CN117345660A - Method, device, equipment and storage medium for monitoring cavitation state of centrifugal pump - Google Patents
Method, device, equipment and storage medium for monitoring cavitation state of centrifugal pump Download PDFInfo
- Publication number
- CN117345660A CN117345660A CN202311574879.6A CN202311574879A CN117345660A CN 117345660 A CN117345660 A CN 117345660A CN 202311574879 A CN202311574879 A CN 202311574879A CN 117345660 A CN117345660 A CN 117345660A
- Authority
- CN
- China
- Prior art keywords
- centrifugal pump
- cavitation
- liquid medium
- ultrasonic signal
- signal information
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 48
- 238000012544 monitoring process Methods 0.000 title claims abstract description 45
- 239000007788 liquid Substances 0.000 claims abstract description 100
- 238000003062 neural network model Methods 0.000 claims abstract description 66
- 238000007781 pre-processing Methods 0.000 claims abstract description 15
- 238000012545 processing Methods 0.000 claims abstract description 8
- 230000006870 function Effects 0.000 claims description 22
- 239000011159 matrix material Substances 0.000 claims description 9
- 230000010349 pulsation Effects 0.000 claims description 9
- 238000012806 monitoring device Methods 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 7
- 238000004422 calculation algorithm Methods 0.000 claims description 6
- 238000006243 chemical reaction Methods 0.000 claims description 6
- 238000005457 optimization Methods 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000012423 maintenance Methods 0.000 abstract description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 39
- 238000012549 training Methods 0.000 description 12
- 230000002706 hydrostatic effect Effects 0.000 description 7
- 230000002159 abnormal effect Effects 0.000 description 6
- 238000001514 detection method Methods 0.000 description 6
- 210000002569 neuron Anatomy 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000001914 filtration Methods 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 239000000523 sample Substances 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F04—POSITIVE - DISPLACEMENT MACHINES FOR LIQUIDS; PUMPS FOR LIQUIDS OR ELASTIC FLUIDS
- F04D—NON-POSITIVE-DISPLACEMENT PUMPS
- F04D15/00—Control, e.g. regulation, of pumps, pumping installations or systems
- F04D15/0088—Testing machines
Landscapes
- Engineering & Computer Science (AREA)
- Mechanical Engineering (AREA)
- General Engineering & Computer Science (AREA)
- Control Of Non-Positive-Displacement Pumps (AREA)
Abstract
The invention relates to a method, a device, equipment and a storage medium for monitoring cavitation state of a centrifugal pump, belonging to the technical field of centrifugal pump fault monitoring, wherein the monitoring method comprises the following steps: collecting ultrasonic signal information from a liquid medium in a centrifugal pump outlet pipeline; preprocessing the ultrasonic signal information to obtain ultrasonic signal information data, wherein the ultrasonic signal information data comprises waveform signals, counting signals and time difference signals; inputting ultrasonic signal information data into a pre-trained neural network model, calculating real-time state data of a liquid medium, comparing the real-time state data with preset state data, performing logic operation, and judging the cavitation state of the centrifugal pump to obtain a processing result of the ultrasonic signal information. According to the invention, the neural network model is used for calculating the multi-state information in the liquid, so that the cavitation state of the centrifugal pump is detected in a non-contact manner, sensors are not required to be arranged in the centrifugal pump and the pipeline, the use is simple, the maintenance is convenient, and the recognition degree is high.
Description
Technical Field
The application relates to the technical field of centrifugal pump fault monitoring, in particular to a method, a device, equipment and a storage medium for monitoring cavitation state of a centrifugal pump.
Background
When the absolute pressure at the water inlet of the centrifugal pump is reduced to cavitation pressure at the water temperature, water is vaporized, and liquid water forms gas at the water inlet of the centrifugal pump, so that a plurality of small bubbles are formed. These small bubbles break rapidly as the water flows into the high pressure region, and the break generates a large hydraulic impact that repeatedly acts on the impeller of the centrifugal pump at a frequency of tens of thousands of times per second, and over time, the blades of the impeller gradually peel off due to fatigue damage, a phenomenon known as cavitation.
When cavitation is generated, the external characteristics of the water pump are not obviously affected, and after the cavitation is developed to a certain extent, the power, efficiency, flow, lift and other data of the water pump can be suddenly reduced. When cavitation progresses sufficiently, the effective flow area of the water flow is reduced so much that the water flow is interrupted and cannot work, and when the blade is degraded and corroded seriously, serious accidents such as blade fracture or perforation are generated.
