Intelligent vibration reduction control method and system for centrifugal pump

Abstract

The present application relates to the technical field of fluid machinery control, and discloses an intelligent vibration reduction control method and system for centrifugal pumps, the method comprising: obtaining frequency spectrum features from vibration signal data; when the frequency spectrum features exceed a threshold value, extracting mutation parameters, fusing pipeline system coupling data, and generating a multi-pump collaborative state vector; classifying according to the state vector, generating a distributed decision instruction sequence; fusing feedback information according to the instruction sequence, obtaining initial adaptive strategy parameters; adjusting operating parameters according to the initial parameters and calculating efficiency, and iteratively optimizing until the efficiency meets the standard, obtaining final adaptive strategy parameters; checking stability after deploying the final parameters, and if the check fails, returning to the cycle optimization. Through multi-pump collaboration and double closed-loop optimization, the present application can improve the stability, efficiency and reliability of the collaborative operation of the pump group.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fluid machinery control, in particular to an intelligent vibration reduction control method and system for centrifugal pumps. BACKGROUND

[0002] At present, as the core equipment for industrial fluid transportation, the running stability of centrifugal pumps is the key to ensuring the continuity and safety of production, and is widely used in energy, chemical industry, manufacturing and other fields. The vibration control technology of centrifugal pumps aims to suppress harmful vibrations generated during the operation of the pump group through real-time monitoring, data analysis and dynamic control, which is of great significance to prolong the service life of equipment, reduce energy consumption and prevent systematic failures. This technology usually relies on vibration sensors to collect state data, combines signal processing and control theory, and adjusts the operating parameters of the pump to maintain stable operation of the system.

[0003] In one prior art, the vibration reduction method mainly relies on the independent control of a single pump, which suppresses vibration by adjusting its speed or adding independent vibration reduction devices. The traditional scheme usually assumes that the operating environment of each pump is relatively independent, ignoring the mutual influence of fluid dynamic coupling and pressure fluctuations between multiple pump systems through pipelines. This limitation leads to a significant decline in vibration control effect when facing complex scenarios of multiple centrifugal pumps operating in parallel or series, making it difficult to form a system-level optimization strategy. More importantly, due to the lack of real-time sharing and collaborative decision-making mechanism of vibration information between pump groups, when a pump produces vibration due to load mutation, the vibration may be transmitted through the pipeline system, even causing resonance of other pumps, thereby affecting the running stability of the entire system.

[0004] In summary, the prior art has the problem of insufficient vibration control capability when multiple pumps operate cooperatively. SUMMARY

[0005] The present application provides an intelligent vibration reduction control method and system for centrifugal pumps to solve the problem of insufficient vibration control capability when multiple pumps operate cooperatively.

[0006] In a first aspect, to solve the above technical problems, the present application provides an intelligent vibration reduction control method for centrifugal pumps, comprising:

[0007] Obtain vibration signal data and perform frequency domain analysis to obtain frequency spectrum features;

[0008] Compare the frequency spectrum features with a preset vibration threshold value, if greater than the preset vibration threshold value, extract mutation parameters, obtain pipeline system coupling data, and fuse the mutation parameters and the pipeline system coupling data to generate a multi-pump cooperative state vector;

[0009] According to the multi-pump cooperative state vector, perform classification processing to generate a distributed decision instruction sequence;

[0010] If the distributed decision instruction sequence indicates a coordination mechanism activation, adjustment signaling and feedback information fusion are performed to obtain initial adaptive strategy parameters;

[0011] According to the initial adaptive strategy parameters, operation parameter adjustment and calculation efficiency indicators are performed, and the initial adaptive strategy parameters are iteratively updated and the adjustment and calculation are repeatedly performed until the efficiency indicator is greater than a preset efficiency threshold, to obtain final adaptive strategy parameters;

[0012] According to the final adaptive strategy parameters, new vibration signal data is distributed and obtained, and updated frequency spectrum characteristics are obtained;

[0013] According to the updated frequency spectrum characteristics, stability verification is performed, and if the verification result is that there is an abnormality or system instability caused by speed adjustment, the step of iteratively updating the initial adaptive strategy parameters is returned to perform until the efficiency indicator is greater than the preset efficiency threshold and the verification result passes.

[0014] Preferably, the vibration signal data is obtained, and frequency domain analysis is performed to obtain frequency spectrum characteristics, including:

[0015] Vibration signal data is collected from multiple centrifugal pumps to obtain original vibration data;

[0016] According to the original vibration data, frequency domain transformation processing is performed to obtain frequency spectrum characteristics.

[0017] Preferably, the frequency spectrum characteristics are compared with a preset vibration threshold, and if greater than the preset vibration threshold, a mutation parameter is extracted, pipeline system coupling data is obtained, and the mutation parameter and the pipeline system coupling data are fused to generate a multi-pump cooperative state vector, including:

[0018] The frequency spectrum characteristics are compared with a preset vibration threshold, and if the frequency spectrum characteristics are greater than the preset vibration threshold, the pump group is marked as high-risk, and a mutation parameter is extracted;

[0019] Pipeline system coupling data is obtained from a pipeline system;

[0020] The mutation parameter and the pipeline system coupling data are fused to generate a multi-pump cooperative state vector.

[0021] Preferably, according to the multi-pump cooperative state vector, classification processing is performed to generate a distributed decision instruction sequence, including:

[0022] Feature values are extracted from the multi-pump cooperative state vector to determine whether they exceed a preset cooperative threshold;

[0023] If the preset synergy threshold is exceeded, the multi-pump synergy state vector is input into a pre-trained model for classification to obtain a synergy state classification result.

[0024] According to the synergy state classification result, an instruction sequence is generated to obtain a distributed decision instruction sequence.

[0025] Preferably, if the distributed decision instruction sequence indicates that the coordination mechanism is activated, an adjustment signal is sent and feedback information is fused to obtain initial adaptive strategy parameters, including:

[0026] According to the distributed decision instruction sequence, a speed adjustment signal is sent to the high-risk pump group to obtain and get adjusted vibration signal data;

[0027] According to the adjusted vibration signal data, a frequency spectrum feature update and a difference operation are performed to obtain a dynamic change index;

[0028] If the dynamic change index exceeds a preset change threshold, feedback data is obtained and obtained;

[0029] According to the dynamic change index and the feedback data, fusion and weighted average are performed to obtain initial adaptive strategy parameters.

