the following steps: Calculate the Pearson correlation coefficient between each first eigenmode component and the original signal; The first eigenmode component whose Pearson correlation coefficient is greater than a threshold is selected as a sensitive eigenmode component.

5. The risk assessment method for three-frequency fusion of mud pumps according to claim 1 is characterized in that: Before constructing the SVDD hypersphere model based on the evaluation feature vector set, the method further includes: The key parameters of the SVDD hypersphere model are preset, and the particle swarm optimization algorithm is used to optimize the key parameters of the SVDD hypersphere model; Among them, the key parameters of the SVDD hypersphere model include: penalty factor and kernel function parameters.

6. The risk assessment method for three-frequency fusion of mud pumps according to claim 1 is characterized in that: The SVDD hypersphere model is used to analyze the operation vibration signal and complete the risk assessment of each measuring point position of the mud pump to be tested, including the following steps: The operating vibration signals of each measuring point under the working state of the tested mud pump are subjected to variational modal decomposition and modal components are obtained to form a risk feature vector set; The risk feature vector set is input into the SVDD hypersphere model to form sample points, and the distance from the sample points to the center of the hypersphere is calculated; The distance is converted into healthiness according to Laida's law to conduct risk assessment on the location of the mud pump measurement point to be tested.

7. A risk assessment system for mud pump triple frequency fusion, characterized in that: include: A data collection unit is used to monitor the vibration signals of each measuring point under the healthy operation state of the mud pump; A data sorting unit is used to perform variational modal decomposition on the vibration signal and obtain modal components, and extract the permutation entropy of the modal components to form an evaluation feature vector set; A model building unit, used for building a SVDD hypersphere model based on the evaluation feature vector set; The risk assessment unit is used to monitor the operating vibration signals of each measuring point under the working state of the mud pump to be tested, and adopts the SVDD hypersphere model to analyze the operating vibration signals to complete the risk assessment of each measuring point of the mud pump to be tested.

8. The risk assessment system for mud pump triple frequency fusion according to claim 7 is characterized in that: Also includes: The parameter optimization unit is used to preset the key parameters of the SVDD hypersphere model and optimize the key parameters of the SVDD hypersphere model using a particle swarm optimization algorithm; The parameter optimization unit is also used to preset the key parameters of the SVDD hypersphere model and use the particle swarm optimization algorithm to optimize the key parameters of the SVDD hypersphere model.

9. The risk assessment system for mud pump triple frequency fusion according to claim 8, characterized in that: The data sorting unit is specifically used for: The vibration signal is decomposed into the first eigenmode function by using variational mode, and the sensitive eigenmode function components are screened out; Constructing the denoised original signal based on the sensitive intrinsic mode function components; The original signal is denoised using variational mode decomposition as the second eigenmode function; Calculate the permutation entropy of the second eigenmode function component and extract the eigenvector of the vibration signal at each measuring point; The feature vectors are combined in the order of the measuring points to form an evaluation feature vector set.

10. An electronic device, comprising at least one processor and at least one memory, wherein the memory is data-connected to the processor, wherein: The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 6.

Images