Therefore, it is necessary to monitor the centrifugal pump for cavitation and to eliminate it. In the related art, mechanical vibration, noise, ultrasonic waves, liquid pressure, temperature and other data generated during the operation of the water pump are generally monitored, and then the data are analyzed to determine the cavitation condition of the centrifugal pump. However, in the related art described above, for example, the pump head value, the pump body vibration value, the pump driving power value, and other data are obtained through the pump inlet pressure sensor, the pump outlet pressure sensor, the pump body vibration sensor, the sound sensor, the pump driving power sensor, and the signal collector for collecting the foregoing information, and the obtained data are compared with the prestored data in real time to determine the cavitation state, so that there is a problem that a large number of sensors need to be installed in the centrifugal pump or the pipeline, and usually, one data needs to be matched with a special sensor, if the sensor is damaged, the centrifugal pump needs to be stopped during replacement, and even the centrifugal pump needs to be disassembled for maintenance.
Therefore, another way is to judge the severity of cavitation of the centrifugal pump by monitoring the size, density, etc. of bubbles in the liquid in the pipeline, for example, by using a capacitance method and a photoelectric method to monitor the state of bubbles in the pipeline of the centrifugal pump. However, if the cavitation condition is judged by monitoring bubbles through a photoelectric method and a capacitance method, for example, a capacitance polar plate is respectively arranged at two sides of a pipeline to measure capacitance change between the two polar plates, and the change condition of a liquid medium in the pipeline is calculated, the problems that the acquisition of cavitation key data is difficult to obtain, the data amount is small, the variety of the data is so small that manual intelligent automatic monitoring cannot be adopted, and the recognition degree of the cavitation state of the centrifugal pump is low are existed.
Disclosure of Invention
In order to solve the problems, the application provides a method, a device, equipment and a storage medium for monitoring the cavitation state of a centrifugal pump, which can detect bubbles in a liquid medium through ultrasonic waves to obtain real-time state data of a data liquid medium for judging the cavitation state of the centrifugal pump, and accurately judge whether the centrifugal pump belongs to dangerous cavitation.
In a first aspect, the present application provides a method for monitoring cavitation state of a centrifugal pump, which adopts the following technical scheme:
collecting ultrasonic signal information from a liquid medium in a centrifugal pump outlet pipeline;
preprocessing the ultrasonic signal information to obtain ultrasonic signal information data, wherein the ultrasonic signal information data comprises a waveform signal, a counting signal and a time difference signal;
inputting ultrasonic signal information data into a pre-trained neural network model, calculating real-time state data of a liquid medium, comparing the real-time state data with preset state data, performing logic operation, and judging the cavitation state of the centrifugal pump to obtain a processing result of the ultrasonic signal information;
and outputting the processing result of the ultrasonic signal information.
By adopting the technical scheme, the real-time state data of the liquid medium can be calculated according to the acquired ultrasonic signal information data, so that the arrangement of sensors in the centrifugal pump and the pipeline is avoided; the information quantity carried by the ultrasonic waves is large enough to provide enough training data for the neural network model; after the neural network model is trained to achieve satisfactory detection precision, the cavitation state of the centrifugal pump can be monitored in real time.
Optionally, the preprocessing step includes:
collecting voltage signals in ultrasonic signal information;
analog-to-digital conversion is carried out on the voltage signal;
and comparing the voltage signal after analog-digital conversion with a reference signal to obtain the waveform signal, the counting signal and the time difference signal.
By adopting the technical scheme, the acquired ultrasonic signal information is converted into the digital signal which can be identified by a computer, and then the neural network model can calculate the real-time state data of the liquid medium and judge the cavitation state of the centrifugal pump.
Optionally, the real-time status data includes a time point when the bubble appears in the liquid medium, a density of the bubble, a size of the bubble, a duration of the bubble, a pulsating pressure of the liquid medium, and a temperature of the liquid medium; and calculating the waveform signal, the counting signal and the time difference signal through a neural network model to obtain the occurrence time point of the bubbles, the density of the bubbles, the size of the bubbles, the duration of the bubbles, the pulsation pressure of the liquid medium and the temperature of the liquid medium.
By adopting the technical scheme, the time point when the bubble appears can be obtained by counting the continuity of the waveform signal; the average value and standard deviation of the waveform signal amplitude are counted to obtain the density of bubbles, and air bubbles and liquid vapor bubbles can be identified by comparing the standard deviation of the amplitude, so that more accurate time points of bubble occurrence are obtained; counting abnormal time-based pulses in the signal, and calculating the size of the bubbles; the abnormal sampling times in the signals are counted, and the duration time of the bubbles can be calculated; according to the time difference of the ultrasonic waves, the flow velocity of the liquid medium and the hydrostatic sound velocity can be calculated, the pulsation pressure of the liquid medium can be further calculated according to the flow velocity of the liquid medium, the pressure pulsation of the liquid medium can be obtained by counting the change of the pulsation pressure of the liquid medium within a period of time, cavitation is generated due to the fact that liquid steam bubbles break in a high-pressure area, therefore, the pressure pulsation of the liquid medium is very critical state data in cavitation monitoring, and the temperature of the liquid medium can be further calculated according to the hydrostatic sound velocity.
Optionally, the preset state data includes a time point when the bubbles appear when the dangerous cavitation appears and when the non-dangerous cavitation appears, the density of the bubbles, the size of the bubbles, the duration of the bubbles, the pulsation pressure of the liquid medium and the temperature of the liquid medium; the step of obtaining the preset state data comprises the following steps:
presetting a plurality of time periods and a plurality of liquid medium flow rates, and taking the working condition of the centrifugal pump which is in a preset liquid medium flow rate in a preset time period as a preset working condition;
collecting ultrasonic signal information of liquid medium in a centrifugal pump outlet pipeline under each preset working condition;
judging cavitation state of the centrifugal pump under each preset working condition and calculating state data of the liquid medium according to corresponding ultrasonic signal information;
and storing the cavitation state of the centrifugal pump and the state data of the corresponding liquid medium under all preset working conditions as preset state data.
By adopting the technical scheme, the preset state data of the centrifugal pump to be monitored is obtained, and can be used for training a neural network model and can be used as preset state data to be compared with real-time state data in actual monitoring; the plurality of working conditions are set, so that more accurate preset state data can be obtained, and the detection precision of the model is improved.
Optionally, the neural network model is a feedforward neural network model, and comprises an input layer, a hidden function layer and an output layer; the input layer is used for transmitting the waveform signals, the counting signals and the time difference signals into the hidden function layer; the hidden function layer is used for calculating real-time state data, comparing the real-time state data with preset state data and performing logic operation; the output layer is used for outputting an operation result outside the feedforward neural network model;
the step of obtaining the operation result comprises the following steps:
calculating real-time state data;
comparing the calculated real-time state data with preset state data under preset working conditions, and performing logic operation on the comparison result to obtain an operation result of dangerous cavitation/non-dangerous cavitation;
the step of calculating real-time status data includes:
calculating the occurrence time point of bubbles in the liquid medium and the density of the bubbles according to the waveform signals;
calculating the size of bubbles in the liquid medium and the duration of bubbles according to the counting signals;
the pulsating pressure of the liquid medium and the temperature of the liquid medium are calculated from the time difference signal.
By adopting the technical scheme, the feedforward neural network model can automatically judge the cavitation state of the centrifugal pump according to the real-time state data.
Optionally, before the input layer receives the waveform signal, the count signal and the time difference signal, the method further includes:
the feedforward neural network model randomly initializes the weight matrix of the feedforward neural network model and selects a nonlinear optimization algorithm as an optimization algorithm; after the output layer outputs the operation result, the method further comprises:
and calculating a loss function by the feedforward neural network model, debugging a weight matrix by using a back propagation method, and stopping iteration when the feedforward neural network model iterates to a preset number of times.
By adopting the technical scheme, the weight matrix of the feedforward neural network can be optimized, the method is more suitable for the centrifugal pump to be tested, and the detection precision of the feedforward neural network is improved.
In a second aspect, the present application provides a centrifugal pump cavitation state monitoring device, which adopts the following technical scheme:
the acquisition module is used for acquiring ultrasonic signal information of the liquid medium in the centrifugal pump outlet pipeline;
the data preprocessing module is used for preprocessing the ultrasonic signal information to obtain ultrasonic signal information data, wherein the ultrasonic signal information data comprises a waveform signal, a counting signal and a time difference signal;
the calculation module is used for inputting the ultrasonic signal information data into a neural network model, and the neural network model is used for calculating the real-time state data of the liquid medium;
and the comparison module is used for comparing the real-time state data with preset state data to obtain a monitoring result of the cavitation state of the centrifugal pump.
In a third aspect, the present application provides a computer device, which adopts the following technical scheme:
a computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the method of monitoring the cavitation condition of a centrifugal pump according to any one of the first aspects when the program is executed.
In a fourth aspect, the present application provides a computer readable storage medium, which adopts the following technical scheme:
a computer readable storage medium storing a computer program capable of being loaded by a processor and executing the method for monitoring the cavitation state of a centrifugal pump according to any one of the first aspects.
In summary, the present application includes at least one of the following beneficial technical effects:
the real-time state data of the liquid medium can be calculated according to the acquired ultrasonic signal information data, so that the arrangement of sensors in the centrifugal pump and the pipeline is avoided; the information quantity carried by the ultrasonic waves is large enough to provide enough training data for the neural network model; after the neural network model is trained to achieve satisfactory detection precision, the cavitation state of the centrifugal pump can be monitored in real time.
Drawings
Fig. 1 is a schematic flow chart of a method for monitoring cavitation status of a centrifugal pump according to one embodiment of the present application.
Fig. 2 is a second flow chart of a method for monitoring cavitation status of a centrifugal pump according to one embodiment of the present application.
Fig. 3 is a third flow chart of a method for monitoring cavitation of a centrifugal pump according to one embodiment of the present application.
FIG. 4 is a schematic flow diagram of a feed-forward neural network model in the present application.
FIG. 5 is a flow chart of the feedforward neural network training step in the present application.
FIG. 6 is a flow diagram of one embodiment of the present application.
Fig. 7 is a block diagram of a centrifugal pump cavitation status monitoring device according to one embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be further described in detail with reference to fig. 1 to 7 and the embodiments. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the present application.
The application provides a method for monitoring cavitation state of a centrifugal pump.
As shown in fig. 1, the method comprises the following steps:
step S101, acquiring preset state data, referring to fig. 2, specifically the steps are as follows:
s1011, presetting a plurality of time periods and a plurality of liquid medium flow rates, and taking a centrifugal pump working condition under a preset liquid medium flow rate in a preset time period as a preset working condition;
s1012, collecting ultrasonic signal information from liquid medium in a centrifugal pump outlet pipeline under each preset working condition;
s1013, judging cavitation state of the centrifugal pump under each preset working condition and calculating state data of the liquid medium according to the corresponding ultrasonic signal information;
s1014, storing cavitation states of the centrifugal pump and state data of corresponding liquid media under all preset working conditions as preset state data;
in this embodiment, in order to obtain more comprehensive data, nine working conditions including a high flow rate, a medium flow rate and a low flow rate of the centrifugal pump are preset in three time periods including the morning, the evening, and the evening, the liquid medium state information and the cavitation state of the centrifugal pump under each preset working condition are obtained, the states of the liquid medium are affected by different degrees in consideration of different external factors such as the atmospheric pressure and the temperature in different time periods, the time period from 6 to 11 is divided into the morning, the flow rate of the liquid is set to be three of the high flow rate, the medium flow rate and the low flow rate, the time period from 11 to 19 is divided into the evening, the flow rate of the liquid is set to be three of the high flow rate, the medium flow rate and the low flow rate, the time period from 19 to 6 is divided into the evening, and the flow rate of the liquid is set to be three of the high flow rate, the medium flow rate and the low flow rate, and of course, the time period division and the flow rate setting can be adjusted according to actual conditions, for example, the time division of each hour is divided into one time period, and the flow rate of the liquid is set to be two of the high flow rate and the low flow rate.
In this embodiment, ultrasonic waves are emitted to the liquid medium in the centrifugal pump outlet pipeline at an oblique beam angle under nine preset working conditions, ultrasonic signal information of the liquid medium in the centrifugal pump outlet pipeline is collected, and state information of the liquid medium under each preset working condition is obtained, and the cavitation state of the centrifugal pump comprises two types of dangerous cavitation and non-dangerous cavitation, however, the cavitation state of the centrifugal pump can also be divided according to the actual situation of the centrifugal pump, for example, the cavitation state of the centrifugal pump is divided into non-dangerous cavitation, first-stage dangerous cavitation, second-stage dangerous cavitation and third-stage dangerous cavitation, and the cavitation state of the centrifugal pump is judged manually in this embodiment, and the obtained state information of the liquid medium under the preset working condition and the cavitation state of the centrifugal pump are used as training data of a feedforward neural network model;
of course, the information of the operation time, the vibration frequency, the flow pressure, the lift, the voltage, the flow, the power and the like of the centrifugal pump under the corresponding working conditions can be obtained, and the cavitation state of the centrifugal pump can be further judged by combining the information with the calculation result obtained by the follow-up feedforward neural network model;
the transmission monitoring is carried out by adopting the oblique beam angle to emit ultrasonic waves, compared with the vertical emission ultrasonic waves, the transmission monitoring has stronger precision and anti-interference capability, the oblique beam angle can be 30 degrees, 45 degrees or 60 degrees and other angles, the ultrasonic wave can be emitted by using devices such as an ultrasonic transducer, an ultrasonic probe and the like, the detection effect can be achieved by installing an ultrasonic generating device on the outer wall of a pipeline, and the non-contact monitoring is realized;
in this embodiment, since cavitation is generated by the collapse of liquid vapor bubbles, and the centrifugal pump may contain a large amount of air bubbles in the inlet pipe due to its own operation principle, the excessive air bubbles may cause interference, so that the bubble state of the liquid medium in the outlet pipe is selectively detected;
in this embodiment, six types of information for judging the cavitation state can be obtained from the ultrasonic wave, including a time point when the bubble appears, a density of the bubble, a size of the bubble, a duration of the bubble, a pulsating pressure of the liquid medium and a temperature of the liquid medium, and the information obtained by the photoelectric method and the capacitive method is more monolithic, the data amount is small, the data types are few, and it is difficult to obtain key data such as the liquid pressure pulsation, and the like, so that the feedforward neural network model is insufficient to be trained to achieve the purpose of automatically judging the cavitation state.
Step S102, preprocessing ultrasonic signal information to obtain ultrasonic signal information data, wherein the ultrasonic signal information data comprises waveform signals, counting signals and time difference signals;
in this embodiment, as shown in fig. 3, the specific preprocessing step in step S102 includes:
step S1021, digital filtering is carried out on the ultrasonic signal information;
step S1022, collecting voltage signals in the ultrasonic signal information;
step S1023, performing analog-to-digital conversion on the voltage signal, and comparing the voltage signal after analog-to-digital conversion with a reference signal to obtain a waveform signal, a counting signal and a time difference signal;
in the embodiment, the acquired ultrasonic signal information is converted into a digital signal which can be identified by a computer, and then the feedforward neural network model can calculate the state data of the liquid medium and judge the cavitation state of the centrifugal pump; the digital filtering technology can eliminate abnormal values and weak signals, and increases the amount of available data, and the more the available data is, the better the training effect of the feedforward neural network model is.
Step S103, a feedforward neural network model is established and trained;
in the embodiment, before inputting data to a feedforward neural network model, randomly initializing a weight matrix of the feedforward neural network model and selecting a nonlinear optimization algorithm as an optimization algorithm;
in this embodiment, as shown in fig. 4, the feedforward neural network model includes an input layer, a first hidden function layer, a second hidden function layer and an output layer, where the input layer is used to transmit a waveform signal, a count signal and a time difference signal to the first hidden function layer, the first hidden function layer is used to calculate state data of a liquid medium, and the second hidden function layer is used to compare the calculated state data of the liquid medium with state data under a preset working condition, and perform a logic operation on the comparison result to obtain an operation result of dangerous cavitation/non-dangerous cavitation;
in this embodiment, as shown in fig. 5, the step of training the feedforward neural network model in step S103 specifically includes:
s1031, an input layer of the feedforward neural network model receives ultrasonic signal information data and completes classification transfer of waveform signals, counting signals and time difference signals;
s1032, calculating state data of the liquid medium according to the input waveform signals, the counting signals and the time difference signals by the first hidden function layer of the feedforward neural network model;
s1033, performing logic operation on the state data of the liquid medium and the state data under the preset working condition by a second hidden function layer of the feedforward neural network model to obtain a logic result of the cavitation state;
s1034, outputting a final result of the network judgment by an output layer of the feedforward neural network model, transmitting the final result to a man-machine interface layer in an acousto-optic mode, calculating a loss function by the feedforward neural network model, debugging a weight matrix by using a back propagation method, and stopping iteration when the feedforward neural network model iterates to a preset number of times;
s1035, iterating the feedforward neural network model to a preset number of times to finish training;
in this embodiment, the first hidden function layer of the feedforward neural network model calculates state data of the liquid medium according to the input waveform signal, the count signal and the time difference signal, wherein the state data is indirect state data, specifically:
counting the continuity of the waveform signals to obtain the time point when the bubble appears;
the average value and standard deviation of the amplitude of the waveform signal are counted to obtain the density of the bubbles, and in addition, because the standard deviation of the amplitude of the air bubbles is different from the amplitude of the liquid vapor bubbles, for example, the standard deviation of the amplitude of the air bubbles is larger than the standard deviation of the amplitude of the water vapor bubbles, the types of the bubbles can be identified by comparing the standard deviations of the amplitudes, and more accurate time points of occurrence of the liquid bubbles can be obtained;
counting abnormal time-base pulses in the signal, and calculating the size of bubbles;
timing the abnormal sampling times in the signal, and calculating the duration of the bubbles;
according to the time difference between ultrasonic wave emission and receiving, the flow velocity and the hydrostatic sound velocity of the liquid are calculated, the pulsating pressure of the liquid can be further calculated according to the flow velocity of the liquid, the temperature of the liquid can be further calculated according to the hydrostatic sound velocity, and the flow velocity V of the liquid in the application is availableIt indicates that the speed of sound C of the hydrostatic sound is available +.>Where θ is the angle of emission of the ultrasonic oblique beam, t d Is the time of downstream propagation of ultrasonic wave, t u The ultrasonic counter-current propagation time is the distance between the transmitting end and the receiving end of the ultrasonic generating device, and after the hydrostatic sound velocity C is obtained, for example, when the liquid medium is water, the water temperature can be obtained according to a hydrostatic sound velocity-water temperature relation table.
Step S104, collecting ultrasonic signal information of a liquid medium in a centrifugal pump outlet pipeline and preprocessing; in this embodiment, real-time status information of the liquid medium in the pump outlet conduit is collected.
And step 105, inputting ultrasonic signal information into a pre-training feedforward neural network model to calculate and obtain the cavitation state of the centrifugal pump.
In this embodiment, the pre-trained feedforward neural network model can calculate in real time whether the cavitation state of the centrifugal pump is dangerous cavitation or non-dangerous cavitation, and the working personnel can further judge the working state of the centrifugal pump according to the cavitation state result obtained by calculation and combining the information such as the working time, the vibration frequency, the flow pressure, the lift, the voltage, the flow rate, the power and the like of the centrifugal pump.
The present application is further illustrated below with reference to examples.
An embodiment of the present application provides a method for monitoring cavitation state of a centrifugal pump, as shown in fig. 6, the method is executed by a device, the device is connected with an ultrasonic transducer, the ultrasonic transducer is installed on an outer wall of a water outlet pipeline of the centrifugal pump, and emits an ultrasonic signal to water flow in the water outlet pipeline of the centrifugal pump at an angle of 45 ° in a diagonal beam, the method includes:
step S201, collecting ultrasonic signal information of water flow in a water outlet pipeline of a centrifugal water pump;
for this embodiment, the centrifugal pump is the centrifugal pump, which is the most common centrifugal pump, and because of the working principle of the centrifugal pump itself, there are many air bubbles in the water inlet pipe, and the object of monitoring is the vapor bubbles that cause cavitation, therefore, the ultrasonic signal information from the water flow in the water outlet pipe of the centrifugal pump is collected. If the centrifugal oil pump of the oil pipeline is monitored, the set working condition of the centrifugal oil pump is required to be tested, and the data is obtained and used as training data of the neural network model.
Step S202, preprocessing ultrasonic signal information to obtain ultrasonic signal information data, wherein the ultrasonic signal information data comprises waveform signals, counting signals and time difference signals;
for the embodiment, the digital filtering technology is also used for processing the acquired ultrasonic signal information, so that abnormal values and weaker signals are eliminated, and the monitoring precision is improved.
Step S203, inputting ultrasonic signal information data into a neural network model, wherein the neural network model is used for calculating real-time state data of water flow, comparing the real-time state data with preset state data, performing logic operation, and judging the cavitation state of the centrifugal pump;
for the embodiment, the real-time state data includes the time point when the air bubble appears in the liquid medium, the density of the air bubble, the size of the air bubble, the duration of the air bubble, the pulsating pressure of the liquid medium and the temperature of the liquid medium, and the preset working conditions are the state information of the water flow and the corresponding cavitation state under three working conditions of high flow rate, medium flow rate and low flow rate at night, so that the operation condition of the centrifugal water pump is more difficult to observe at night, and compared with the operation condition of the centrifugal water pump in the morning and the afternoon, the air pressure and the water temperature at night also have larger change, and the neural network model is more needed to assist the staff in monitoring.
Step S204, obtaining the processing result of the ultrasonic signal information.
For the embodiment, the neural network model is a feedforward neural network model, the feedforward neural network model is trained in advance, a weight matrix of the feedforward neural network model is initialized randomly before training, a counter propagation method is adopted to adjust the weight matrix of the feedforward neural network model during training, iteration is performed to preset times, final detection accuracy reaches an expected value, three neurons are arranged on an input layer of the feedforward neural network model, six neurons are arranged on a first hidden function layer, three neurons are arranged on a second hidden function layer, two neurons are arranged on an output layer, a time point when bubbles appear, the density of the bubbles, the size of the bubbles, the duration of the bubbles, the pulsating pressure of water flow and the temperature of the water flow are calculated by the first hidden function layer, a monitoring result is obtained by comparing the second hidden function layer, the output layer outputs the monitoring result to the outside of the feedforward neural network model, and the processing result of ultrasonic signal information is dangerous cavitation or non-dangerous cavitation.
By adopting the embodiment of the method for monitoring the cavitation state of the centrifugal pump, ultrasonic waves are emitted to detect the water vapor bubbles of water flow in the water outlet pipeline of the centrifugal pump, ultrasonic information is collected and calculated by the feedforward neural network model, so that the monitoring result of dangerous cavitation or non-dangerous cavitation of the centrifugal pump is obtained, the monitoring result is used for reference of staff, the staff can judge the working state of the centrifugal pump more accurately after combining the information such as the working time, the vibration frequency, the flow pressure, the lift, the voltage, the flow, the power and the like of the centrifugal pump, and certainly, the staff can be informed of the dangerous cavitation in time to treat the dangerous cavitation by means of audible and visual alarm, communication with the staff terminal and the like after the monitoring result of dangerous cavitation of the centrifugal pump is obtained.
The application also provides a cavitation state monitoring device of the centrifugal pump.
As shown in fig. 7, a cavitation state monitoring device for a centrifugal pump includes:
the acquisition module is used for acquiring ultrasonic signal information of water flow in a water outlet pipeline of the centrifugal water pump;
the data preprocessing module is used for preprocessing the ultrasonic signal information to obtain ultrasonic signal information data, wherein the ultrasonic signal information data comprises waveform signals, counting signals and time difference signals;
the calculation module is used for inputting the ultrasonic signal information data into a feedforward neural network model, and the feedforward neural network model is used for calculating real-time state data of water flow;
and the comparison module is used for comparing the real-time state data with preset state data to obtain a monitoring result of the cavitation state of the centrifugal water pump.
In the embodiments provided herein, it should be understood that the provided methods and apparatus may be implemented in other ways. For example, the system embodiments described above are merely illustrative; for example, a division of a module is merely a logical function division, and there may be another division manner in actual implementation, for example, multiple modules may be combined or may be integrated into another device, or some features may be omitted or not performed.
The centrifugal pump cavitation state monitoring device can realize any one of the above-mentioned methods, and the specific working process of each module in the centrifugal pump cavitation state monitoring device can refer to the corresponding process in the above-mentioned method embodiment.
The embodiment of the application also discloses a computer device.
Computer apparatus comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor implementing a method of monitoring the cavitation condition of a centrifugal pump as described above when executing the computer program.
The embodiment of the application also discloses a computer readable storage medium.
A computer readable storage medium storing a computer program capable of being loaded by a processor and executing any one of the methods of monitoring a cavitation condition of a centrifugal pump as described above.
Wherein a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device; program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.
In the foregoing embodiments, the descriptions of the embodiments are focused on, and for those portions of one embodiment that are not described in detail, reference may be made to the related descriptions of other embodiments.
The foregoing description of the preferred embodiments of the present application is not intended to limit the scope of the application, in which any feature disclosed in this specification (including abstract and drawings) may be replaced by alternative features serving the same, equivalent or similar purpose, unless expressly stated otherwise. That is, each feature is one example only of a generic series of equivalent or similar features, unless expressly stated otherwise.
Claims (9)
1. A method for monitoring cavitation of a centrifugal pump, the method comprising:
collecting ultrasonic signal information from a liquid medium in a centrifugal pump outlet pipeline;
preprocessing the ultrasonic signal information to obtain ultrasonic signal information data, wherein the ultrasonic signal information data comprises a waveform signal, a counting signal and a time difference signal;
inputting ultrasonic signal information data into a pre-trained neural network model, calculating real-time state data of a liquid medium, comparing the real-time state data with preset state data, performing logic operation, and judging the cavitation state of the centrifugal pump to obtain a processing result of the ultrasonic signal information;
and outputting the processing result of the ultrasonic signal information.
2. A method for monitoring cavitation of a centrifugal pump according to claim 1, wherein the preprocessing step includes:
collecting voltage signals in ultrasonic signal information;
analog-to-digital conversion is carried out on the voltage signal;
and comparing the voltage signal after analog-digital conversion with a reference signal to obtain the waveform signal, the counting signal and the time difference signal.
3. The method for monitoring cavitation of a centrifugal pump according to claim 1, wherein: the real-time state data comprises the time point when the bubbles appear in the liquid medium, the density of the bubbles, the size of the bubbles, the duration of the bubbles, the pulsation pressure of the liquid medium and the temperature of the liquid medium;
and calculating the waveform signal, the counting signal and the time difference signal through a neural network model to obtain the occurrence time point of the bubbles, the density of the bubbles, the size of the bubbles, the duration of the bubbles, the pulsation pressure of the liquid medium and the temperature of the liquid medium.
4. A method for monitoring cavitation status of a centrifugal pump according to claim 3, wherein the preset status data includes a time point when the dangerous cavitation occurs and when the non-dangerous cavitation occurs, a density of the bubbles, a size of the bubbles, a duration of the bubbles, a pulsating pressure of the liquid medium, and a temperature of the liquid medium;
the step of obtaining the preset state data comprises the following steps:
presetting a plurality of time periods and a plurality of liquid medium flow rates, and taking the working condition of the centrifugal pump which is in a preset liquid medium flow rate in a preset time period as a preset working condition;
collecting ultrasonic signal information of liquid medium in a centrifugal pump outlet pipeline under each preset working condition;
judging cavitation state of the centrifugal pump under each preset working condition and calculating state data of the liquid medium according to corresponding ultrasonic signal information;
and storing the cavitation state of the centrifugal pump and the state data of the corresponding liquid medium under all preset working conditions as preset state data.
5. A method for monitoring cavitation of a centrifugal pump according to claim 3, wherein: the neural network model is a feedforward neural network model and comprises an input layer, a hidden function layer and an output layer; the input layer is used for transmitting the waveform signals, the counting signals and the time difference signals into the hidden function layer; the hidden function layer is used for calculating real-time state data, comparing the real-time state data with preset state data and performing logic operation; the output layer is used for outputting an operation result outside the feedforward neural network model;
the step of obtaining the operation result comprises the following steps:
calculating real-time state data;
comparing the calculated real-time state data with preset state data under preset working conditions, and performing logic operation on the comparison result to obtain an operation result of dangerous cavitation/non-dangerous cavitation;
the step of calculating real-time status data includes:
calculating the occurrence time point of bubbles in the liquid medium and the density of the bubbles according to the waveform signals;
calculating the size of bubbles in the liquid medium and the duration of bubbles according to the counting signals;
the pulsating pressure of the liquid medium and the temperature of the liquid medium are calculated from the time difference signal.
6. The method of claim 5, further comprising, before the input layer receives the waveform signal, the count signal, and the time difference signal:
the feedforward neural network model randomly initializes the weight matrix of the feedforward neural network model and selects a nonlinear optimization algorithm as an optimization algorithm;
after the output layer outputs the operation result, the method further comprises:
and calculating a loss function by the feedforward neural network model, debugging a weight matrix by using a back propagation method, and stopping iteration when the feedforward neural network model iterates to a preset number of times.
7. A centrifugal pump cavitation state monitoring device, characterized in that the monitoring device comprises:
the acquisition module is used for acquiring ultrasonic signal information of the liquid medium in the centrifugal pump outlet pipeline;
the data preprocessing module is used for preprocessing the ultrasonic signal information to obtain ultrasonic signal information data, wherein the ultrasonic signal information data comprises a waveform signal, a counting signal and a time difference signal;
the calculation module is used for inputting the ultrasonic signal information data into a neural network model, and the neural network model is used for calculating the real-time state data of the liquid medium;
and the comparison module is used for comparing the real-time state data with preset state data to obtain a monitoring result of the cavitation state of the centrifugal pump.
8. A computer device, characterized by: comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the method of monitoring the cavitation status of a centrifugal pump according to any of claims 1-6 when the program is executed.
9. A computer-readable storage medium, characterized by: a computer program which can be loaded by a processor and which performs the method for monitoring the cavitation status of a centrifugal pump according to any of claims 1-6.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2023109111738 | 2023-07-24 | ||
| CN202310911173.8A CN116816694A (en) | 2023-07-24 | 2023-07-24 | Method, device, equipment and storage medium for monitoring cavitation state of centrifugal pump |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117345660A true CN117345660A (en) | 2024-01-05 |
Family
ID=88122128
Family Applications (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310911173.8A Withdrawn CN116816694A (en) | 2023-07-24 | 2023-07-24 | Method, device, equipment and storage medium for monitoring cavitation state of centrifugal pump |
| CN202311574879.6A Pending CN117345660A (en) | 2023-07-24 | 2023-11-23 | Method, device, equipment and storage medium for monitoring cavitation state of centrifugal pump |
Family Applications Before (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310911173.8A Withdrawn CN116816694A (en) | 2023-07-24 | 2023-07-24 | Method, device, equipment and storage medium for monitoring cavitation state of centrifugal pump |
Country Status (1)
| Country | Link |
|---|---|
| CN (2) | CN116816694A (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117231524B (en) * | 2023-11-14 | 2024-01-26 | 浙江嘉源和达水务有限公司 | Pump cavitation state monitoring and diagnosing method and system |
-
2023
- 2023-07-24 CN CN202310911173.8A patent/CN116816694A/en not_active Withdrawn
- 2023-11-23 CN CN202311574879.6A patent/CN117345660A/en active Pending
Also Published As
| Publication number | Publication date |
|---|---|
| CN116816694A (en) | 2023-09-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117345660A (en) | Method, device, equipment and storage medium for monitoring cavitation state of centrifugal pump | |
| CN104976518A (en) | Subsea pipeline leakage monitoring system | |
| CN105240187B (en) | The method that turbine state monitoring and fault diagnosis are realized based on chaology | |
| CN101592288B (en) | Method for identifying pipeline leakage | |
| CN108804740B (en) | Long-distance pipeline pressure monitoring method based on integrated improved ICA-KRR algorithm | |
| CN207215437U (en) | A kind of ultrasonic wave stress detection device | |
| KR102005138B1 (en) | Device abnormality presensing method and system using thereof | |
| CN103048330A (en) | Visual detecting device for surface crack of underwater structure | |
| CN104019849A (en) | Steel penstock and expansion joint operating parameter automatic monitoring system and evaluation method thereof | |
| CN101718396A (en) | Method and device for detecting leakage of fluid conveying pipeline based on wavelet and mode identification | |
| CN110700810B (en) | Drilling platform safety system for testing and monitoring method thereof | |
| CN116625683A (en) | Wind turbine generator system bearing fault identification method, system and device and electronic equipment | |
| CN111695465A (en) | Pipe network fault diagnosis and positioning method and system based on pressure wave mode identification | |
| CN101360980A (en) | Method and device for detecting the position of a pulsed mechanical effect on a plant component | |
| KR20150071392A (en) | Diagnosis system for checking a clogging in a pipeline using acoustic transfer matrix | |
| CN103575482A (en) | Equipment for diagnosing blockage of lead pipe and method for diagnosing blockage of lead pipe | |
| JPH0484754A (en) | Method for evaluating soundness of material | |
| CN102174992A (en) | Pressure curve feature extraction method for pressure pipeline | |
| CN112857798A (en) | Multi-axis mechanical transmission system quality evaluation method based on frequency spectrum analysis | |
| CN107281687B (en) | Fire hydrant water condition detection method | |
| Zhu et al. | Blocking detection of underground drainage pipe based on CEEMD and GG clustering | |
| Liu et al. | Cavitations monitoring and diagnosis of hydropower turbine on line based on vibration and ultrasound acoustic | |
| CN106401989B (en) | A kind of centrifugal pump cavitation monitoring device | |
| CN116642957B (en) | Foundation pile ultrasonic detection interpretation method based on multiple acoustic parameters | |
| CN202794110U (en) | Ship pipe system leakage detecting device based on ultrasonic surface wave transmission property |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |