CN112052551A - Method and system for identifying surge operation fault of fan - Google Patents
Method and system for identifying surge operation fault of fan Download PDFInfo
- Publication number
- CN112052551A CN112052551A CN201911021826.5A CN201911021826A CN112052551A CN 112052551 A CN112052551 A CN 112052551A CN 201911021826 A CN201911021826 A CN 201911021826A CN 112052551 A CN112052551 A CN 112052551A
- Authority
- CN
- China
- Prior art keywords
- frequency data
- cnn
- time
- layer
- fault
- 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.)
- Granted
Links
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
- F04D27/00—Control, e.g. regulation, of pumps, pumping installations or pumping systems specially adapted for elastic fluids
- F04D27/001—Testing thereof; Determination or simulation of flow characteristics; Stall or surge detection, e.g. condition monitoring
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E10/00—Energy generation through renewable energy sources
- Y02E10/70—Wind energy
- Y02E10/72—Wind turbines with rotation axis in wind direction
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Mechanical Engineering (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention provides a method and a system for identifying surge operation faults of a fan. The identification method comprises the following steps: firstly, acquiring a field device acoustic signal of fan equipment in the working process; preprocessing the field device sound signal to obtain two-dimensional time-frequency data of the field device sound signal; then, identifying the two-dimensional time-frequency data by adopting a CNN network to obtain a first identification result; identifying the two-dimensional time frequency data by adopting an LSTM network to obtain a second identification result; and finally, based on a D-S evidence theory, performing information fusion on the first identification result and the second identification result to obtain an identification result of the surge operation fault of the fan. The method utilizes the field device acoustic signal to diagnose the surge fault of the fan, does not need to install a sensor on the fan, adopts the convolutional neural network and the long-time memory network to diagnose the fault occurrence probability of the characteristics of the surge acoustic signal, and uses the D-S evidence theory to carry out fusion diagnosis of two results of a decision layer, thereby improving the accuracy of diagnosis.
Description
Technical Field
The invention relates to the technical field of equipment fault detection, in particular to a method and a system for identifying a surge operation fault of a fan.
Background
Along with the continuous dilatation of power plant's unit, when satisfying power plant's load requirement, guarantee that the demand of all kinds of power equipment steady operation also increases thereupon. The boiler is characterized in that fan-type equipment such as a feeding fan, a draught fan and a primary fan is used for controlling the pressure of a hearth and ensuring the normal operation of the boiler, and the boiler is an important auxiliary equipment for the high-efficiency operation of a power plant. When the stability of the fan equipment is low, or the opening of the valve fluctuates greatly to cause the running rotating speed of the fan to be not matched with the rotating airflow, the fan is easy to surging and break down, the fan works in an unstable area, pressure and flow pulsation with the characteristics of wide frequency domain and high amplitude are generated, and periodic convective aerodynamic noise is caused. The surge fault brings great potential safety hazard to the stable operation of the unit, so the real-time monitoring of the running state of the fan and the early warning of the surge fault are very important. The existing surge fault early warning method mainly comprises the steps that 1, a starting signal of a fan, a combustion mark in a furnace, an operation current detection signal, an outlet pressure detection signal and an opening degree feedback signal of an outlet adjusting valve are collected through the outlet adjusting valve of the fan and a pressure detection device and are sent to a computer control system for calculation and analysis, and whether the fan has a surge phenomenon or not is identified; 2. performing spectral subtraction denoising on the preprocessed voice signal, forming an input of a 3D convolutional neural Network by adopting a Mel Frequency Cepstral Coefficient (Mel Frequency Cepstral Coefficient) and data stacking, and performing voiceprint recognition by utilizing the Convolutional Neural Network (CNN); 3. the running sound of the automobile engine is processed based on time-frequency two-dimensional processing, fault judgment is carried out according to AlexNet, and meanwhile LSTM is used for identifying the running time state of the engine so as to assist in more accurate fault judgment. However, the prior art has the following technical defects: 1. when surging occurs, the fan has the characteristics of low-load working conditions, severe vibration and the like, and most of the current power plants use the combination of information sensors such as pressure, valve opening, vibration and the like as the basis for judging whether surging occurs. This approach requires multiple sensors to monitor the device in real time, adding to the cost of the device. Meanwhile, the installation requirements of the sensors are different, the placement positions are different, and the difficulty in installation and maintenance of the sensors is increased. 2. The current equipment fault detection and early warning method mostly utilizes deep knowledge of fault characteristics, and identifies and classifies the fault characteristics based on the traditional information processing method, the method needs higher equipment understanding and operation experience, and the generalization is poor, when the equipment has other faults or when a plurality of faults are mixed to appear and the like, the fault type is often difficult to identify, or the identification rate is poor. 3. The equipment fault identification method usually has higher requirements on accuracy rate, and a fault monitoring and early warning algorithm is required to have low false rejection rate and low false acceptance rate. The signal of equipment surge often is periodic variation, and the fault characteristic at surge initial stage takes place and the operation characteristic of fan turn-down load is more similar, only carries out the fault diagnosis on the one hand from static angle and can't reach the requirement of accurate fault early warning, and the rate of accuracy is lower.
Disclosure of Invention
The invention aims to provide a method and a system for identifying a surge operation fault of a fan, so as to improve the identification accuracy rate on the basis of not installing a sensor on fan equipment.
In order to achieve the purpose, the invention provides the following scheme:
a method for identifying a surge operation fault of a fan comprises the following steps:
acquiring a field device acoustic signal of the fan device in the working process;
preprocessing the field device acoustic signal to obtain two-dimensional time-frequency data of the field device acoustic signal;
adopting a CNN network to identify the two-dimensional time-frequency data to obtain a first identification result;
identifying the two-dimensional time frequency data by adopting an LSTM network to obtain a second identification result;
and based on a D-S evidence theory, performing information fusion on the first identification result and the second identification result to obtain an identification result of the surge operation fault of the fan.
