CN115342063A - Centrifugal pump energy consumption prediction method based on empirical mode decomposition and BP neural network - Google Patents
Centrifugal pump energy consumption prediction method based on empirical mode decomposition and BP neural network Download PDFInfo
- Publication number
- CN115342063A CN115342063A CN202211072414.6A CN202211072414A CN115342063A CN 115342063 A CN115342063 A CN 115342063A CN 202211072414 A CN202211072414 A CN 202211072414A CN 115342063 A CN115342063 A CN 115342063A
- Authority
- CN
- China
- Prior art keywords
- centrifugal pump
- signal
- energy consumption
- pump
- energy
- 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
- 238000005265 energy consumption Methods 0.000 title claims abstract description 59
- 238000000034 method Methods 0.000 title claims abstract description 28
- 238000000354 decomposition reaction Methods 0.000 title claims abstract description 26
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 15
- 238000005457 optimization Methods 0.000 claims abstract description 7
- 238000012549 training Methods 0.000 claims abstract description 6
- 239000013598 vector Substances 0.000 claims description 22
- 239000000463 material Substances 0.000 claims description 4
- 238000012935 Averaging Methods 0.000 claims description 2
- 239000003989 dielectric material Substances 0.000 claims description 2
- 238000010606 normalization Methods 0.000 abstract description 2
- 238000012545 processing Methods 0.000 abstract description 2
- 230000004913 activation Effects 0.000 description 4
- 238000011156 evaluation Methods 0.000 description 2
- 238000013210 evaluation model Methods 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 239000012530 fluid Substances 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F05—INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
- F05D—INDEXING SCHEME FOR ASPECTS RELATING TO NON-POSITIVE-DISPLACEMENT MACHINES OR ENGINES, GAS-TURBINES OR JET-PROPULSION PLANTS
- F05D2270/00—Control
- F05D2270/70—Type of control algorithm
- F05D2270/709—Type of control algorithm with neural networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Mechanical Engineering (AREA)
- Control Of Non-Positive-Displacement Pumps (AREA)
Abstract
The invention discloses a centrifugal pump energy consumption prediction method based on empirical mode decomposition and a BP neural network, which comprises the following steps: acquiring a plurality of operating state parameters of the centrifugal pump under a plurality of operating conditions; respectively carrying out empirical mode decomposition on the collected pump body vibration signal and the collected noise signal to obtain a new signal; calculating the heat energy lost when the centrifugal pump runs according to the collected temperature of the whole pump; obtaining loss eigenvectors of vibration energy, heat energy and sound energy after normalization processing, and using the loss eigenvectors as input of a centrifugal pump energy consumption prediction model; establishing a centrifugal pump energy consumption prediction model based on a BP neural network, and updating the weight through an RMSprop optimization algorithm; and inputting a plurality of running state parameters of the centrifugal pump to be tested into the trained centrifugal pump energy consumption prediction model to obtain the energy consumption value of the centrifugal pump. According to the method, empirical mode decomposition is carried out on the vibration signals and the noise signals, the RMSprop optimization algorithm is adopted to update the weight value of the centrifugal pump energy consumption prediction model during training, and the accuracy of subsequent centrifugal pump energy consumption prediction is improved.
Description
Technical Field
The invention relates to the field of centrifugal pump energy consumption evaluation, in particular to a centrifugal pump energy consumption prediction method based on empirical mode decomposition and a BP neural network.