[0030] Preferably, according to the initial adaptive strategy parameters, the running parameter adjustment and the calculation efficiency index are performed, and the initial adaptive strategy parameters are updated and adjusted and calculated repeatedly until the efficiency index is greater than a preset efficiency threshold, to obtain final adaptive strategy parameters, including:

[0031] According to the initial adaptive strategy parameters, running parameter adjustment is performed to obtain and obtain optimized pump group vibration data;

[0032] According to the optimized pump group vibration data, a current efficiency index is calculated and obtained;

[0033] If the current efficiency index is less than a preset efficiency threshold, the feedback data is fused and iteratively adjusted and corrected until the current efficiency index is greater than the preset efficiency threshold, to obtain final adaptive strategy parameters.

[0034] Preferably, according to the final adaptive strategy parameters, new vibration signal data is distributed and obtained to obtain updated frequency spectrum features, including:

[0035] The final adaptive strategy parameters are distributed, and vibration signal data is reacquired to obtain new vibration signal data;

[0036] data integrity of the new vibration signal data is calculated, and if the data integrity is greater than a preset integrity threshold, the vibration signal data is transformed to obtain updated spectral features.

[0037] Preferably, according to the updated spectral features, stability verification is performed, and if the verification result is that there is an abnormality or system instability caused by speed adjustment, the step of cyclically iterating to update the initial adaptive strategy parameters is returned to be executed until the efficiency index is greater than the preset efficiency threshold and the verification result passes, including:

[0038] The updated spectral features are input into a pre-trained model for classification processing, and if the classification processing indicates that the stability verification fails, verification failure data is generated;

[0039] According to the verification failure data, the dynamic change index is integrated and weightedly averaged to obtain a fusion feature;

[0040] The initial adaptive strategy parameters are updated according to the fusion feature, and the step of cyclically iterating to update the initial adaptive strategy parameters is returned to be executed until the efficiency index reaches a preset efficiency threshold.

[0041] In a second aspect, the present application provides an intelligent vibration reduction control system for a centrifugal pump, comprising:

[0042] A data acquisition module is configured to acquire vibration signal data and perform frequency domain analysis to obtain spectral features.

[0043] An evaluation and diagnosis module is configured to compare the spectral features with a preset vibration threshold, extract mutation parameters if the spectral features are greater than the preset vibration threshold, acquire pipeline system coupling data, and fuse the mutation parameters and the pipeline system coupling data to generate a multi-pump coordination state vector.

[0044] A decision instruction generation module is configured to perform classification processing according to the multi-pump coordination state vector to obtain a distributed decision instruction sequence.

[0045] An initial strategy formulation module is configured to perform adjustment signal transmission and feedback information fusion to obtain initial adaptive strategy parameters if the distributed decision instruction sequence indicates that a coordination mechanism is activated.

[0046] An iteration and optimization module is configured to perform running parameter adjustment and calculate an efficiency index according to the initial adaptive strategy parameters, cyclically iterate to update the initial adaptive strategy parameters, and repeatedly perform adjustment and calculation until the efficiency index is greater than a preset efficiency threshold to obtain final adaptive strategy parameters.

[0047] A deployment and verification module is configured to distribute and acquire new vibration signal data according to the final adaptive strategy parameter, and obtain updated spectral features;

[0048] A checking and correcting module is configured to perform stability checking according to the updated spectral features, and if the checking result is that there is an abnormality or system instability caused by speed adjustment, return to perform the step of cyclically updating the initial adaptive strategy parameter until the efficiency index is greater than the preset efficiency threshold and the checking result is passed.

[0049] In a third aspect, the present application further provides an electronic device comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor implements the intelligent vibration reduction control method of the centrifugal pump according to any one of the above embodiments when executing the computer program.

[0050] In a fourth aspect, the present application further provides a computer readable storage medium comprising a stored computer program, wherein the computer readable storage medium controls the device where the computer readable storage medium is located to perform the intelligent vibration reduction control method of the centrifugal pump according to any one of the above embodiments when the computer program is running.

[0051] Compared with the prior art, the present application has the following beneficial effects:

[0052] (1) The present application forms a synergy vector capable of reflecting the overall operation state of the multi-pump system by fusing the mutation parameters of high-risk pump groups and the coupling data of pressure, flow and the like of the entire pipeline system, and makes a synergy decision accordingly, solving the problem that the prior art can only control a single pump independently and ignores the fluid dynamic coupling effect between pumps, thereby significantly improving the stability and control accuracy of the pump group operation under complex synergy working conditions.

[0053] (2) The present application calculates the efficiency index in real time after the initial speed adjustment, compares it with the preset threshold, and if it does not meet the standard, updates the initial adaptive strategy parameter by fusing feedback data, adjusting weights and correcting errors, and repeats the optimization process until the standard is met, solving the problem that the vibration reduction strategy in the prior art is relatively fixed and difficult to adapt to dynamic changes in load, thereby enhancing the adaptive ability of the system and achieving higher operating efficiency while ensuring the vibration reduction effect.

[0054] (3) The present application sets the final stability check and the correction mechanism of returning to the core optimization cycle after failure, after the optimal strategy is deployed, the system will perform the final stability check, if the check fails, not only will be corrected, but also will trigger the whole process to return to the core loop optimization step, until the system meets the dual standards of efficiency and stability, provides a final safety "bottom line" for the system, ensures that any adjustment will not sacrifice system stability, thereby greatly improving the reliability and operation safety of the whole pump group system, effectively avoiding the potential risks caused by improper adjustment. BRIEF DESCRIPTION OF DRAWINGS

[0055] Figure 1 is a kind of intelligent vibration control method flowchart of centrifugal pump provided by the first embodiment of the present application;

[0056] Figure 2 is a kind of intelligent vibration control system structure schematic diagram of centrifugal pump provided by the second embodiment of the present application. DETAILED DESCRIPTION

[0057] The technical solutions in the embodiments of the present application will be described clearly and completely in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0058] Referring to Figure 1 The first embodiment of the present application provides an intelligent vibration control method for centrifugal pump, comprising the following steps:

[0059] S11, obtain vibration signal data and perform frequency domain analysis to obtain frequency spectrum characteristics;

[0060] S12, compare the frequency spectrum characteristics with the preset vibration threshold value, if greater than the preset vibration threshold value, extract mutation parameters, obtain pipeline system coupling data, and fuse the mutation parameters and the pipeline system coupling data to generate a multi-pump cooperative state vector;

[0061] S13, according to the multi-pump cooperative state vector, perform classification processing to generate a distributed decision instruction sequence;

[0062] S14, if the distributed decision instruction sequence indicates that the coordination mechanism is activated, perform adjustment signal sending and feedback information fusion to obtain initial adaptive strategy parameters;

[0063] S15, according to the initial adaptive strategy parameter, running parameter adjustment and calculation efficiency index, loop iteration update the initial adaptive strategy parameter and repeat adjustment and calculation, until the efficiency index is greater than the preset efficiency threshold, get the final adaptive strategy parameter;

[0064] S16, according to the final adaptive strategy parameter, distribute and obtain new vibration signal data, get the updated frequency spectrum characteristics;

[0065] S17, according to the updated frequency spectrum characteristics, stability check, if the check result is abnormal or system instability caused by speed adjustment, return to execute the loop iteration update the initial adaptive strategy parameter step, until the efficiency index is greater than the preset efficiency threshold and the check result passes.

[0066] In step S11, the vibration signal data is obtained, and frequency domain analysis is carried out to obtain frequency spectrum characteristics, including:

[0067] The vibration signal data is collected from a plurality of centrifugal pumps to obtain the original vibration data;

[0068] According to the original vibration data, frequency domain transformation processing is carried out to obtain the frequency spectrum characteristics.

[0069] First, the vibration signal data is collected from a plurality of centrifugal pumps. In an embodiment, a high-precision acceleration sensor array is pre-installed on the bearing seat or pump shell of each centrifugal pump, and a signal acquisition protocol containing a time synchronization mechanism is used for periodic data acquisition. The process of using the signal acquisition protocol is as follows: a central time server is preset in the system, and all sensor nodes are synchronized to the server through the network time protocol (NTP), and the local clock is synchronized to the millisecond level. Each sensor performs periodic data acquisition at a sampling frequency of 1000 Hz, i.e. records the vibration acceleration instantaneous value every 1 millisecond, and encapsulates the instantaneous value and the synchronized high-precision time stamp into a data packet, which is transmitted through industrial Ethernet and collected, thereby obtaining and obtaining the original vibration data. The original vibration data is a multi-dimensional time series data set stored in a time series database, wherein each dimension corresponds to a collection sequence of a sensor, which is composed of a series of time stamps accurate to the millisecond level and corresponding vibration acceleration amplitudes.

[0070] Then, for each time series in the original vibration data, the specific operation of processing by using the fast Fourier transform is that, first, a Hanning window function is applied to the time series for pretreatment, and the specific process is that, a data segment of N sampling points is intercepted from the time series, and a Hanning window coefficient sequence with the same length N is generated, then each sampling point in the data segment is multiplied by the Hanning window coefficient at the corresponding position point by point, to obtain a windowed data segment with the amplitude of both ends approaching to zero, so as to reduce the frequency spectrum leakage generated in the subsequent transformation.

[0071] Then, the windowed data segment is taken as the input of the fast Fourier transform algorithm; the core butterfly operation of the algorithm is to decompose an N-point discrete Fourier transform (DFT) into two N / 2-point DFTs, and the specific operation is to divide the N-point input sequence into two groups according to the parity, calculate the N / 2-point DFT of each group respectively, and then combine the results of the two N / 2-point DFTs into an N-point DFT result through one complex multiplication and one complex addition and subtraction, and this process is recursively performed until it is decomposed into the most basic 2-point DFT operation, and finally the time domain data segment is efficiently decomposed into the superposition of sine and cosine harmonics on N / 2 discrete frequency components.

[0072] Finally, the modulus square of each frequency component corresponding complex result is calculated to obtain the energy value at the frequency point. The set of all frequency points and their corresponding energy values constitutes the spectrum feature. The spectrum feature is a feature vector quantifying the distribution of vibration energy at each frequency, and its specific form is a set of key-value pairs of frequency values and corresponding energy amplitudes, which clearly represents the energy concentration at key frequency components such as fundamental frequency and multiple frequency.

[0073] In step S12, the spectrum feature is compared with a preset vibration threshold, and if it is greater than the preset vibration threshold, a mutation parameter is extracted, pipeline system coupling data is obtained, and the mutation parameter and the pipeline system coupling data are fused to generate a multi-pump cooperative state vector, including:

[0074] The spectrum feature is compared with a preset vibration threshold, and if the spectrum feature is greater than the preset vibration threshold, the pump group is marked as high risk, and a mutation parameter is extracted;

[0075] Pipeline system coupling data is obtained from the pipeline system;

[0076] The mutation parameter and the pipeline system coupling data are fused to generate a multi-pump cooperative state vector.

[0077] Specifically, the spectral characteristics obtained in the previous step are first compared with a preset spectral energy vibration threshold. The spectral energy vibration threshold is set based on historical data statistical analysis. The specific setting process is as follows: historical vibration data of all pump groups under normal operating conditions are collected for at least one month. After spectral analysis, the 95th percentile of the energy amplitude is calculated for key frequency points such as the fundamental frequency and second harmonic. The threshold vector composed of these percentiles is used as the preset vibration threshold.

[0078] During comparison, if the energy amplitude of any key frequency point in the spectral characteristics is greater than its corresponding threshold component, the vibration is determined to be abnormal, and the centrifugal pump is marked as a high-risk pump group. Then, abrupt change parameters are extracted through time series analysis. If the energy amplitude of all key frequency points does not exceed the threshold, the pump group is determined to be operating normally, and the system returns to step S11 to continue monitoring. The specific operation of the time series analysis is as follows: First, model parameters are selected. The vibration amplitude sequence of M time points (M is set to 512 according to experimental statistics) before the trigger threshold is taken as a sample. Multiple autoregressive AR models of different orders are fitted to this sample sequence, and the Akaike Information Criterion (AIC) value of each model is calculated. The calculation formula is AIC = 2k - 2ln(L), where k is the number of model parameters (i.e., model order p), and L is the maximum likelihood function value of the model. Finally, the order p that minimizes the AIC value is selected as the optimal model order. Then, model training is performed. An AR(p) model is established based on the selected order p, and its mathematical expression is... Based on the autocorrelation function of the sample sequences, the autoregressive coefficients in the model are calculated. Finally, the model is used, and the vector of autoregressive coefficients obtained from training is output as the mutation parameter, which quantifies the intrinsic dynamic characteristics of the current vibration state.

[0079] At the same time, pressure sensors and flow meters installed in the pipeline system are used to acquire and obtain pipeline system coupling data in real time. The pipeline system coupling data is a data structure that includes the current pipeline pressure value, fluid flow value and corresponding timestamp.

[0080] Finally, the mutation parameters are fused with the pipeline system coupling data to generate a multi-pump collaborative state vector. The specific fusion process involves first performing Z-score normalization on mutation parameters, pressure data, and flow data with different sources and dimensions. Specifically, the mean μ and standard deviation σ of each data type over a past period are calculated. Then, for each currently acquired data value x, the normalized value is calculated by dividing the difference between the current value and the mean by the standard deviation, resulting in a dimensionless standardized dataset.

[0081] Then the principal component analysis method is used to process the standardized data set, and the operation is to calculate the covariance matrix of the data set, solve the eigenvalues and eigenvectors, select the first k eigenvectors with cumulative variance contribution rate exceeding 85% to form a projection matrix, and finally project the standardized data set on the matrix to obtain a k-dimensional multi-pump coordination state vector. Each dimension of the vector represents a system-level main operating mode, such as load balancing degree or overall pressure fluctuation trend.

[0082] In step S13, according to the multi-pump coordination state vector, a classification processing is performed to generate a distributed decision instruction sequence, including:

[0083] The eigenvalue is extracted from the multi-pump coordination state vector to determine whether it exceeds the preset coordination threshold;

[0084] If the preset coordination threshold is exceeded, the multi-pump coordination state vector is input into a pre-trained model for classification to obtain a coordination state classification result;

[0085] According to the coordination state classification result, an instruction sequence is generated to obtain a distributed decision instruction sequence.

[0086] Specifically, first, the vector norm of the multi-pump coordination state vector obtained in the previous step is calculated as the eigenvalue representing the coordination running deviation degree, and the specific calculation process is as follows: the component value of each dimension in the vector is squared, then the square results of all component values are summed, and finally the square root of the sum is taken to obtain the vector norm.

[0087] Then, the eigenvalue is compared with the preset multi-pump coordination deviation threshold. The setting basis of the multi-pump coordination deviation threshold is historical data statistical analysis, and the specific setting process is as follows: 100 groups of pump groups in normal coordination and known fault coordination states in historical operation are collected, the norm of each vector is calculated, and the norm value that best distinguishes the normal and fault states is selected as the threshold, for example, set to 0.75. If the calculated eigenvalue does not exceed the coordination threshold, it is determined that the pump group coordination running state is normal, and the system returns to step S11 for continuous monitoring; if the preset coordination threshold is exceeded, the multi-pump coordination state vector is input into a pre-trained support vector machine (SVM) classification model for classification processing. The specific operation of using the support vector machine model is as follows: the model has been trained by a large number of historical coordination state vector data labeled with "normal coordination", "load imbalance", "coordination fault" and the like, and has constructed an optimal hyperplane for dividing different state categories in a multi-dimensional feature space; when a new vector is input, the model will judge which side of the hyperplane the vector falls on, and output the corresponding coordination state classification result.

[0088] Finally, according to the cooperative state classification result, a distributed computing method is used to generate an instruction sequence to obtain a distributed decision instruction sequence. The specific generation process is that the classification results such as "load imbalance" or "cooperative failure" are distributed to edge computing nodes, and each node parses the state result into a control instruction containing specific parameters for a specific pump group according to a preset rule base. The setting basis of the preset rule base is the equipment operation mechanism, which contains the mapping relationship between each cooperative failure state and a set of optimal control actions. The mapping relationship is based on the pump group performance curve, the fluid mechanics model, and the historical operation and maintenance data of experienced equipment engineers. For example, for the result of "load imbalance", the instruction sequence "instruction 1: reduce the speed of pump 1 by 10%; instruction 2: increase the flow of pump 2 by 15%; instruction 3: balance the system pipeline pressure to 1.1 MPa" can be generated. The instruction sequence is then issued to each pump group controller through MQTT or other protocols to form the distributed decision instruction sequence.

[0089] In step S14, if the distributed decision instruction sequence indicates the activation of the coordination mechanism, adjustment signal sending and feedback information fusion are performed to obtain initial adaptive strategy parameters, including:

[0090] According to the distributed decision instruction sequence, a speed adjustment signal is sent to the high-risk pump group, and adjusted vibration signal data is obtained;

[0091] According to the adjusted vibration signal data, frequency spectrum feature updating and difference operation are performed to obtain a dynamic change index;

[0092] If the dynamic change index exceeds a preset change threshold, feedback data is obtained;

[0093] According to the dynamic change index and the feedback data, fusion and weighted average are performed to obtain initial adaptive strategy parameters.

[0094] Specifically, first, the coordination mechanism activation signal is extracted from the distributed decision instruction sequence obtained in the previous step, and according to the target speed value contained in the instruction, such as reducing the speed from 1500 rpm to 1350 rpm, a speed adjustment signal is issued to the controller of the high-risk pump group through a wireless module. After the speed adjustment is completed, the vibration signal data after adjustment is obtained by using the signal acquisition protocol and processing method in step S11, and the updated frequency spectrum feature is obtained by processing.

[0095] Subsequently, according to the adjusted vibration signal data, a spectrum feature update and a difference operation are performed to obtain a dynamic change index, and a specific calculation process is as follows: an energy amplitude difference operation is performed on each frequency point between the current updated spectrum feature and the spectrum feature before the speed adjustment to obtain a difference spectrum, then an absolute value sum of energy amplitude differences of all frequency points in the difference spectrum is calculated, and the sum value is the dynamic change index quantifying the real-time change intensity of the vibration mode.

[0096] Then, the dynamic change index is compared with a preset dynamic change threshold, and the preset dynamic change threshold is set according to experimental statistics, that is, through multiple experiments, a normal convergence range of the index under effective vibration reduction adjustment is counted, and an upper limit value thereof is taken as the threshold, for example, 0.3. If the dynamic change index does not exceed the threshold, it indicates that the current adjustment is effective and the system tends to be stable, and no initial adaptive strategy parameter needs to be generated, and a subsequent step is directly entered. If the dynamic change index exceeds the preset change threshold, it indicates that the system response is intense or unstable, at this time, operating state data containing real-time speed, load current and the like are obtained from a sensor network of the pump group, and environmental noise sound wave data are obtained from a microphone array installed around the pipeline, and the two parts of data jointly constitute the feedback data.

[0097] Finally, according to the dynamic change index and the feedback data, a fusion processing is performed to generate the initial adaptive strategy parameter, and a specific process is as follows: first, a median filtering algorithm is used to smooth each time sequence in the feedback data, and a specific operation is as follows: a sliding window with a size of N (N is an odd number, for example, 5) is set, and the window moves along the time sequence point by point; at each window position, N data points in the window are read, and the N data points are arranged in ascending order; a value at a middle position (i.e., the (N+1) / 2th position) after sorting is selected, and the median value is taken as a processing result to replace a value at a center point position of the original sequence, thereby generating a smoothed time sequence with abnormal disturbances removed.

[0098] Then, key statistics such as a mean value of the load current are extracted from the smoothed operating state data, and an average sound pressure level is extracted from the environmental noise data; finally, the dynamic change index, the mean value of the load current, the average sound pressure level and the like are spliced into a multi-dimensional fusion feature vector, and the vector is input into a preset mapping function, the mapping function is established based on offline simulation and historical experimental data, and the mapping function can map the input fusion feature vector into a specific control parameter value, and the output control parameter value is the initial adaptive strategy parameter used for subsequent fine adjustment.

[0099] In step S15, according to the initial adaptive strategy parameters, the running parameter adjustment and the efficiency index calculation are performed, the initial adaptive strategy parameters are iteratively updated, and the adjustment and calculation are repeatedly performed until the efficiency index is greater than the preset efficiency threshold, and the final adaptive strategy parameters are obtained, including:

[0100] According to the initial adaptive strategy parameters, the running parameter adjustment is performed, the optimized pump group vibration data are obtained, and the optimized pump group vibration data are obtained.

[0101] According to the optimized pump group vibration data, the current efficiency index is calculated and obtained.

[0102] If the current efficiency index is less than the preset efficiency threshold, the feedback data are fused, and the weight adjustment and error correction are iteratively performed until the current efficiency index is greater than the preset efficiency threshold, and the final adaptive strategy parameters are obtained.

[0103] Specifically, first, according to the initial adaptive strategy parameters obtained in the last step, the running parameters of the high-risk pump group are fine-tuned, and the running parameters mainly include the driving motor speed of the pump group. The fine-tuning rule and specific operation process are as follows: the initial adaptive strategy parameters are taken as an adjustment proportion factor, multiplied by a preset reference speed adjustment step (for example, 50 rpm) to obtain the actual speed adjustment amount of this time, and the controller adjusts the pump group speed according to the adjustment amount, for example, the speed is further adjusted from 1350 rpm to 1320 rpm. After the adjustment is completed, the optimized pump group vibration data are immediately obtained through the acceleration sensor.

[0104] Subsequently, according to the optimized pump group vibration data and in combination with the real-time input power data obtained from the power sensor, the current efficiency index is calculated and obtained. The specific calculation process is as follows: first, the vibration loss power is calculated by performing energy spectrum integration on the optimized vibration data; second, the effective output power is obtained by subtracting the vibration loss power from the input total power; and finally, the final ratio, i.e., the current efficiency index, is obtained by dividing the effective output power by the input total power.

[0105] It should be noted that the vibration loss power calculated by energy spectrum integration is a relative quantity or an indicative value, which is used to represent the trend of vibration intensity change, and its absolute value is not directly equal to the actual loss of mechanical power (watt). In practical applications, the calculation of the efficiency index can use one of the following alternatives, 1. The efficiency index can be defined as the difference between the reference vibration energy and the current vibration energy divided by the reference vibration energy. Wherein, the reference vibration energy can be the vibration energy spectrum integration value measured by the system under the rated steady state working condition, this index directly quantifies the percentage of vibration reduction; 2. Combined with the model and characteristic curve of the pump, an empirical correction relationship between the vibration characteristic value (such as overall vibration speed) and the pump efficiency is established, and the efficiency index can be defined as the efficiency correction factor calculated according to the vibration under the current working condition; those skilled in the art should understand that the core of the efficiency index is to provide a relative criterion that can be used to quantitatively evaluate the vibration reduction effect and system performance in the iterative optimization process, and its specific mathematical form can not be limited to the above definition.

[0106] Then, the current efficiency index is compared with a preset efficiency threshold, and the setting of the preset efficiency threshold is based on the equipment performance standard and historical operation data, for example, according to the equipment performance curve and combined with the historical optimal operation condition, the efficiency threshold is set to 0.85. If the current efficiency index is greater than or equal to the preset efficiency threshold, the iterative cycle is terminated, and the current adaptive strategy parameter is output as the final adaptive strategy parameter.

[0107] If the current efficiency index is less than the preset efficiency threshold, a parameter updating process is started, and Kalman filtering is used to fuse the feedback data, and its specific process includes two stages of prediction and update. In the prediction stage, the model predicts the feedback data value at the current time according to the state at the last time; in the update stage, the model uses the actual measured feedback data value to correct the prediction result in the prediction stage, so as to obtain a fused feedback data which is closer to the true value and has filtered out random noise; then, weight adjustment is carried out according to the fused feedback data, and its specific process and rules are that a set of initial weights (such as 0.6 and 0.4) are set for each channel (such as load current, environmental noise) in the feedback data, if the decrease of the efficiency index and the data fluctuation of a certain channel have strong correlation in time, then the weight of this channel is adjusted by a preset proportion (such as 10%) higher, and the weights of other channels are adjusted lower at the same time, so as to ensure that the total weight is 1.

[0108] Next, error correction is performed. The specific correction process is as follows: First, the difference between the current efficiency index and the preset efficiency threshold is calculated to obtain the efficiency deviation. Then, based on the adjusted weight coefficient, the efficiency deviation is proportionally allocated to each fused feedback data channel to correct the data so that the corrected data can better explain the reduction in efficiency. Finally, the initial adaptive strategy parameters are updated by multiplying the corrected feedback data by a correction factor, and the updated parameters are used for the next round of operating parameter adjustment. This cycle is repeated until the efficiency index meets the threshold requirement.

[0109] In step S16, based on the final adaptive strategy parameters, new vibration signal data is distributed and acquired to obtain updated spectral features, including:

[0110] The final adaptive strategy parameters are distributed, and the vibration signal data is reacquired to obtain new vibration signal data.

[0111] The data integrity of the new vibration signal data is calculated. If the data integrity is greater than a preset integrity threshold, the vibration signal data is transformed to obtain updated spectral features.

[0112] Specifically, the final adaptive strategy parameters obtained in the previous step are first distributed through a distribution protocol that includes a synchronization mechanism and a shared topology. The specific process is as follows: the central coordination node broadcasts a synchronization signal to all pump group controllers. At the same time that each controller receives the signal, it updates the initial adaptive strategy parameters stored locally to ensure the consistency of the control strategy of the entire pump group system. After the parameter distribution is completed, the signal acquisition protocol described in step S11 is immediately used to reacquire vibration signal data from each centrifugal pump to obtain new vibration signal data.

[0113] Subsequently, the data integrity of the new vibration signal data is calculated. The specific calculation process is as follows: within a preset time window, the total number of data packets theoretically expected to be received is compared with the total number of data packets actually received; the ratio of these two values ​​is the data integrity. Then, this data integrity is compared with a preset integrity threshold. The integrity threshold is set based on empirical statistics of communication network quality; for example, in an industrial Ethernet environment, the threshold is set to 98% based on historical packet loss rates. If the calculated data integrity is greater than or equal to the integrity threshold, the data is considered valid, and the new vibration signal data is transformed using the Fast Fourier Transform method described in step S11 to obtain the updated spectral characteristics. If the data integrity is less than the threshold, the acquired data is considered invalid, and the system discards this batch of data and initiates a new data acquisition request.

[0114] In step S17, a stability check is performed based on the updated spectral characteristics. If the check result indicates an anomaly or system instability caused by speed adjustment, the process returns to the step of iteratively updating the initial adaptive strategy parameters until the efficiency index is greater than the preset efficiency threshold and the check result passes. This includes:

[0115] The updated spectral features are input into a pre-trained model for classification. If the classification indicates that the stability check has failed, then check failure data is generated.

[0116] Based on the verification failure data, the dynamic change indicators are integrated and weighted averaged to obtain the fusion feature;

[0117] The initial adaptive strategy parameters are updated based on the fusion features, and the process returns to the step of iteratively updating the initial adaptive strategy parameters until the efficiency index reaches a preset efficiency threshold.

[0118] Specifically, the updated spectral features obtained in the previous step are first used as input and fed into a pre-trained Support Vector Machine (SVM) classification model for final stability verification. The model training process involves using historical spectral feature data containing thousands of groups labeled as "stable operation" or "cooperative failure" as the training set, selecting a radial basis function as the kernel function, and determining the optimal penalty coefficient C and kernel function parameters through grid search and cross-validation. .

[0119] The specific implementation process is as follows: First, define a C consisting of multiple pairs. A two-dimensional parametric grid composed of values, for example, the candidate value set for C is [0.1, 1, 10]. The candidate value set is [0.01, 0.1, 1]. Secondly, k-fold cross-validation is used to evaluate the performance of each parameter combination in the grid. Specifically, the training set is randomly divided into 5 disjoint subsets. Four subsets are used in turn to train the SVM model, and the remaining subset is used for validation. This process is repeated 5 times until each subset has been used as a validation set at least once. Finally, the average classification accuracy of the parameter combination in these 5 validations is calculated. Finally, all C and ... values ​​in the grid are traversed. Combinations were selected to maximize the average accuracy of the 5-fold cross-validation (C, The optimal hyperparameters for the model are a combination of these parameters. The goal of training is to use these optimal hyperparameters to train an optimal hyperplane on the entire training set that can separate the two classes of data points with the maximum margin.

[0120] The specific operation of the stability check processing is to map the current to-be-checked updated frequency spectrum feature into the high-dimensional feature space of the model, judge which side of the optimal hyperplane it falls on, and output the classification processing result of "running stable" or "cooperative failure". If the classification processing result of the model is "running stable", it indicates that the current intelligent vibration reduction control is successful, and the whole process ends, and the system returns to the normal monitoring state, that is, the system only performs the periodic data acquisition and frequency spectrum feature generation described in step S11, and does not trigger the subsequent abnormal processing and parameter adjustment process.

[0121] If the classification processing indicates that the stability check fails, the check failure data containing the abnormal frequency spectrum feature and the failure classification label is generated. Then, according to the check failure data, the dynamic change index calculated and stored in step S14 is integrated, and the specific integration process is as follows: first, the total energy value or the energy of the main abnormal frequency peak value is extracted from the abnormal frequency spectrum feature in the check failure data to obtain a scalar form of the spectrum abnormality index; then, the spectrum abnormality index and the dynamic change index are weighted and averaged, and the setting basis of the weight coefficient is the contribution degree of each index to the system instability in historical data analysis, for example, 0.6 and 0.4 respectively, the spectrum abnormality index is multiplied by 0.6, the dynamic change index is multiplied by 0.4, and the sum of the two products is obtained. A scalar value that can fully reflect the current instability state is obtained, that is, the fusion feature.

[0122] Finally, the final adaptive strategy parameter is corrected according to the fusion feature, and the correction rule and specific operation are as follows: the fusion feature is taken as a penalty term multiplied by a preset decay coefficient (for example, 0.1) to obtain a correction amount, and then the correction amount is subtracted from the current final adaptive strategy parameter to generate an updated adaptive strategy parameter that is more inclined to conservative stable operation; then, the system takes the updated adaptive strategy parameter as a new initial value, returns and re-executes the loop iteration optimization process described in step S15 until the efficiency index and the stability check result finally obtained both meet the preset requirements.

[0123] It should be noted that the various thresholds (such as vibration threshold, cooperation threshold, change threshold, efficiency threshold, integrity threshold, etc.) described in the present application are examples based on an exemplary device and working condition. In actual application, the specific values of these thresholds need to be calibrated and adjusted according to the specific model of the target centrifugal pump, the configuration of the pipeline system, the historical operation data and the desired control performance. The core of the present application is the control logic and architecture, not the specific threshold values.

[0124] To sum up, the application realizes adaptive cooperative control of centrifugal pump set vibration by constructing a multi-pump cooperative state vector and adopting a double closed-loop optimization mechanism, and significantly improves the stability and efficiency of system operation.

[0125] With reference to Figure 2 The second embodiment of the application provides an intelligent vibration reduction control system of a centrifugal pump, comprising:

[0126] A data acquisition module is configured to acquire vibration signal data and perform frequency domain analysis to obtain frequency spectrum characteristics.

[0127] An evaluation and diagnosis module is configured to compare the frequency spectrum characteristics with a preset vibration threshold value, extract a mutation parameter if the frequency spectrum characteristics are greater than the preset vibration threshold value, acquire pipeline system coupling data, and fuse the mutation parameter and the pipeline system coupling data to generate a multi-pump cooperative state vector.

[0128] A decision instruction generation module is configured to perform classification processing according to the multi-pump cooperative state vector to obtain a distributed decision instruction sequence.

[0129] An initial strategy formulation module is configured to perform adjustment signal sending and feedback information fusion if the distributed decision instruction sequence indicates that a coordination mechanism is activated to obtain initial adaptive strategy parameters.

[0130] An iteration and optimization module is configured to perform running parameter adjustment and calculation of an efficiency index according to the initial adaptive strategy parameters, cyclically iteratively update the initial adaptive strategy parameters, and repeatedly perform adjustment and calculation until the efficiency index is greater than a preset efficiency threshold value to obtain final adaptive strategy parameters.

[0131] A deployment and verification module is configured to distribute the final adaptive strategy parameters and acquire new vibration signal data to obtain updated frequency spectrum characteristics.

[0132] A checking and correction module is configured to perform stability checking according to the updated frequency spectrum characteristics, and if the checking result is that there is an abnormality or system instability caused by speed adjustment, return to perform the step of cyclically iteratively updating the initial adaptive strategy parameters until the efficiency index is greater than the preset efficiency threshold value and the checking result passes.

[0133] It should be noted that the intelligent vibration reduction control system of a centrifugal pump provided by the embodiments of the application is used to perform all process steps of the intelligent vibration reduction control method of a centrifugal pump of the above embodiments, and the working principles and beneficial effects of the two are one-to-one corresponding, and thus will not be repeated.

[0134] This invention also provides an electronic device. The electronic device includes a processor, a memory, and a computer program stored in the memory and executable on the processor, such as an intelligent vibration damping control program for a centrifugal pump. When the processor executes the computer program, it implements the steps in the aforementioned embodiments of the intelligent vibration damping control methods for centrifugal pumps, for example... Figure 1 The step S11 shown. Alternatively, when the processor executes the computer program, it implements the functions of each module / unit in the above-described device embodiments, such as the data acquisition module.

[0135] For example, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the electronic device.

[0136] The electronic device may be a desktop computer, laptop, handheld computer, or smart tablet, etc. The electronic device may include, but is not limited to, a processor and memory. Those skilled in the art will understand that the above components are merely examples of electronic devices and do not constitute a limitation on the electronic device. It may include more or fewer components than described above, or combine certain components, or different components. For example, the electronic device may also include input / output devices, network access devices, buses, etc.

[0137] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the electronic device, connecting all parts of the electronic device via various interfaces and lines.

[0138] The memory can be used to store the computer program and / or modules, and the processor realizes various functions of the electronic device by running or executing the computer program and / or modules stored in the memory, and calling data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required by a function (such as a sound playing function, an image playing function, etc.), and the like; and the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), and the like. In addition, the memory can include a high-speed random access memory, and can also include a nonvolatile memory, for example, a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state memory devices.

[0139] The modules / units integrated in the electronic device, if realized in the form of software function units and sold or used as independent products, can be stored in a computer readable storage medium. Based on this understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. The computer program can realize the steps of the above-mentioned various method embodiments when executed by a processor. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate forms, etc. The computer readable medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the contents included in the computer readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, the computer readable medium does not include electrical carrier signals and telecommunication signals.

[0140] It should be noted that the apparatus embodiments described above are merely illustrative, and the units described as separate units can or can not be physically separate, and the units displayed as units can or can not be physical units, i.e. can be located in one place, or can be distributed to multiple network units. Part or all of the modules can be selected to achieve the purpose of the embodiment scheme according to actual needs. In addition, the connection relationship between the modules in the apparatus embodiment provided by the present application indicates that there is a communication connection between them, which can be implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement it without creative labor.

[0141] The above specific embodiments further illustrate the purpose, technical scheme and beneficial effects of the present application. It should be understood that the above description is only a specific embodiment of the present application and is not intended to limit the protection scope of the present application. It is particularly pointed out that any modification, equivalent replacement, improvement, etc. made by those skilled in the art within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method of intelligent vibration reduction control of a centrifugal pump, characterized by, The method comprises the following steps: acquiring vibration signal data and performing frequency domain analysis to obtain frequency spectrum characteristics; comparing the frequency spectrum characteristics with a preset vibration threshold value, if greater than the preset vibration threshold value, extracting mutation parameters, acquiring pipeline system coupling data, and fusing the mutation parameters and the pipeline system coupling data to generate a multi-pump collaborative state vector; performing classification processing according to the multi-pump collaborative state vector to generate a distributed decision instruction sequence; if the distributed decision instruction sequence indicates that the coordination mechanism is activated, then performing adjustment signal sending and feedback information fusion to obtain initial adaptive strategy parameters; according to the initial adaptive strategy parameters, performing operation parameter adjustment and calculating efficiency indicators, and cyclically updating the initial adaptive strategy parameters and repeatedly performing adjustment and calculation until the efficiency indicator is greater than a preset efficiency threshold value, to obtain final adaptive strategy parameters; according to the final adaptive strategy parameters, performing distribution and acquiring new vibration signal data to obtain updated frequency spectrum characteristics; according to the updated frequency spectrum characteristics, performing stability checking, if the checking result is that there is an abnormality or system instability caused by speed adjustment, then returning to perform the step of cyclically updating the initial adaptive strategy parameters until the efficiency indicator is greater than the preset efficiency threshold value and the checking result passes. wherein the extracting mutation parameters comprises: First, the model parameter selection is carried out, the vibration amplitude sequence of the M time points before the trigger threshold is intercepted as a sample sequence, a plurality of autoregressive AR models of different orders are fitted for the sample sequence, and the Akaike information criterion value of each model is calculated respectively, and finally the order which makes the Akaike information criterion value minimum is selected as the best model order; then the model training is carried out, the AR model is established according to the selected order, and the autoregressive coefficients in the model are calculated based on the autocorrelation function of the sample sequence ; finally, the vector composed of the autoregressive coefficients obtained by training is output as the mutation parameter, and the mutation parameter quantifies the internal dynamic characteristics of the current vibration state; wherein the pipeline system coupling data includes current time pipeline pressure value, fluid flow value and corresponding time stamp.

2. The intelligent vibration damping control method of a centrifugal pump according to claim 1, characterized by, The method comprises the following steps: acquiring vibration signal data and performing frequency domain analysis to obtain frequency spectrum characteristics; acquiring vibration signal data from multiple centrifugal pumps to obtain original vibration data; 3. The intelligent vibration damping control method of a centrifugal pump according to claim 1, characterized by, performing frequency domain transformation processing according to the original vibration data to obtain frequency spectrum characteristics. The method comprises the following steps: comparing the frequency spectrum characteristics with a preset vibration threshold value, if greater than the preset vibration threshold value, extracting mutation parameters, acquiring pipeline system coupling data, and fusing the mutation parameters and the pipeline system coupling data to generate a multi-pump collaborative state vector; comparing the frequency spectrum characteristics with a preset vibration threshold value, if the frequency spectrum characteristics are greater than the preset vibration threshold value, marking as a high-risk pump group, and extracting mutation parameters; 4. The intelligent vibration damping control method of a centrifugal pump according to claim 1, characterized by, acquiring pipeline system coupling data from a pipeline system; fusing the mutation parameters and the pipeline system coupling data to generate a multi-pump collaborative state vector. The method comprises the following steps: extracting feature values from the multi-pump collaborative state vector to determine whether they exceed a preset collaborative threshold value; 5. The intelligent vibration damping control method of a centrifugal pump according to claim 3, characterized by, if the feature values exceed the preset collaborative threshold value, inputting the multi-pump collaborative state vector into a pre-trained model to perform classification to obtain a collaborative state classification result; according to the collaborative state classification result, performing instruction sequence generation to obtain a distributed decision instruction sequence. The method comprises the following steps: if the distributed decision instruction sequence indicates that the coordination mechanism is activated, then performing adjustment signal sending and feedback information fusion to obtain initial adaptive strategy parameters. According to the distributed decision instruction sequence, a rotating speed adjustment signal is sent to the high-risk pump group, and adjusted vibration signal data is obtained; According to the adjusted vibration signal data, frequency spectrum feature updating and difference operation are performed to obtain a dynamic change index; If the dynamic change index exceeds a preset change threshold, feedback data is obtained; According to the dynamic change index and the feedback data, fusion and weighted average are performed to obtain an initial adaptive strategy parameter.

6. The intelligent vibration damping control method of a centrifugal pump according to claim 5, characterized by, According to the initial adaptive strategy parameter, running parameter adjustment and efficiency index calculation are performed, the initial adaptive strategy parameter is updated iteratively, and adjustment and calculation are repeated until the efficiency index is greater than a preset efficiency threshold, to obtain a final adaptive strategy parameter, including: According to the initial adaptive strategy parameter, running parameter adjustment is performed, and optimized pump group vibration data is obtained; According to the optimized pump group vibration data, a current efficiency index is calculated and obtained; If the current efficiency index is less than a preset efficiency threshold, the feedback data is fused, and weight adjustment and error correction are iteratively performed until the current efficiency index is greater than the preset efficiency threshold, to obtain the final adaptive strategy parameter.

7. The intelligent vibration damping control method of a centrifugal pump according to claim 1, characterized by, According to the final adaptive strategy parameter, new vibration signal data is obtained by distribution and acquisition, and updated frequency spectrum features are obtained, including: The final adaptive strategy parameter is distributed, and vibration signal data is re-acquired to obtain new vibration signal data; The data integrity of the new vibration signal data is calculated, and if the data integrity is greater than a preset integrity threshold, the vibration signal data is transformed to obtain updated frequency spectrum features.

8. The intelligent vibration damping control method of a centrifugal pump according to claim 5, characterized by, According to the updated frequency spectrum features, stability verification is performed, and if the verification result is that there is an abnormality or system instability caused by rotating speed adjustment, the step of iteratively updating the initial adaptive strategy parameter is returned to be executed until the efficiency index is greater than the preset efficiency threshold and the verification result passes, including: The updated frequency spectrum features are input into a pre-trained model for classification processing, and if the classification processing indicates that the stability verification fails, verification failure data is generated; According to the verification failure data, the dynamic change index is integrated and weighted averaged to obtain a fusion feature; According to the fusion feature, the initial adaptive strategy parameter is updated, and the step of iteratively updating the initial adaptive strategy parameter is returned to be executed until the efficiency index reaches the preset efficiency threshold.

9. An intelligent vibration damping control system for a centrifugal pump, characterized by An intelligent vibration reduction control method for a centrifugal pump as claimed in any one of claims 1 to 8, comprising: A data acquisition module for acquiring vibration signal data and performing frequency domain analysis to obtain frequency spectrum features; An evaluation and diagnosis module for comparing the frequency spectrum features with a preset vibration threshold, extracting a mutation parameter if the frequency spectrum features are greater than the preset vibration threshold, acquiring pipeline system coupling data, and fusing the mutation parameter and the pipeline system coupling data to generate a multi-pump cooperative state vector; A decision instruction generation module is configured to perform classification processing according to the multi-pump coordination state vector to obtain a distributed decision instruction sequence; An initial strategy formulation module is configured to, if the distributed decision instruction sequence indicates that the coordination mechanism is activated, perform adjustment signal sending and feedback information fusion to obtain initial adaptive strategy parameters; An iteration and optimization module is configured to perform running parameter adjustment and calculation of an efficiency index according to the initial adaptive strategy parameters, cyclically iteratively update the initial adaptive strategy parameters and repeatedly perform adjustment and calculation until the efficiency index is greater than a preset efficiency threshold, and obtain final adaptive strategy parameters; A deployment and verification module is configured to perform distribution according to the final adaptive strategy parameters and obtain new vibration signal data to obtain updated frequency spectrum features; A check and correction module is configured to perform stability checking according to the updated frequency spectrum features, and if the checking result is that there is an abnormality or system instability caused by speed adjustment, return to perform the step of cyclically iteratively updating the initial adaptive strategy parameters until the efficiency index is greater than the preset efficiency threshold and the checking result passes.

Images