Optionally, the preprocessing the field device acoustic signal to obtain two-dimensional time-frequency data of the field device acoustic signal includes:
carrying out Hamming window filtering and discrete Fourier transform on the field device acoustic signal to obtain frequency domain data of the field device acoustic signal;
and establishing two-dimensional time-frequency data of the field device acoustic signal by taking the time domain data of the field device acoustic signal as one dimension and the frequency domain data as the other dimension.
Optionally, the identifying the two-dimensional time-frequency data by using the CNN network to obtain a first identification result specifically includes:
respectively acquiring a normal sound signal when the fan equipment normally operates and a fault sound signal when the fan equipment operates at the initial surge stage, and constructing a normal time-frequency data set and a fault time-frequency data set;
training a CNN model according to the normal time-frequency data set and the fault time-frequency data set to obtain a trained CNN model;
and inputting the two-dimensional time-frequency data into the trained CNN model, and obtaining the output of the trained CNN model as a first recognition result.
Optionally, the training a CNN model according to the normal time-frequency data set and the fault time-frequency data set to obtain a trained CNN model specifically includes:
initializing convolution kernel weights and additive offset vectors of each layer of the CNN model;
respectively inputting the normal time-frequency data in the normal time-frequency data set and the fault time-frequency data in the fault time-frequency data set into a first convolution layer of a CNN (convolutional neural network) model to obtain an output result of the first convolution layer;
inputting the output result of the first convolution layer into a first pooling layer of the CNN model to obtain the output result of the first pooling layer;
inputting the output result of the first pooling layer into a second convolution layer of the CNN model to obtain the output result of the second convolution layer;
inputting the output result of the second convolutional layer into a second pooling layer of the CNN model to obtain the output result of the second pooling layer;
inputting the output result of the second pooling layer into the full-connection layer of the CNN model to obtain the output result of the CNN model;
calculating the error between the output result of the CNN model and the target output result to obtain a CNN prediction error;
judging whether the CNN prediction error is smaller than a first error threshold value or not to obtain a first judgment result;
if the first judgment result shows that the CNN prediction error is not smaller than a first error threshold, judging whether the CNN training frequency is smaller than a first training frequency threshold to obtain a second judgment result;
if the second judgment result indicates that the CNN training times are smaller than the first training time threshold, updating convolution kernel weights and additive offset vectors of all layers of the CNN model, and returning to the step of inputting normal time-frequency data in the normal time-frequency data set and fault time-frequency data in the fault time-frequency data set into a first convolution layer of the CNN model respectively to obtain an output result of the first convolution layer;
and if the first judgment result shows that the CNN prediction error is smaller than a first error threshold value or the second judgment result shows that the CNN training times are not smaller than a first training time threshold value, outputting the trained CNN model.
Optionally, the identifying the two-dimensional time-frequency data by using the LSTM network to obtain a second identification result specifically includes:
respectively acquiring a normal sound signal when the fan equipment normally operates and a fault sound signal when the fan equipment operates at the initial surge stage, and constructing a normal time-frequency data set and a fault time-frequency data set;
training an LSTM model according to the normal time-frequency data set and the fault time-frequency data set to obtain a trained LSTM model;
and inputting the two-dimensional time-frequency data into the trained LSTM model, and obtaining the output of the trained LSTM model as a second recognition result.
Optionally, training the LSTM model according to the normal time-frequency data set and the fault time-frequency data set to obtain a trained LSTM model, which specifically includes:
initializing weight values and additive bias vectors of each layer of the LSTM model;
respectively inputting the normal time-frequency data set and the fault time-frequency data set into an input layer of the LSTM model to obtain the output of the input layer;
inputting the output of the input layer into a first hidden layer of the LSTM model to obtain the output of the first hidden layer;
inputting the output of the first hidden layer into a second hidden layer of the LSTM model to obtain the output of the second hidden layer;
inputting the output of the second hidden layer into a full connection layer of the LSTM model to obtain an output result of the LSTM model;
calculating the error between the output result of the LSTM model and the target output result to obtain an LSTM prediction error;
judging whether the LSTM prediction error is smaller than a second error threshold value or not to obtain a third judgment result;
if the third judgment result shows that the LSTM prediction error is not smaller than the second error threshold, judging whether the LSTM training frequency is smaller than the second training frequency threshold or not to obtain a fourth judgment result;
if the fourth judgment result indicates that the LSTM training times are smaller than a second training time threshold, updating the weight and additive bias vectors of each layer of the LSTM model, and returning to the step of inputting the normal time-frequency data set and the fault time-frequency data set into the input layer of the LSTM model respectively to obtain the output of the input layer;
and if the third judgment result shows that the LSTM prediction error is smaller than a second error threshold value or the fourth judgment result shows that the LSTM training frequency is not smaller than a second training frequency threshold value, outputting the trained LSTM model.
A fan surge operation fault identification system, the identification system comprising:
the field device sound signal acquisition module is used for acquiring a field device sound signal of the fan device in the working process;
the preprocessing module is used for preprocessing the field device acoustic signal to obtain two-dimensional time-frequency data of the field device acoustic signal;
the CNN network identification module is used for identifying the two-dimensional time-frequency data by adopting a CNN network to obtain a first identification result;
the LSTM network identification module is used for identifying the two-dimensional time-frequency data by adopting an LSTM network to obtain a second identification result;
and the identification result fusion module is used for carrying out information fusion on the first identification result and the second identification result based on a D-S evidence theory to obtain an identification result of the surge operation fault of the fan.
Optionally, the preprocessing module specifically includes:
the filtering and Fourier transform submodule is used for carrying out Hamming window filtering and discrete Fourier transform on the field equipment acoustic signal to obtain frequency domain data of the field equipment acoustic signal;
and the two-dimensional time-frequency data establishing submodule is used for establishing the two-dimensional time-frequency data of the field equipment acoustic signal by taking the time domain data of the field equipment acoustic signal as one dimension and the frequency domain data as the other dimension.
Optionally, the CNN network identification module specifically includes:
the sample data acquisition submodule is used for respectively acquiring a normal sound signal when the fan equipment normally operates and a fault sound signal when the fan equipment operates at the initial surge stage, and constructing a normal time-frequency data set and a fault time-frequency data set;
the CNN model training submodule is used for training a CNN model according to the normal time-frequency data set and the fault time-frequency data set to obtain a trained CNN model;
and the CNN network identification submodule is used for inputting the two-dimensional time-frequency data into the trained CNN model and obtaining the output of the trained CNN model as a first identification result.