Background
The centrifugal pump is used as the most widely applied fluid machinery, according to statistics, the power consumption of the pump accounts for about 20% of the generated energy in China, and the energy consumption of the pump is a main component of industrial energy consumption. The centrifugal pump has the common problems of high energy consumption and low electric energy utilization rate, the energy transmitted to the medium by the centrifugal pump is far more than the energy required by the transmission of the medium, and the rest energy is distributed in vibration loss, heat energy loss and sound energy loss. The energy consumption process of the centrifugal pump is complex, if the energy consumption of each part of the centrifugal pump loss can be accurately evaluated, the reason of high energy consumption can be effectively analyzed, corresponding solving measures can be found, the energy-saving consciousness of users and producers can be enhanced, the energy waste is reduced, the electric energy utilization rate is improved, and the process flow is integrally improved. At present, energy consumption evaluation and prediction of a centrifugal pump mainly take data point fitting as a main part, and the energy consumption dependency relationship of the centrifugal pump is obtained by analyzing the nonlinear relationship between different influencing factors and energy consumption, so that the energy consumption of the centrifugal pump is predictedTo pairThe degree of dependence of the output and input parameters is large.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides the following technical scheme:
a centrifugal pump energy consumption prediction method based on empirical mode decomposition and a BP neural network comprises the following steps:
s1: acquiring a plurality of operating state parameters of the centrifugal pump in a certain period of time under a plurality of operating conditions including rated conditions; the operating state parameters at least comprise the temperature of the whole pump, a pump body vibration signal and a noise signal;
s2: respectively decomposing the collected pump body vibration signal and the noise signal by using an empirical mode decomposition method to obtain a plurality of IMF signals and a residual signal, discarding the residual signal, adding all the residual IMF signals to obtain a new signal, and extracting the energy characteristic of the new signal to obtain the energy characteristic of the pump body vibration signal and the energy characteristic of the noise signal;
calculating the heat energy lost when the centrifugal pump runs according to the collected temperature of the whole pump;
normalizing the operation state parameters except the whole pump temperature, the pump body vibration signal and the noise signal, and taking the energy characteristic of the pump body vibration signal, the energy characteristic of the noise signal and the heat energy lost when the centrifugal pump operates as labels, wherein the operation state parameters after normalization respectively form a vibration energy loss characteristic vector, a heat energy loss characteristic vector and a sound energy loss characteristic vector;
s3: establishing a centrifugal pump energy consumption prediction model based on a BP neural network, and updating a weight through an RMSprop optimization algorithm; respectively inputting the vibration energy loss characteristic vector, the heat energy loss characteristic vector and the sound energy loss characteristic vector obtained in the step S2 into a centrifugal pump energy consumption prediction model to finish training the centrifugal pump energy consumption prediction model;
s4: inputting a plurality of running state parameters of the centrifugal pump to be tested in a certain running state within a certain time period into the trained centrifugal pump energy consumption prediction model to respectively obtain vibration energy loss, heat energy loss and sound energy loss; the energy loss of the final centrifugal pump, i.e. the energy consumption value, is the sum of the three losses.
Further, the operation state parameter in the step S1 includes a torque x i1 Inlet pressure x i2 Outlet pressure x i3 Flow x i4 Temperature x of the entire pump i5 Motor speed x i6 Input current x i7 Input voltage x i8 And pump body vibration signal x i9 Noise signal x i10 。
Further, in step S2, decomposing the collected pump body vibration signal and the collected noise signal by using an empirical mode decomposition method, respectively, to obtain a plurality of IMF signals, which specifically includes:
(1) Solving upper and lower extreme points of the signal;
(2) Connecting all upper extreme points by using a cubic spline interpolation method to form an upper envelope curve, and connecting all lower extreme points to form a lower envelope curve;
(3) Averaging values of corresponding points on the upper envelope line and the lower envelope line to obtain an average value curve;
(4) Subtracting the value of the corresponding point on the mean curve from the original signal to obtain a difference signal; judging whether the difference signal meets the characteristic condition, if so, taking the difference signal as a first-order IMF1 component; if not, returning to the step (1);
the characteristic conditions are as follows: the average value of all points on the average curve is 0; in one period, the difference between the number of local extreme points of the upper envelope curve, the lower envelope curve and the mean curve and the number of curve zero points is one at most;
(5) Subtracting the IMF1 signal of the previous order from the original signal, returning the obtained signal to the step (1) for decomposition to obtain a second-order IMF2 component of the original signal; and by analogy, obtaining the rest IMF components until the obtained signal is a monotone function, and finishing the decomposition process.
Further, the step S2 is performed according to the collected whole pump temperature x i5 Calculating the heat energy delta Q lost when the centrifugal pump operates comprises the following steps:
ΔQ＝Q 1 +Q 2
Q 1 ＝m 1 c 1 Δt
Q 2 ＝m 2 c 2 Δt
in the formula, Q 1 M is the heat lost during the operation of the pump body of the centrifugal pump 1 For the mass of the pump body of the centrifugal pump,. DELTA.t is the difference between the temperatures before and after operation, c 1 The specific heat capacity of the material of the pump body of the centrifugal pump; q 2 M is the heat lost during the operation of the centrifugal pump 2 Mass of the medium in the pump body of the centrifugal pump, c 2 Is a centrifugalThe specific heat capacity of the dielectric material in the pump body.
Further, the centrifugal pump energy consumption prediction model based on the BP neural network in the step S3 includes an input layer, a hidden layer, and an output layer;
configuring the number of nodes of an input layer as q and the dimension of a characteristic vector;
the number of nodes of an output layer is configured to be j, namely, the energy consumption value of the centrifugal pump is only output;
the number of nodes configuring the hidden layer is
Wherein q is the number of nodes of the input layer, j is the number of nodes of the output layer, p is the number of nodes of the hidden layer, a is any integer from 1 to 10,
to round the symbol down.
The invention has the beneficial effects that:
1. a neural network model is adopted to establish a mapping relation between the running state parameters of the centrifugal pump and the energy loss of the centrifugal pump, so that the energy loss generated by the centrifugal pump can be effectively predicted.
2. The vibration signals and the noise signals are decomposed by adopting an empirical mode, residual signals are removed, the obtained IMF components form effective signals, the energy of the effective signals is extracted, and the accuracy of energy consumption prediction of the subsequent centrifugal pump is improved.
3. When the energy consumption prediction model of the centrifugal pump is trained, the weight is updated by adopting the RMSprop optimization algorithm, so that the accuracy of the energy consumption prediction model can be improved.
Drawings
FIG. 1 is a flow chart of energy consumption prediction for a centrifugal pump.
FIG. 2 is a model for energy consumption prediction.
Detailed Description
The present invention will be described in detail below with reference to the accompanying drawings and preferred embodiments, and the objects and effects of the present invention will become more apparent, and the present invention will be described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
As shown in FIG. 1, the method for predicting the energy consumption of the centrifugal pump based on empirical mode decomposition and the BP neural network comprises the following steps:
step one, collecting running state data of the centrifugal pump.
1.1, determining a plurality of commonly used operating states of a certain centrifugal pump, wherein the operating states comprise operating states under rated working conditions and partial working conditions, and the operating states are marked as states i, i =1,2, \ 8230;, n.
1.2 determining 10 operating state parameters of the centrifugal pump within a certain time period t under a certain operating state, specifically recording as: torque x i1 Inlet pressure x i2 Outlet pressure x i3 Flow rate x i4 Temperature x of the entire pump i5 Motor speed x i6 Input current x i7 Input voltage x i8 Vibration signal x of pump body i9 Of a noise signal x i10 。
1.3 simulation experiment is carried out on a certain centrifugal pump, the centrifugal pump is respectively operated under a plurality of operation states in the step 1.1, the operation state parameters in the step 1.2 are collected in a certain time period t, wherein the temperature x of the whole pump i5 Measured by an infrared temperature sensor, a pump body vibration signal x i9 Measured by an acceleration sensor and the remaining parameters by the corresponding sensors.
1.4 to eliminate the interference of external factors, the centrifugal pump is operated to repeat 3 times of tests under each operation state.
Step two, preprocessing the collected various parameter data
2.1 pairs of data x in step 1.2 i1 ，x i2 ，x i3 ，x i4 ，x i6 ，x i7 ，x i8 The average values are respectively taken.
2.2 Pump body vibration Signal x collected i9 And a noise signal x i10 The data in (1) is a continuous signal.
Using Empirical Mode Decomposition (EMD) on the pump body vibration signal x i9 Sum noise signalx i10 Processing to vibrate the signal x i9 For example, the specific steps are as follows:
(1) Determining a vibration signal x i9 The upper and lower extreme points of (1);
(2) Respectively connecting all the upper extreme points and the lower extreme points by using a cubic spline interpolation method to form an upper envelope line x and a lower envelope line x i91 、x i92 ；
(3) Adding the values of all the points on the envelope curve formed by connecting the two extreme points to obtain an average value curve h i9 ：
(4) Original pump body vibration signal x i9 Subtracting the value of the corresponding point on the mean curve to obtain a signal S i91 。
S i91 ＝x i9 -h i9
The characteristic conditions of the IMF component are: the average value of all points on the average value curve formed by connecting the upper extreme point and the lower extreme point must be 0; the number of local extremum points of the curve in a period needs to be different from the number of zero-crossing points by one at most.
If the signal S i91 If the characteristic conditions are not met, the decomposition is ended, and the step (1) of the step 2.2 is returned. If the characteristic condition is satisfied, the signal S i91 As the first order IMF1.
(5) The original pump body vibration signal x i9 Subtracting the IMF1 signal of the previous order to obtain a signal U i91 Then returning to the step (1) of the step 2.2 for decomposition to obtain an original pump body vibration signal x i9 The second order IMF2 component of (a). Repeating the steps to obtain the rest components: s i91 ，S i92 ，…，S i9n . Up to S i9n The decomposition process ends for a monotonic function. Obtaining an original vibration signal x i9 All IMF components of (a).
2.3EMD decomposition to obtain several IMF signals and a residual signal, discarding the residual signal, adding all the obtained IMF signals to obtain new signal x ib 。
2.4 extraction of x according to the formula ib Corresponding energy characteristics, denoted E ib ：
Where b =9,10,i is the corresponding operating state in step 1.1, i =1,2, \8230, and n, t are the sensor sampling times.
Obtaining the energy characteristic E of the vibration signal of the pump body i9 Energy characteristic E of noise signal i10 。
2.5 according to the whole pump temperature x i5 Calculating the heat energy delta Q lost when the centrifugal pump operates:
ΔQ＝Q 1 +Q 2
Q 1 ＝m 1 c 1 Δt
Q 2 ＝m 2 c 2 Δt
in the formula, Q 1 The heat quantity m lost when the pump body of the centrifugal pump operates 1 For the mass of the pump body of the centrifugal pump,. DELTA.t is the difference between the temperatures before and after operation, c 1 The specific heat capacity of the material of the pump body of the centrifugal pump; q 2 M is the heat lost during operation of the centrifugal pump 2 Mass of the medium in the pump body of the centrifugal pump, c 2 Is the specific heat capacity of the medium material in the pump body of the centrifugal pump.
2.6 constructing eigenvectors for vibrational energy losses
In the formula, x 1 Dividing the collected operating state parameter by the whole pump temperature x i5 And pump body vibration signal x i9 Noise signal x i10 External operating state parameter, y 1 Is a loss of vibrational energy.
2.7 constructing a feature vector for heat energy loss:
in the formula, y 2 Is a loss of thermal energy.
2.8 constructing a feature vector for acoustic energy loss:
in the formula, y 3 Is a loss of acoustic energy.
2.9 vibration energy losses D for step 2.6, step 2.7, step 2.8 i1 Heat energy loss D i2 Acoustic energy loss D i3 Normalizing the operation state parameters to [0,1 ]]In the range of (a) to (b),
in which i represents the respective operating state parameter, x i As raw data of corresponding state parameters, x i ' is data normalized by corresponding state parameter, x imin Is the minimum value, x, in the raw data corresponding to the state parameter imax Which is the maximum value in the raw data corresponding to the state parameter.
2.10 obtaining normalized feature vector D i1 ′，D i2 ′，D i3 ′。
Three-step centrifugal pump energy consumption prediction model establishment
3.1 configuring relevant parameters of the prediction model, and specifically comprising the following steps:
(1) And configuring the number of nodes of the input layer as q, namely the dimension of the feature vector.
(2) And the number of the nodes of the output layer is j, namely the energy consumption value of the centrifugal pump.
(3) And (3) configuring the number of the nodes of the hidden layer by combining an empirical formula and a trial and error method:
in the formula, p is the number of hidden layer nodes, q is the number of input layer nodes, j is the number of output layer nodes (namely 1), and a is any integer from 1 to 10. The architecture of the centrifugal pump energy consumption prediction model is shown in fig. 2.
3.2 selecting an activation function to carry out data training on the energy consumption prediction model of the centrifugal pump, and respectively inputting the normalized vibration energy loss characteristic vector D obtained in the step 2.10 into the energy consumption prediction model of the centrifugal pump i1 ', heat energy loss eigenvector D i2 ', acoustic energy loss eigenvector D i3 ', and updating the weight value through an RMSprop optimization algorithm. The method comprises the following specific steps:
(1) Selecting Relu activation function to configure the hidden layer, wherein each node input of the hidden layer is R j ，w qp For the data connection between the input layer and the hidden layer:
where x is output layer output data, b 1 Activation threshold for hidden layer node, w qp The weights of the input layer and the hidden layer are connected, q is the number of nodes of the input layer, and p is the number of nodes of the hidden layer.
(2) The output of each node of the hidden layer is:
y j ＝f(R j )
(3) The output of each node of the output layer is:
in the formula, b 2 Being the activation threshold of the output layer node, w jp Are weights between the hidden layer and the output layer.
(4) Calculating the output error t m ：
In the formula, Z m For the actual output of the mth neuron during training,
is the desired output of the mth neuron.
3.3 comparing the output error t of the step (4) in the step 3.2 m And the set error T m Contrast if t m ≤T m If the energy consumption of the centrifugal pump does not meet the requirement of the error, the energy consumption of the centrifugal pump is predicted, and if the energy consumption of the centrifugal pump does not meet the requirement of t m ≤T m If the condition is met, updating the weight value through the RMSprop optimization algorithm, returning to the step (1) in the step 3.2 for inheritance, and continuing until the error meets the requirement of t m ≤T m . The RMSprop algorithm is shown below:
s＝βs+(1-β)dw 2
where s is the exponentially weighted average, w is the weight, α is a constant, β is set to 0.999, ε is set to 10 -8 。
Step four centrifugal pump energy consumption prediction implementation
The step one to the step three are carried out on the centrifugal pump to be predicted to obtain a trained energy consumption evaluation model of the centrifugal pump, a torque signal, an inlet pressure signal, an outlet pressure signal, a flow signal, a whole pump temperature signal, a motor rotating speed signal and an input current signal are measured according to the working condition to be predicted and input into the energy evaluation model of the centrifugal pump to respectively obtain the predicted vibration energy loss y of the centrifugal pump 1 Heat energy loss y 2 Loss of sound energy y 3 . The energy loss, i.e. the energy consumption value of the final centrifugal pump is y = y 1 +y 2 +y 3 。
The method provided by the invention takes torque signals, inlet pressure signals, outlet pressure signals, flow signals, whole pump temperature signals, motor rotating speed signals, input current signals, input voltage signals, pump body vibration signals, pump body noise signals and other signals of the centrifugal pump under different operating conditions as original data, EMD decomposition is carried out on the vibration signals and the noise signals, all IMFs are added to obtain effective vibration and noise signals, then the energy of the effective vibration and noise signals is calculated, and meanwhile, the heat energy loss is calculated. The method comprises the steps of combining effective vibration, noise signals, heat loss and torque signals, inlet pressure signals, outlet pressure signals, flow signals, whole pump temperature signals, motor rotating speed signals and input current signals to construct feature vectors, normalizing the feature vectors, inputting the normalized feature vectors into a centrifugal pump energy consumption prediction model to conduct model training and prediction, and finally obtaining a trained centrifugal pump energy consumption prediction model for actual detection.
It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and although the invention has been described in detail with reference to the foregoing examples, it will be apparent to those skilled in the art that various changes in the form and details of the embodiments may be made and equivalents may be substituted for elements thereof. All modifications, equivalents and the like which come within the spirit and principle of the invention are intended to be included within the scope of the invention.
Claims (5)
1. A centrifugal pump energy consumption prediction method based on empirical mode decomposition and a BP neural network is characterized by comprising the following steps:
s1: acquiring a plurality of operating state parameters of the centrifugal pump in a certain period of time under a plurality of operating conditions including rated conditions; the operating state parameters at least comprise the temperature of the whole pump, a pump body vibration signal and a noise signal;
s2: decomposing the collected pump body vibration signal and the noise signal by using an empirical mode decomposition method to obtain a plurality of IMF signals and a residual signal, discarding the residual signal, adding all the residual IMF signals to obtain a new signal, and extracting the energy characteristic of the new signal to obtain the energy characteristic of the pump body vibration signal and the energy characteristic of the noise signal;
calculating the heat energy lost when the centrifugal pump operates according to the collected temperature of the whole pump;
normalizing the operation state parameters except the whole pump temperature, the pump body vibration signal and the noise signal, and taking the energy characteristic of the pump body vibration signal, the energy characteristic of the noise signal and the heat energy lost when the centrifugal pump operates as labels, wherein the normalized operation state parameters respectively form a vibration energy loss characteristic vector, a heat energy loss characteristic vector and a sound energy loss characteristic vector;
s3: establishing a centrifugal pump energy consumption prediction model based on a BP neural network, and updating the weight through an RMSprop optimization algorithm; respectively inputting the vibration energy loss characteristic vector, the heat energy loss characteristic vector and the sound energy loss characteristic vector obtained in the step S2 into a centrifugal pump energy consumption prediction model to finish training the centrifugal pump energy consumption prediction model;
s4: inputting a plurality of operating state parameters of the centrifugal pump to be tested in a certain operating state within a certain time period into a trained centrifugal pump energy consumption prediction model to respectively obtain vibration energy loss, heat energy loss and sound energy loss; the energy loss of the final centrifugal pump, i.e. the energy consumption value, is the sum of the three losses.
2. The method of claim 1, wherein the operating state parameter in step S1 comprises torque x i1 Inlet pressure x i2 Outlet pressure x i3 Flow x i4 Temperature x of the entire pump i5 Motor speed x i6 Input current x i7 Input voltage x i8 And pump body vibration signal x i9 Noise signal x i10 。
3. The method for predicting the energy consumption of the centrifugal pump based on the empirical mode decomposition and the BP neural network according to claim 1, wherein in the step S2, the collected pump body vibration signal and the collected noise signal are decomposed by using an empirical mode decomposition method to obtain a plurality of IMF signals, and specifically comprises:
(1) Solving upper and lower extreme points of the signal;
(2) Connecting all upper extreme points by using a cubic spline interpolation method to form an upper envelope line, and connecting all lower extreme points to form a lower envelope line;
(3) Averaging values of corresponding points on the upper envelope line and the lower envelope line to obtain an average curve;
(4) Subtracting the value of the corresponding point on the mean curve from the original signal to obtain a difference signal; judging whether the difference signal meets the characteristic condition, if so, taking the difference signal as a first-order IMF1 component; if not, returning to the step (1);
the characteristic conditions are as follows: the average value of all points on the average value curve is 0; in one period, the difference between the number of local extreme points of the upper envelope curve, the lower envelope curve and the mean curve and the number of curve zero points is one at most;
(5) Subtracting the IMF1 signal of the previous order from the original signal, and returning the obtained signal to the step (1) for decomposition to obtain a second-order IMF2 component of the original signal; and by analogy, obtaining the rest IMF components until the obtained signal is a monotone function, and finishing the decomposition process.
4. The method for predicting the energy consumption of a centrifugal pump based on Empirical Mode Decomposition (EMD) and BP neural network according to claim 1, wherein the step S2 is performed according to the collected whole pump temperature x i5 Calculating the heat energy delta Q lost when the centrifugal pump operates comprises the following steps:
ΔQ＝Q 1 +Q 2
Q 1 ＝m 1 c 1 Δt
Q 2 ＝m 2 c 2 Δt
in the formula, Q 1 The heat quantity m lost when the pump body of the centrifugal pump operates 1 Δ t is the difference between the temperature before and after operation, c 1 The specific heat capacity of the material of the pump body of the centrifugal pump; q 2 M is the heat lost during operation of the centrifugal pump 2 Mass of the medium in the pump body of the centrifugal pump, c 2 Is a centrifugalThe specific heat capacity of the dielectric material in the pump body.
5. The centrifugal pump energy consumption prediction method based on Empirical Mode Decomposition (EMD) and BP neural network as claimed in claim 1, wherein the BP neural network based centrifugal pump energy consumption prediction model in step S3 comprises an input layer, an implicit layer and an output layer;
configuring the number of nodes of an input layer as q and the dimension of a characteristic vector;
the number of nodes of an output layer is configured to be j, namely, the energy consumption value of the centrifugal pump is only output;
the number of nodes configuring the hidden layer is
Wherein q is the number of input layer nodes, j is the number of output layer nodes, p is the number of hidden layer nodes, a is any integer from 1 to 10,
to round the symbol down.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211072414.6A CN115342063A (en) | 2022-09-02 | 2022-09-02 | Centrifugal pump energy consumption prediction method based on empirical mode decomposition and BP neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211072414.6A CN115342063A (en) | 2022-09-02 | 2022-09-02 | Centrifugal pump energy consumption prediction method based on empirical mode decomposition and BP neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115342063A true CN115342063A (en) | 2022-11-15 |
Family
ID=83955506
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211072414.6A Pending CN115342063A (en) | 2022-09-02 | 2022-09-02 | Centrifugal pump energy consumption prediction method based on empirical mode decomposition and BP neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115342063A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106126896A (en) * | 2016-06-20 | 2016-11-16 | 中国地质大学（武汉） | The mixed model wind speed forecasting method learnt based on empirical mode decomposition and the degree of depth and system |
| CN109647899A (en) * | 2018-12-26 | 2019-04-19 | 北京科技大学 | More specification rolled piece power consumption forecasting procedures in a kind of hot strip rolling finishing stands |
| CN111709585A (en) * | 2020-06-19 | 2020-09-25 | 国网上海市电力公司 | Air conditioner load prediction method and device and storage medium |
| CN112098128A (en) * | 2020-09-08 | 2020-12-18 | 天津大学 | Power mechanical equipment fault and energy consumption analysis method based on noise and vibration |
| CN113807615A (en) * | 2021-10-14 | 2021-12-17 | 北京嘉洁能科技股份有限公司 | Electric heating energy consumption prediction method and prediction system based on BP neural network |
| CN114046259A (en) * | 2021-11-12 | 2022-02-15 | 利欧集团浙江泵业有限公司 | Centrifugal pump frequency conversion control method based on double neural network model |
| CN114547987A (en) * | 2022-04-22 | 2022-05-27 | 中国计量大学 | Centrifugal pump turbine performance prediction method based on improved artificial neural network |
-
2022
- 2022-09-02 CN CN202211072414.6A patent/CN115342063A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106126896A (en) * | 2016-06-20 | 2016-11-16 | 中国地质大学（武汉） | The mixed model wind speed forecasting method learnt based on empirical mode decomposition and the degree of depth and system |
| CN109647899A (en) * | 2018-12-26 | 2019-04-19 | 北京科技大学 | More specification rolled piece power consumption forecasting procedures in a kind of hot strip rolling finishing stands |
| CN111709585A (en) * | 2020-06-19 | 2020-09-25 | 国网上海市电力公司 | Air conditioner load prediction method and device and storage medium |
| CN112098128A (en) * | 2020-09-08 | 2020-12-18 | 天津大学 | Power mechanical equipment fault and energy consumption analysis method based on noise and vibration |
| CN113807615A (en) * | 2021-10-14 | 2021-12-17 | 北京嘉洁能科技股份有限公司 | Electric heating energy consumption prediction method and prediction system based on BP neural network |
| CN114046259A (en) * | 2021-11-12 | 2022-02-15 | 利欧集团浙江泵业有限公司 | Centrifugal pump frequency conversion control method based on double neural network model |
| CN114547987A (en) * | 2022-04-22 | 2022-05-27 | 中国计量大学 | Centrifugal pump turbine performance prediction method based on improved artificial neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108920863B (en) | Method for establishing energy consumption estimation model of robot servo system | |
| CN115096590B (en) | Rolling bearing fault diagnosis method based on IWOA-ELM | |
| Hu et al. | A Quasi-ARMAX approach to modelling of non-linear systems | |
| CN112232244A (en) | Fault diagnosis method for rolling bearing | |
| CN109829136B (en) | Method and system for predicting residual life of degradation equipment with random jump | |
| CN110377942B (en) | Multi-model space-time modeling method based on finite Gaussian mixture model | |
| CN108919755B (en) | Distributed fault detection method based on multiple nonlinear cross relation models | |
| CN111709577B (en) | RUL prediction method based on long-range correlation GAN-LSTM | |
| CN112733692B (en) | Fault prediction method and device based on integrated hybrid model and related equipment | |
| CN108492026B (en) | Soft measurement method based on integrated orthogonal component optimization regression analysis | |
| CN112001115B (en) | Soft measurement modeling method of semi-supervised dynamic soft measurement network | |
| CN115342063A (en) | Centrifugal pump energy consumption prediction method based on empirical mode decomposition and BP neural network | |
| Galvão et al. | Linear-wavelet networks | |
| CN113203953A (en) | Lithium battery residual service life prediction method based on improved extreme learning machine | |
| CN114519405B (en) | Process industry multi-sensor data collaborative analysis method and system | |
| CN116821828A (en) | Multi-dimensional time sequence prediction method based on industrial data | |
| CN110794676A (en) | CSTR process nonlinear control method based on Hammerstein-Wiener model | |
| CN111241749A (en) | Permanent magnet synchronous motor chaos prediction method based on reserve pool calculation | |
| CN115901255A (en) | Bearing fault diagnosis method based on parameter optimization VMD-SVM | |
| Larimore et al. | ADAPT-LPV software for identification of nonlinear parameter-varying systems | |
| CN115183969A (en) | Method and system for estimating BWBN model parameters | |
| CN115099268A (en) | Wheeled robot intelligent fault diagnosis method and system based on graph convolution network | |
| CN114357864A (en) | Phase modulator state evaluation method and evaluation system based on fuzzy reasoning | |
| CN114384427A (en) | Lithium ion battery capacity nonlinear degradation prediction method based on quantum neural network | |
| CN112036088A (en) | Superheated steam temperature prediction method based on multi-innovation random gradient optimization |
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 |