Optionally, the CNN model training sub-module specifically includes:
the initialization unit is used for initializing convolution kernel weights and additive offset vectors of each layer of the CNN model;
the first convolution layer training unit is used for respectively inputting the normal time-frequency data in the normal time-frequency data set and the fault time-frequency data in the fault time-frequency data set into a first convolution layer of a CNN (convolutional neural network) model and acquiring an output result of the first convolution layer;
the first pooling layer training unit is used for inputting the output result of the first convolution layer into a first pooling layer of the CNN model to obtain the output result of the first pooling layer;
the second convolutional layer training unit is used for inputting the output result of the first pooling layer into a second convolutional layer of the CNN model and acquiring the output result of the second convolutional layer;
the second pooling layer training unit is used for inputting the output result of the second convolutional layer into a second pooling layer of the CNN model and acquiring the output result of the second pooling layer;
the full-connection layer training unit is used for inputting the output result of the second pooling layer into the full-connection layer of the CNN model and acquiring the output result of the CNN model;
the prediction error calculation unit is used for calculating the error between the output result of the CNN model and the target output result to obtain a CNN prediction error;
the first judgment unit is used for judging whether the CNN prediction error is smaller than a first error threshold value or not to obtain a first judgment result;
a second judging unit, configured to, if the first judgment result indicates that the CNN prediction error is not smaller than the first error threshold, judge whether the CNN training frequency is smaller than the first training frequency threshold, to obtain a second judgment result;
a parameter updating unit, configured to update a convolution kernel weight and an additive offset vector of each layer of the CNN model if the second determination result indicates that the CNN training times are smaller than a first training time threshold, and return to the step "input the normal time-frequency data in the normal time-frequency data set and the fault time-frequency data in the fault time-frequency data set to a first convolution layer of the CNN model, respectively, to obtain an output result of the first convolution layer";
and the trained CNN model output unit is used for outputting the trained CNN model if the first judgment result shows that the CNN prediction error is smaller than a first error threshold or the second judgment result shows that the CNN training frequency is not smaller than a first training frequency threshold.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
the invention provides a method and a system for identifying surge operation faults of a fan. The identification method comprises the following steps: firstly, acquiring a field device acoustic signal of fan equipment in the working process; preprocessing the field device acoustic signal to obtain two-dimensional time-frequency data of the field device acoustic signal; then, identifying the two-dimensional time-frequency data by adopting a CNN network to obtain a first identification result; identifying the two-dimensional time frequency data by adopting an LSTM network to obtain a second identification result; and finally, based on a D-S evidence theory, performing information fusion on the first identification result and the second identification result to obtain an identification result of the surge operation fault of the fan. The method utilizes the field equipment acoustic signal to diagnose the fan surge fault, does not need to install a sensor on the fan, adopts a Convolutional Neural Network (CNN) and a Long Short-term Memory Network (LSTM) to diagnose the fault occurrence probability of the characteristics of the surge acoustic signal, and utilizes a D-S evidence theory to carry out fusion diagnosis of two results of a decision layer, thereby improving the diagnosis accuracy.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.
FIG. 1 is a flow chart of a method for identifying a surge operation fault of a wind turbine according to the present invention;
fig. 2 is a flowchart of CNN network training provided by the present invention;
FIG. 3 is a flow chart of LSTM network training provided by the present invention;
fig. 4 is a diagram illustrating a fast storage structure of the LSTM network according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention aims to provide a method and a system for identifying a surge operation fault of a fan, so as to improve the identification accuracy rate on the basis of not installing a sensor on fan equipment.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
As shown in fig. 1, the present invention provides a method for identifying a surge operation failure of a wind turbine, wherein the method comprises the following steps:
and acquiring the field equipment acoustic signal of the fan equipment in the working process.
Preprocessing the field device acoustic signal to obtain two-dimensional time-frequency data of the field device acoustic signal; the method specifically comprises the following steps:
carrying out Hamming window filtering and discrete Fourier transform on the field device acoustic signal to obtain frequency domain data of the field device acoustic signal; and establishing two-dimensional time-frequency data of the field device acoustic signal by taking the time domain data of the field device acoustic signal as one dimension and the frequency domain data as the other dimension.
Adopting a CNN network to identify the two-dimensional time-frequency data to obtain a first identification result; the method specifically comprises the following steps:
respectively acquiring a normal sound signal when the fan equipment normally operates and a fault sound signal when the fan equipment operates at the initial surge stage, and constructing a normal time-frequency data set and a fault time-frequency data set; training a CNN model according to the normal time-frequency data set and the fault time-frequency data set to obtain a trained CNN model; and inputting the two-dimensional time-frequency data into the trained CNN model, and obtaining the output of the trained CNN model as a first recognition result.
Identifying the two-dimensional time frequency data by adopting an LSTM network to obtain a second identification result; the method specifically comprises the following steps:
respectively acquiring a normal sound signal when the fan equipment normally operates and a fault sound signal when the fan equipment operates at the initial surge stage, and constructing a normal time-frequency data set and a fault time-frequency data set; training an LSTM model according to the normal time-frequency data set and the fault time-frequency data set to obtain a trained LSTM model; and inputting the two-dimensional time-frequency data into the trained LSTM model, and obtaining the output of the trained LSTM model as a second recognition result.
Based on a D-S evidence theory, performing information fusion on the first identification result and the second identification result to obtain an identification result of the surge operation fault of the fan; the fusion mode is shown as the following formula:
wherein, P (normal) is the fusion probability result of the normal operation of the fan, P (fault) is the fusion probability result of the surge operation of the fan, K is a normalization constant, PA(normal) is the CNN recognition probability, P, of the normal operation of the fanA(fault) CNN recognition probability, P, for a surging operation of a wind turbineB(normal) is the LSTM recognition probability, P, of the normal operation of the fanB(fault) is the LSTM recognition probability of a surge operation of a wind turbine.
The method comprises the following steps of respectively acquiring a normal sound signal when the fan equipment normally operates and a fault sound signal when the fan equipment surges and operates in an initial stage, constructing a normal time-frequency data set and a fault time-frequency data set, and specifically comprising the following steps of:
1-1 utilizes the adapter to gather the acoustic signal of fan equipment, obtains two kinds of digital sample signals of fixed duration: normal sound signals (hereinafter referred to as "normal signals") of normal operation of the fan and normal sound signals (hereinafter referred to as "fault signals") of the initial surge stage of the fan;
1-2, a plant status data set is established from a plurality of normal signals and fault signals, the fault signals and normal signals collectively referred to as sample signals in the plant status data set.
Step 2: obtaining two-dimensional time-frequency data
And carrying out short-time Fourier transform on each sample signal in the data set to obtain a plurality of frequency spectrums of each signal, and combining the frequency spectrums with the time domain spectrum to obtain two-dimensional time-frequency data serving as the input characteristics of the identification network. The method comprises the following specific steps:
2-1 performing initial N-point Hamming window filtering on a sample signal with the total length of M points in the equipment state data set, wherein N < M and a Hamming window function is as follows:
where n is each signal point within each frame of the audio signal.
2-2, performing discrete Fourier transform on N points in the window to obtain frequency domain data of the sample signal at the moment;
2-3, moving the window N points to the right, repeating the step 2-1 and the step 2-2, and if the signal length after moving is less than the N points, performing zero filling until all the sample signals of the M points are processed. Performing Hamming window filtering and Fourier transformation on the next moment of the sample signal until frequency domain data acquisition of the sample signal at all moments is completed;
2-4, for the sample signal, constructing two-dimensional time-frequency data as signal characteristics by taking time domain data as one dimension and corresponding frequency domain data as the other dimension;
2-5 repeating the steps 2-1 to 2-4 to obtain a normal time frequency data set of the normal signals and a fault time frequency signal data set of the fault signals in the data set.
As shown in fig. 2, the training a CNN model according to the normal time-frequency data set and the fault time-frequency data set to obtain a trained CNN model specifically includes:
initializing convolution kernel weights and additive offset vectors of each layer of the CNN model; specifically, the convolution kernel weight is initialized with a random value, and the additive bias vector is initialized with a zero value. Respectively inputting the normal time-frequency data in the normal time-frequency data set and the fault time-frequency data in the fault time-frequency data set into a first convolution layer of a CNN (convolutional neural network) model to obtain an output result of the first convolution layer; inputting the output result of the first convolution layer into a first pooling layer of the CNN model to obtain the output result of the first pooling layer; inputting the output result of the first pooling layer into a second convolution layer of the CNN model to obtain the output result of the second convolution layer; inputting the output result of the second convolutional layer into a second pooling layer of the CNN model to obtain the output result of the second pooling layer; inputting the output result of the second pooling layer into the full-connection layer of the CNN model to obtain the output result of the CNN model; calculating the error of the output result of the CNN model and the target output result (the target output result is the actual fault category of the training sample, for example, A, B, C fault categories are available, if the actual fault category is A, the target output result is 1, 0, 0, namely the probability that the fault category is A is 1, and the probability that the fault category is B and C is 0; and the output result of the CNN model is the probability that the state of the training sample belongs to each fault type), so as to obtain the CNN prediction error; judging whether the CNN prediction error is smaller than a first error threshold value or not to obtain a first judgment result; if the first judgment result shows that the CNN prediction error is not smaller than a first error threshold, judging whether the CNN training frequency is smaller than a first training frequency threshold to obtain a second judgment result; if the second judgment result indicates that the CNN training times are smaller than the first training time threshold, updating convolution kernel weights and additive offset vectors of all layers of the CNN model, and returning to the step of inputting normal time-frequency data in the normal time-frequency data set and fault time-frequency data in the fault time-frequency data set into a first convolution layer of the CNN model respectively to obtain an output result of the first convolution layer; and if the first judgment result shows that the CNN prediction error is smaller than a first error threshold value or the second judgment result shows that the CNN training times are not smaller than a first training time threshold value, outputting the trained CNN model.
The method comprises the following specific steps:
3-1, initializing each parameter of the convolution network, initializing the convolution kernel weight by adopting a random value, and initializing the additive offset vector by adopting a zero value;
3-2 inputting a normal time-frequency data set and a fault time-frequency data set to carry out the characteristic training of the convolutional layer, wherein the formula is as follows:
Xl＝A(Xl-1×Wl+bl)
wherein XlFor convolutional layer output, A is the activation function, Xl-1For local areas of input data, WlAs a convolution kernel weight, blIs an additive offset vector;
after the convolution layer of the whole two-dimensional time frequency data is output, the output result is input into a pooling layer for down sampling, and the operation of the pooling layer is shown as a formula: xl＝S(Xl-1) Wherein X islFor pooled layer output, S is the downsampling rule for maximum pooling, Xl-1The local region of the input data is processed until the pooling operation of the convolution output graph data is completed.
3-3, repeating the step 3-2 to finish the second convolution and pooling operation;
3-4, inputting the result of the step 3-3 into a full connection layer, and obtaining an output value y through a ReLU activation function and a Softmax functiondI.e. the output of the CNN model;
3-5 determining the error between the output value and the target value y (target output result) by using Euclidean distance
3-6, judging whether the error is smaller than an expected value (a first error threshold), if so, returning the error to the network, sequentially obtaining the error of each layer, updating the weight and the additive offset vector of each layer of the CNN network, and repeating the steps 3-2 to 3-6 to train the network for the next time; and if the error is smaller than the expected value or the CNN network training times reach a first training time threshold, finishing the network training to obtain the trained CNN network. The invention updates the convolution kernel weight and the additive offset vector of each layer of the CNN model in the following ways: and (4) reversely transmitting the error along each layer of the CNN, calculating the error of parameters such as convolution kernel weight values and additive offset vectors of each layer, obtaining the modified value of each parameter through gradient descent, and obtaining the updated parameter by subtracting the modified value from the original value.
As shown in fig. 3, the training of the LSTM model according to the normal time-frequency data set and the fault time-frequency data set to obtain the trained LSTM model specifically includes:
initializing weight values and additive bias vectors of each layer of the LSTM model; respectively inputting the normal time-frequency data set and the fault time-frequency data set into an input layer of the LSTM model to obtain the output of the input layer; inputting the output of the input layer into a first hidden layer of the LSTM model to obtain the output of the first hidden layer; inputting the output of the first hidden layer into a second hidden layer of the LSTM model to obtain the output of the second hidden layer; inputting the output of the second hidden layer into a full connection layer of the LSTM model to obtain an output result of the LSTM model; calculating the error between the output result of the LSTM model and the target output result to obtain an LSTM prediction error; judging whether the LSTM prediction error is smaller than a second error threshold value or not to obtain a third judgment result; if the third judgment result shows that the LSTM prediction error is not smaller than the second error threshold, judging whether the LSTM training frequency is smaller than the second training frequency threshold or not to obtain a fourth judgment result; if the fourth judgment result indicates that the LSTM training times are smaller than a second training time threshold, updating the weight and additive bias vectors of each layer of the LSTM model, and returning to the step of inputting the normal time-frequency data set and the fault time-frequency data set into the input layer of the LSTM model respectively to obtain the output of the input layer; and if the third judgment result shows that the LSTM prediction error is smaller than a second error threshold value or the fourth judgment result shows that the LSTM training frequency is not smaller than a second training frequency threshold value, outputting the trained LSTM model. The method comprises the following specific steps:
4-1, initializing each parameter weight and bias of the LSTM network;
4-2, inputting the normal time-frequency data set and the fault time-frequency data set to an input layer of the LSTM network, wherein all frequency domain data of a certain time-frequency data in the same time domain dimension represent an input node, and the time-frequency data can be transmitted to the next layer as a characteristic;
4-3, processing the time-frequency data characteristics in two hidden layers. Each hidden layer has a plurality of time step processing modules, each time step module has a time sequence computing capability, the time step modules corresponding to the two hidden layers have a layer sequence computing capability, as shown in fig. 4, each time step module has a memory block structure including an input gate, a forgetting gate and an output gate, and the computation relationship of each time step is as follows:
wherein the content of the first and second substances,
for the purpose of inputting the output of the gate,
in order to forget the output of the gate,
is the output of the output gate, and is,
the cell parameters are memorized for the current time step,
the cell parameters are stored for the last time step,
the local feature input region corresponding to the time step for the previous hidden layer,
in a local feature input area of a time step on the same hidden layer, sigm represents a nonlinear Sigmoid activation function, Tanh represents a Tanh activation function,
to represent
And
tan h activation output of, WlWeight matrix representing the current hidden layer, blAn additive vector bias representing the current hidden layer,
is the output of the current time step.
Transmitting the output to the next layer until the hidden layer calculation is completed;
4-4, inputting the output result of the step 4-3 into the full connection layer, and obtaining an output value y 'through a ReLU activation function and a Softmax function'dNamely the output result of the LSTM network;
4-5 calculating the error between the output value and the target value y' (target output result) by using Euclidean distance
4-6, judging whether the error is smaller than a desired value (a second error threshold), if so, calculating the error item value of each neuron in a reverse direction, wherein error transmission needs to be carried out in two directions, and one is in reverse propagation along time, namely, calculating the error item of each previous moment from the current moment t; the other one is to spread the error term to a previous hidden layer, calculate the parameter gradient and update the parameter value according to the corresponding error term, repeat the steps 4-2 to 4-6, carry on the next network training; and if the error is smaller than the expected value or the network training times reach the maximum value, finishing the network training to obtain the trained LSTM network. The specific steps of updating the weight and the additive bias vector of each layer of the LSTM model comprise: the specific error transfer mode is consistent with the error transfer mode in the convolution kernel weight of each layer of the CNN model and the updating mode of the additive bias vector, and particularly, the LSTM network carries out error transfer along two directions, one is back propagation along time, namely, from the current t moment, the error item of each moment before is calculated; the other is to propagate the error term to an upper hidden layer, calculate the parameter gradient and update the parameter value according to the corresponding error term.
The invention also provides a fan surge operation fault identification system, which comprises the following components:
the field device sound signal acquisition module is used for acquiring a field device sound signal of the fan device in the working process; the method specifically comprises the following steps:
the filtering and Fourier transform submodule is used for carrying out Hamming window filtering and discrete Fourier transform on the field equipment acoustic signal to obtain frequency domain data of the field equipment acoustic signal; and the two-dimensional time-frequency data establishing submodule is used for establishing the two-dimensional time-frequency data of the field equipment acoustic signal by taking the time domain data of the field equipment acoustic signal as one dimension and the frequency domain data as the other dimension.
The preprocessing module is used for preprocessing the field device acoustic signal to obtain two-dimensional time-frequency data of the field device acoustic signal;
the CNN network identification module is used for identifying the two-dimensional time-frequency data by adopting a CNN network to obtain a first identification result; the method specifically comprises the following steps:
the sample data acquisition submodule is used for respectively acquiring a normal sound signal when the fan equipment normally operates and a fault sound signal when the fan equipment operates at the initial surge stage, and constructing a normal time-frequency data set and a fault time-frequency data set; the CNN model training submodule is used for training a CNN model according to the normal time-frequency data set and the fault time-frequency data set to obtain a trained CNN model; and the CNN network identification submodule is used for inputting the two-dimensional time-frequency data into the trained CNN model and obtaining the output of the trained CNN model as a first identification result.
The CNN model training submodule specifically includes: the initialization unit is used for initializing convolution kernel weights and additive offset vectors of each layer of the CNN model; the first convolution layer training unit is used for respectively inputting the normal time-frequency data in the normal time-frequency data set and the fault time-frequency data in the fault time-frequency data set into a first convolution layer of a CNN (convolutional neural network) model and acquiring an output result of the first convolution layer; the first pooling layer training unit is used for inputting the output result of the first convolution layer into a first pooling layer of the CNN model to obtain the output result of the first pooling layer; the second convolutional layer training unit is used for inputting the output result of the first pooling layer into a second convolutional layer of the CNN model and acquiring the output result of the second convolutional layer; the second pooling layer training unit is used for inputting the output result of the second convolutional layer into a second pooling layer of the CNN model and acquiring the output result of the second pooling layer; the full-connection layer training unit is used for inputting the output result of the second pooling layer into the full-connection layer of the CNN model and acquiring the output result of the CNN model; the prediction error calculation unit is used for calculating the error between the output result of the CNN model and the target output result to obtain a CNN prediction error; the first judgment unit is used for judging whether the CNN prediction error is smaller than a first error threshold value or not to obtain a first judgment result; a second judging unit, configured to, if the first judgment result indicates that the CNN prediction error is not smaller than the first error threshold, judge whether the CNN training frequency is smaller than the first training frequency threshold, to obtain a second judgment result; a parameter updating unit, configured to update a convolution kernel weight and an additive offset vector of each layer of the CNN model if the second determination result indicates that the CNN training times are smaller than a first training time threshold, and return to the step "input the normal time-frequency data in the normal time-frequency data set and the fault time-frequency data in the fault time-frequency data set to a first convolution layer of the CNN model, respectively, to obtain an output result of the first convolution layer"; and the trained CNN model output unit is used for outputting the trained CNN model if the first judgment result shows that the CNN prediction error is smaller than a first error threshold or the second judgment result shows that the CNN training frequency is not smaller than a first training frequency threshold.
The LSTM network identification module is used for identifying the two-dimensional time-frequency data by adopting an LSTM network to obtain a second identification result;
and the identification result fusion module is used for carrying out information fusion on the first identification result and the second identification result based on a D-S evidence theory to obtain an identification result of the surge operation fault of the fan.
The method has the advantages that the cost and inconvenience brought by the installation and the maintenance of multiple sensors are reduced by combining an acoustic processing technology, and the algorithm of single information judgment is easier to realize. The traditional audio information processing technology and the deep learning method are combined, and the fault recognition rate is greatly improved. Decision fusion is carried out on the recognition results of the CNN network and the LSTM network through a D-S evidence theory, and fault recognition is carried out from the static judgment of the CNN network and the dynamic memory of the LSTM network, so that the accuracy of information judgment is improved, and the fault early warning and prediction capability is enhanced.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. For the system disclosed by the embodiment, the description is relatively simple because the system corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the method part for description.
The principle and the implementation manner of the present invention are explained by applying specific examples, the above description of the embodiments is only used to help understanding the method of the present invention and the core idea thereof, the described embodiments are only a part of the embodiments of the present invention, not all embodiments, and all other embodiments obtained by a person of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the protection scope of the present invention.
Claims (10)
1. A method for identifying a surge operation fault of a fan is characterized by comprising the following steps:
acquiring a field device acoustic signal of the fan device in the working process;
preprocessing the field device acoustic signal to obtain two-dimensional time-frequency data of the field device acoustic signal;
adopting a CNN network to identify the two-dimensional time-frequency data to obtain a first identification result;
identifying the two-dimensional time frequency data by adopting an LSTM network to obtain a second identification result;
and based on a D-S evidence theory, performing information fusion on the first identification result and the second identification result to obtain an identification result of the surge operation fault of the fan.
2. The method for identifying the surge operation fault of the wind turbine according to claim 1, wherein the preprocessing is performed on the field device acoustic signal to obtain two-dimensional time-frequency data of the field device acoustic signal, and specifically comprises:
carrying out Hamming window filtering and discrete Fourier transform on the field device acoustic signal to obtain frequency domain data of the field device acoustic signal;
and establishing two-dimensional time-frequency data of the field device acoustic signal by taking the time domain data of the field device acoustic signal as one dimension and the frequency domain data as the other dimension.
3. The method for identifying the surge operation fault of the fan according to claim 1, wherein the identifying the two-dimensional time-frequency data by using the CNN network to obtain a first identification result specifically comprises:
respectively acquiring a normal sound signal when the fan equipment normally operates and a fault sound signal when the fan equipment operates at the initial surge stage, and constructing a normal time-frequency data set and a fault time-frequency data set;
training a CNN model according to the normal time-frequency data set and the fault time-frequency data set to obtain a trained CNN model;
and inputting the two-dimensional time-frequency data into the trained CNN model, and obtaining the output of the trained CNN model as a first recognition result.
4. The method for identifying the surge operation fault of the fan according to claim 3, wherein the training of the CNN model according to the normal time-frequency data set and the fault time-frequency data set to obtain the trained CNN model specifically comprises:
initializing convolution kernel weights and additive offset vectors of each layer of the CNN model;
respectively inputting the normal time-frequency data in the normal time-frequency data set and the fault time-frequency data in the fault time-frequency data set into a first convolution layer of a CNN (convolutional neural network) model to obtain an output result of the first convolution layer;
inputting the output result of the first convolution layer into a first pooling layer of the CNN model to obtain the output result of the first pooling layer;
inputting the output result of the first pooling layer into a second convolution layer of the CNN model to obtain the output result of the second convolution layer;
inputting the output result of the second convolutional layer into a second pooling layer of the CNN model to obtain the output result of the second pooling layer;
inputting the output result of the second pooling layer into the full-connection layer of the CNN model to obtain the output result of the CNN model;
calculating the error between the output result of the CNN model and the target output result to obtain a CNN prediction error;
judging whether the CNN prediction error is smaller than a first error threshold value or not to obtain a first judgment result;
if the first judgment result shows that the CNN prediction error is not smaller than a first error threshold, judging whether the CNN training frequency is smaller than a first training frequency threshold to obtain a second judgment result;
if the second judgment result indicates that the CNN training times are smaller than the first training time threshold, updating convolution kernel weights and additive offset vectors of all layers of the CNN model, and returning to the step of inputting normal time-frequency data in the normal time-frequency data set and fault time-frequency data in the fault time-frequency data set into a first convolution layer of the CNN model respectively to obtain an output result of the first convolution layer;
and if the first judgment result shows that the CNN prediction error is smaller than a first error threshold value or the second judgment result shows that the CNN training times are not smaller than a first training time threshold value, outputting the trained CNN model.
5. The method for identifying the surge operation fault of the fan according to claim 1, wherein the identifying the two-dimensional time-frequency data by using the LSTM network to obtain a second identification result specifically comprises:
respectively acquiring a normal sound signal when the fan equipment normally operates and a fault sound signal when the fan equipment operates at the initial surge stage, and constructing a normal time-frequency data set and a fault time-frequency data set;
training an LSTM model according to the normal time-frequency data set and the fault time-frequency data set to obtain a trained LSTM model;
and inputting the two-dimensional time-frequency data into the trained LSTM model, and obtaining the output of the trained LSTM model as a second recognition result.
6. The method for identifying the surge operation fault of the fan according to claim 5, wherein the LSTM model is trained according to the normal time-frequency data set and the fault time-frequency data set to obtain the trained LSTM model, and the method specifically comprises the following steps:
initializing weight values and additive bias vectors of each layer of the LSTM model;
respectively inputting the normal time-frequency data set and the fault time-frequency data set into an input layer of the LSTM model to obtain the output of the input layer;
inputting the output of the input layer into a first hidden layer of the LSTM model to obtain the output of the first hidden layer;
inputting the output of the first hidden layer into a second hidden layer of the LSTM model to obtain the output of the second hidden layer;
inputting the output of the second hidden layer into a full connection layer of the LSTM model to obtain an output result of the LSTM model;
calculating the error between the output result of the LSTM model and the target output result to obtain an LSTM prediction error;
judging whether the LSTM prediction error is smaller than a second error threshold value or not to obtain a third judgment result;
if the third judgment result shows that the LSTM prediction error is not smaller than the second error threshold, judging whether the LSTM training frequency is smaller than the second training frequency threshold or not to obtain a fourth judgment result;
if the fourth judgment result indicates that the LSTM training times are smaller than a second training time threshold, updating the weight and additive bias vectors of each layer of the LSTM model, and returning to the step of inputting the normal time-frequency data set and the fault time-frequency data set into the input layer of the LSTM model respectively to obtain the output of the input layer;
and if the third judgment result shows that the LSTM prediction error is smaller than a second error threshold value or the fourth judgment result shows that the LSTM training frequency is not smaller than a second training frequency threshold value, outputting the trained LSTM model.
7. A fan surge operation fault identification system, the identification system comprising:
the field device sound signal acquisition module is used for acquiring a field device sound signal of the fan device in the working process;
the preprocessing module is used for preprocessing the field device acoustic signal to obtain two-dimensional time-frequency data of the field device acoustic signal;
the CNN network identification module is used for identifying the two-dimensional time-frequency data by adopting a CNN network to obtain a first identification result;
the LSTM network identification module is used for identifying the two-dimensional time-frequency data by adopting an LSTM network to obtain a second identification result;
and the identification result fusion module is used for carrying out information fusion on the first identification result and the second identification result based on a D-S evidence theory to obtain an identification result of the surge operation fault of the fan.
8. The blower surge operation fault identification system of claim 7, wherein the preprocessing module specifically comprises:
the filtering and Fourier transform submodule is used for carrying out Hamming window filtering and discrete Fourier transform on the field equipment acoustic signal to obtain frequency domain data of the field equipment acoustic signal;
and the two-dimensional time-frequency data establishing submodule is used for establishing the two-dimensional time-frequency data of the field equipment acoustic signal by taking the time domain data of the field equipment acoustic signal as one dimension and the frequency domain data as the other dimension.
9. The wind turbine surge operation fault identification system of claim 7, wherein the CNN network identification module specifically comprises:
the sample data acquisition submodule is used for respectively acquiring a normal sound signal when the fan equipment normally operates and a fault sound signal when the fan equipment operates at the initial surge stage, and constructing a normal time-frequency data set and a fault time-frequency data set;
the CNN model training submodule is used for training a CNN model according to the normal time-frequency data set and the fault time-frequency data set to obtain a trained CNN model;
and the CNN network identification submodule is used for inputting the two-dimensional time-frequency data into the trained CNN model and obtaining the output of the trained CNN model as a first identification result.
10. The system for identifying a wind turbine surge operation fault according to claim 9, wherein the CNN model training submodule specifically comprises:
the initialization unit is used for initializing convolution kernel weights and additive offset vectors of each layer of the CNN model;
the first convolution layer training unit is used for respectively inputting the normal time-frequency data in the normal time-frequency data set and the fault time-frequency data in the fault time-frequency data set into a first convolution layer of a CNN (convolutional neural network) model and acquiring an output result of the first convolution layer;
the first pooling layer training unit is used for inputting the output result of the first convolution layer into a first pooling layer of the CNN model to obtain the output result of the first pooling layer;
the second convolutional layer training unit is used for inputting the output result of the first pooling layer into a second convolutional layer of the CNN model and acquiring the output result of the second convolutional layer;
the second pooling layer training unit is used for inputting the output result of the second convolutional layer into a second pooling layer of the CNN model and acquiring the output result of the second pooling layer;
the full-connection layer training unit is used for inputting the output result of the second pooling layer into the full-connection layer of the CNN model and acquiring the output result of the CNN model;
the prediction error calculation unit is used for calculating the error between the output result of the CNN model and the target output result to obtain a CNN prediction error;
the first judgment unit is used for judging whether the CNN prediction error is smaller than a first error threshold value or not to obtain a first judgment result;
a second judging unit, configured to, if the first judgment result indicates that the CNN prediction error is not smaller than the first error threshold, judge whether the CNN training frequency is smaller than the first training frequency threshold, to obtain a second judgment result;
a parameter updating unit, configured to update a convolution kernel weight and an additive offset vector of each layer of the CNN model if the second determination result indicates that the CNN training times are smaller than a first training time threshold, and return to the step "input the normal time-frequency data in the normal time-frequency data set and the fault time-frequency data in the fault time-frequency data set to a first convolution layer of the CNN model, respectively, to obtain an output result of the first convolution layer";
and the trained CNN model output unit is used for outputting the trained CNN model if the first judgment result shows that the CNN prediction error is smaller than a first error threshold or the second judgment result shows that the CNN training frequency is not smaller than a first training frequency threshold.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911021826.5A CN112052551B (en) | 2019-10-25 | 2019-10-25 | Fan surge operation fault identification method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911021826.5A CN112052551B (en) | 2019-10-25 | 2019-10-25 | Fan surge operation fault identification method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112052551A true CN112052551A (en) | 2020-12-08 |
| CN112052551B CN112052551B (en) | 2023-05-02 |
Family
ID=73609742
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911021826.5A Active CN112052551B (en) | 2019-10-25 | 2019-10-25 | Fan surge operation fault identification method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112052551B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113566948A (en) * | 2021-07-09 | 2021-10-29 | 中煤科工集团沈阳研究院有限公司 | Fault audio recognition and diagnosis method for robot coal pulverizer |
| CN115306754A (en) * | 2022-10-12 | 2022-11-08 | 中国航发四川燃气涡轮研究院 | Axial flow fan aerodynamic instability identification method based on acoustic array |
| CN116393217A (en) * | 2023-02-24 | 2023-07-07 | 华能曲阜热电有限公司 | Intelligent monitoring method for material level of steel ball coal mill |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170010308A1 (en) * | 2015-07-08 | 2017-01-12 | Sensanna Incorporated | Low loss acoustic wave sensors and tags and high efficiency antennas and methods for remote activation thereof |
| WO2019056471A1 (en) * | 2017-09-19 | 2019-03-28 | 平安科技（深圳）有限公司 | Driving model training method, driver recognition method and apparatus, device, and medium |
| CN110131193A (en) * | 2018-02-02 | 2019-08-16 | 中国航发商用航空发动机有限责任公司 | Aero-engine surge fault monitoring method and system |
-
2019
- 2019-10-25 CN CN201911021826.5A patent/CN112052551B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170010308A1 (en) * | 2015-07-08 | 2017-01-12 | Sensanna Incorporated | Low loss acoustic wave sensors and tags and high efficiency antennas and methods for remote activation thereof |
| WO2019056471A1 (en) * | 2017-09-19 | 2019-03-28 | 平安科技（深圳）有限公司 | Driving model training method, driver recognition method and apparatus, device, and medium |
| CN110131193A (en) * | 2018-02-02 | 2019-08-16 | 中国航发商用航空发动机有限责任公司 | Aero-engine surge fault monitoring method and system |
Non-Patent Citations (6)
| Title |
|---|
| "Time amplifying techniques towards atomic time resolution" * |
| RUOCHEN DUAN 等: "Fault Diagnosis of On-Load Tap-Changer in Converter Transformer Based on Time–Frequency Vibration Analysis" * |
| YONGJIE ZHAI 等: "Multiscale Entropy Feature Extraction Method of Running Power Equipment Sound" * |
| 李粮;田青林;钟明;赖小皇;: "基于Mallat算法的机载失速喘振辨识装置研制" * |
| 李长征;胡智琦;许思琦;: "基于D-S证据融合的压气机气动失稳检测" * |
| 陈敏红: "基于时频分析的轴流风机故障诊断实验研究" * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113566948A (en) * | 2021-07-09 | 2021-10-29 | 中煤科工集团沈阳研究院有限公司 | Fault audio recognition and diagnosis method for robot coal pulverizer |
| CN115306754A (en) * | 2022-10-12 | 2022-11-08 | 中国航发四川燃气涡轮研究院 | Axial flow fan aerodynamic instability identification method based on acoustic array |
| CN115306754B (en) * | 2022-10-12 | 2023-02-17 | 中国航发四川燃气涡轮研究院 | Axial flow fan aerodynamic instability identification method based on acoustic array |
| CN116393217A (en) * | 2023-02-24 | 2023-07-07 | 华能曲阜热电有限公司 | Intelligent monitoring method for material level of steel ball coal mill |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112052551B (en) | 2023-05-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107941537B (en) | A kind of mechanical equipment health state evaluation method | |
| Zhang et al. | Bearing performance degradation assessment using long short-term memory recurrent network | |
| CN109308522B (en) | GIS fault prediction method based on recurrent neural network | |
| CN112052551A (en) | Method and system for identifying surge operation fault of fan | |
| CN107725283A (en) | A kind of fan trouble detection method based on depth belief network model | |
| CN110866448A (en) | Flutter signal analysis method based on convolutional neural network and short-time Fourier transform | |
| CN110309537B (en) | Intelligent health prediction method and system for aircraft | |
| CN113659565B (en) | Online prediction method for frequency situation of new energy power system | |
| CN113567131A (en) | Bearing fault diagnosis method based on S transformation and miniature convolution neural network model | |
| Pan et al. | High-precision bearing signal recovery based on signal fusion and variable stepsize forward-backward pursuit | |
| CN110737948A (en) | method for predicting residual life of aero-engine based on deep FNN-LSTM hybrid network | |
| CN116677569A (en) | Fan tower stress monitoring and early warning method and system | |
| CN105894014B (en) | Abnormal behavior sequential detection method based on multi-factor inconsistency measurement | |
| CN105823801B (en) | A kind of electronic nose drift compensation method based on deepness belief network feature extraction | |
| CN114372492A (en) | Interpretable rolling bearing fault diagnosis method | |
| CN114266201A (en) | Self-attention elevator trapping prediction method based on deep learning | |
| CN112802011A (en) | Fan blade defect detection method based on VGG-BLS | |
| CN113627268A (en) | Model training method, and method and device for detecting fault of speed reducer for mine hoist | |
| US20220269988A1 (en) | Abnormality degree calculation system and abnormality degree calculation method | |
| Pérez-Pérez et al. | Fault detection and isolation in wind turbines based on neuro-fuzzy qLPV zonotopic observers | |
| CN110287594B (en) | Aero-engine state diagnosis method based on neural network algorithm | |
| CN112560252A (en) | Prediction method for residual life of aircraft engine | |
| CN108710920B (en) | Indicator diagram identification method and device | |
| Jung et al. | A Diagnostic Technology of Powertrain Parts that Cause Abnormal Noises Using Artificial Intelligence | |
| CN111048203B (en) | Brain blood flow regulator evaluation device |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |