CN114166318A - Ultrasonic water meter flow data calibration method based on deep learning - Google Patents
Ultrasonic water meter flow data calibration method based on deep learning Download PDFInfo
- Publication number
- CN114166318A CN114166318A CN202210131336.6A CN202210131336A CN114166318A CN 114166318 A CN114166318 A CN 114166318A CN 202210131336 A CN202210131336 A CN 202210131336A CN 114166318 A CN114166318 A CN 114166318A
- Authority
- CN
- China
- Prior art keywords
- data
- flow data
- flow
- data calibration
- water meter
- 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
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01F—MEASURING VOLUME, VOLUME FLOW, MASS FLOW OR LIQUID LEVEL; METERING BY VOLUME
- G01F1/00—Measuring the volume flow or mass flow of fluid or fluent solid material wherein the fluid passes through a meter in a continuous flow
- G01F1/66—Measuring the volume flow or mass flow of fluid or fluent solid material wherein the fluid passes through a meter in a continuous flow by measuring frequency, phase shift or propagation time of electromagnetic or other waves, e.g. using ultrasonic flowmeters
-
- 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
-
- 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
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Electromagnetism (AREA)
- Fluid Mechanics (AREA)
- Measuring Volume Flow (AREA)
Abstract
The invention relates to the technical field of intelligent water affairs, and discloses an ultrasonic water meter flow data calibration method based on deep learning, which comprises the following steps: processing raw data of the water meter by using a metering algorithm to obtain the ultrasonic flight time difference, the absolute flight time of the ultrasonic wave on the upstream and downstream of the transducer and the upstream and downstream signal amplitude of the transducer in a specified measuring time period under different conditions; obtaining the real instantaneous flow of the water flow in the period of time by a standard table method; repeating the previous two steps until N groups of data are obtained and preprocessing the data; creating a flow data calibration model; training and testing the flow data calibration model by using the preprocessed data; and acquiring the ultrasonic flight time difference, the water temperature and the maximum amplitude difference of the upstream and the downstream of the transducer of the flow to be calibrated, and inputting the ultrasonic flight time difference, the water temperature and the maximum amplitude difference into a flow data calibration model to calibrate the flow data. The invention does not need factory calibration, greatly improves the calibration efficiency, realizes data calibration under the conditions of multiple temperatures and complex flow fields, and has strong engineering practicability.
Description
Technical Field
The invention relates to the technical field of intelligent water affairs, in particular to an ultrasonic water meter flow data calibration method based on deep learning, which is mainly used for improving the accuracy of flow data measured by an ultrasonic water meter in an intelligent water affair system.
Background
The traditional water meter flow data calibration method is very easily influenced by flow fields, temperatures and individual differences of water meters, is low in accuracy, and because each water meter has errors in manufacturing, processing and assembling, each water meter needs to be independently subjected to factory calibration, and time is extremely consumed.
With the improvement of computing power, the convolutional neural network capable of generating more accurate results is widely applied to the field of data processing, and can be well applied to the technical field of water meter flow data calibration.
Disclosure of Invention
Aiming at the defects and shortcomings of the prior art, the invention provides the ultrasonic water meter flow data calibration method based on deep learning, which does not need factory calibration, greatly improves the calibration efficiency, realizes the flow data calibration under the conditions of multiple temperatures and complex flow fields, and has strong engineering practicability.
The purpose of the invention can be realized by the following technical scheme:
a deep learning-based ultrasonic water meter flow data calibration method comprises the following steps:
step S1: processing the original time sequence data of the water meter by using a metering algorithm to obtain the ultrasonic flight time difference, water temperature data and upstream and downstream signal amplitudes of the transducer in different batches, different epitopes and different temperatures within a specified measuring time period;
step S2: obtaining the real instantaneous flow of the water flow in the period of time by a standard table method;
step S3: repeating the steps S1 and S2 until N groups of data are obtained, and preprocessing the data;
step S4: creating a flow data calibration model based on a one-dimensional convolutional neural network;
step S5: training and testing the flow data calibration model by using the preprocessed data;
step S6: and acquiring the ultrasonic flight time difference of the flow to be calibrated, the water temperature and the difference between the maximum amplitudes of the upstream and the downstream of the transducer, and inputting the difference into a flow data calibration model to calibrate the flow data.
Further, the data preprocessing step in step S3 specifically includes:
step S31: calculating the difference between the maximum amplitudes of the upstream and the downstream of the transducer in each measurement time period;
step S32: cleaning the flight time difference data and the water temperature data of the ultrasonic water meter in each measurement time period and the difference data of the maximum amplitude values of the upstream and the downstream of the transducer, and eliminating abnormal points;
step S33: interpolation supplement is carried out on the removed data abnormal points to ensure that the data lengths are consistent;
step S34: and carrying out normalization processing on the interpolated data.
Further, in step S5, 70% of the preprocessed data is used as the training set and 30% is used as the testing set.
Further, the flow data calibration model in step S4 adopts the Relu function as the activation function.
Further, the flow data calibration model in step S4 is output probabilistically by using a normalized exponential function.
Further, the step S5 specifically includes:
step S51: establishing a flow data calibration model based on a one-dimensional convolutional neural network, and setting training parameters and a time threshold of the flow data calibration model; the training parameters at least comprise a period, a batch size, iteration times and a learning rate;
step S52: training the flow data calibration model by using training set data;
step S53: judging whether the identification accuracy meets an early stop method every time the time threshold value is passed, if so, entering step S54; otherwise, go to step S52;
step S54: and sending the test set data into a flow data calibration model for testing, verifying whether the accuracy requirement is met, if so, entering a step S6, and otherwise, skipping to S52.
Further, the flow data calibration model in step S5 optimizes the training parameters by using an adaptive matrix estimation algorithm during the training process.
Further, the step S6 is specifically: and after normalizing the obtained ultrasonic flight time difference of the flow to be calibrated, the water temperature and the difference between the maximum amplitudes of the upstream and the downstream of the transducer, inputting a flow data calibration model for data calibration.
Further, N in step S3 is an integer of 200 or more.
The invention has the beneficial technical effects that:
by collecting different temperatures, different flow field conditions, time sequence data of different water meters and real instantaneous flow corresponding to calibration data, the trained flow data calibration model can calibrate the water meter flow data under different conditions without factory calibration, and the calibration efficiency is greatly improved.
The flow data calibration model is created through the convolutional neural network, the created model is the one-dimensional convolutional neural network, the input data dimension is N x 1, the convolutional kernel dimension is M x 1, M is smaller than N, the structure is relatively simple, the performance of the flow data calibration model training is further improved, and the water meter flow data calibration speed is greatly improved.
The data are normalized and then input into a flow data calibration model, the temperature, the time-of-flight difference and the maximum amplitude of the upstream and the downstream of the transducer are subjected to descaler dimensionization and pulled into the same frame for training, the convergence speed of the model is improved, a Dropout layer arranged in the flow data calibration model randomly assigns zero weight to neurons in the network, 50% of the neurons are zero weight due to the fact that the scheme selects the ratio of 0.5, the generalization capability of the flow data calibration model is greatly improved through the operation, and the accuracy of processing invisible data is further improved.
And judging whether to stop the training of the numerical recognition model or not by a premature stop method, namely stopping the training when the recognition accuracy begins to decrease and using the training parameter of the last iteration as the final parameter of the numerical recognition model, so as to avoid overfitting caused by continuous training.
Drawings
FIG. 1 is a general flow diagram of the present invention.
Fig. 2 is a diagram of a flow data calibration model according to an embodiment of the invention.
Fig. 3 is a data diagram of an ultrasonic water meter in an embodiment of the present invention.
Reference numerals: 1 represents the ultrasonic water meter time-of-flight difference; 2 represents water temperature; 3 represents the difference between the maximum amplitudes upstream and downstream of the transducer; and 4 represents the true instantaneous flow rate of the water flow.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is 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 do not limit the invention.
As shown in fig. 1, a deep learning-based ultrasonic water meter flow data calibration method includes the following steps:
step S1: three ultrasonic water meters are selected from different batches to be tested, the epitopes of the three ultrasonic water meters are different, the water temperature is selected from 0-50 ℃, and a group of time sequence data is collected at every 5 ℃. The sampling frequency is set to be 2Hz, the original time sequence data of the water meter is processed by using a metering algorithm, and the ultrasonic wave flight time difference (each group of data comprises 120 data points), the water temperature data (each group of data comprises 120 data points) and the upstream and downstream signal amplitude of the transducer (each group of data comprises 120 data points) in different batches, different epitopes and different temperatures in a specified measurement time period are obtained.
Step S2: obtaining the real instantaneous flow of the water flow in the period of time by a standard table method; the standard meter method is that a standard meter is connected in series on a pipeline to acquire real flow in real time.
Step S3: the obtained set of data is shown in fig. 3, and steps S1 and S2 are repeated until 200 sets of data are obtained and the data are preprocessed.
The data preprocessing step specifically comprises:
step S31: calculating the difference between the maximum amplitudes of the upstream and the downstream of the transducer in each measurement time period;
step S32: cleaning the flight time difference data and the water temperature data of the ultrasonic water meter in each measurement time period and the difference data of the maximum amplitude values of the upstream and the downstream of the transducer, and eliminating abnormal points;
step S33: interpolation supplement is carried out on the removed data abnormal points to ensure that the data lengths are consistent;
step S34: and carrying out normalization processing on the data after interpolation processing, and simultaneously splicing the three sections of data into a group of data with the length of 360.
Step S4: the water meter data belongs to a one-dimensional time sequence, so that a flow data calibration model is created based on a one-dimensional convolution neural network;
as shown in fig. 2, the specific steps are as follows:
step S41: the input data enter a first layer of convolution network, the size of convolution kernels is set to be 10, the step length is set to be 2, the number of the convolution kernels is 100, the SAME is filled in the convolution kernels, the size of a first layer of convolution output matrix is 177 x 100, and the input data are output through a Rule activation function;
the formula of the Rule function is:
f（x）=max（0，x）
wherein x is the convolution output matrix;
step S42: the result of the first layer is transmitted to the convolution of the second layer, the parameter is the same as that of the first layer, the size of the obtained output matrix is 43 x 100, and the output matrix is output through a Rule activation function;
step S43: in order to reduce the output complexity and prevent data overfitting, a maximum pooling layer is added after the second convolution network, the size is 3, the step size is 2, and an output matrix 22 is 100;
step S44: in order to learn the characteristics of higher layers, two layers of convolution networks are used, and the final output size is 2 x 120;
step S45: adding one more pooling layer to further avoid overfitting, wherein the pooling adopts average value pooling, and the output size is 1 x 120;
step S46: the generalization capability of the network is improved through a Dropout layer, an output result is obtained through a full connection layer, and probabilistic output is carried out by adopting a normalized exponential function Softmax; the formula is as follows:
wherein xiJ is the output value of the ith node, J is the number of output nodes, and J = 120.
Step S5: training and testing the flow data calibration model by using the preprocessed data; the method comprises the following specific steps:
step S51: establishing a flow data calibration model based on a one-dimensional convolutional neural network, and setting training parameters and a time threshold of the flow data calibration model; the training parameters include at least an epoch (epoch), a batch size, a number of iterations, a drop rate (dropout rate), and a learning rate; the period represents how many groups of data are needed for performing a generation of training, and the value of the embodiment is 200; the batch size indicates that several groups of data are trained each time, and the value is preferably 100; the discarding rate indicates that part of neurons are randomly allowed not to participate in calculation, namely some features are randomly lost, and the value is preferably 50%; the value of the initial learning rate is preferably 0.001;
step S52: taking 70% of the preprocessed data as a training set and 30% as a test set; training the flow data calibration model by using training set data; the model optimizes the training parameters by using an adaptive matrix estimation algorithm in the training process, namely, secondary gradient correction is introduced to search for a global optimum point, so that the learning effect is better;
step S53: judging whether the identification accuracy meets an early stop method every time the time threshold value is passed, if so, entering step S54; otherwise, go to step S52; in the embodiment, the frequency threshold is preferably 80, the period is preferably 200, that is, after 200 sets of ultrasonic water meter data are trained once, one iteration training is completed, and after 80 iterations, whether the recognition accuracy meets the early stop method is judged.
Step S54: sending the test set data into a flow data calibration model for testing, verifying whether the accuracy requirement is met, if so, entering a step S6, otherwise, skipping to S52; in the embodiment, the data satisfies the accuracy requirement, and the process proceeds to S6.
Step S6: and after normalizing the obtained ultrasonic flight time difference of the flow to be calibrated, the water temperature and the difference between the maximum amplitudes of the upstream and the downstream of the transducer, inputting a flow data calibration model for data calibration.
The above-mentioned embodiments are illustrative of the specific embodiments of the present invention, and are not restrictive, and those skilled in the relevant art can make various changes and modifications to obtain corresponding equivalent technical solutions without departing from the spirit and scope of the present invention, so that all equivalent technical solutions should be included in the scope of the present invention.
Claims (8)
1. A deep learning-based ultrasonic water meter flow data calibration method is characterized by comprising the following steps:
step S1: processing the original time sequence data of the water meter by using a metering algorithm to obtain the ultrasonic flight time difference, water temperature data and upstream and downstream signal amplitudes of the transducer in different batches, different epitopes and different temperatures within a specified measuring time period;
step S2: obtaining the real instantaneous flow of the water flow in the period of time by a standard table method;
step S3: repeating the steps S1 and S2 until N groups of data are obtained, and preprocessing the data, wherein N is an integer greater than or equal to 200;
step S4: creating a flow data calibration model based on a one-dimensional convolutional neural network;
step S5: training and testing the flow data calibration model by using the preprocessed data;
step S6: and acquiring the ultrasonic flight time difference of the flow to be calibrated, the water temperature and the difference between the maximum amplitudes of the upstream and the downstream of the transducer, and inputting the difference into a flow data calibration model to calibrate the flow data.
2. The ultrasonic water meter flow data calibration method based on deep learning of claim 1, wherein: the data preprocessing step in step S3 specifically includes:
step S31: calculating the difference between the maximum amplitudes of the upstream and the downstream of the transducer in each measurement time period;
step S32: cleaning the flight time difference data and the water temperature data of the ultrasonic water meter in each measurement time period and the difference data of the maximum amplitude values of the upstream and the downstream of the transducer, and eliminating abnormal points;
step S33: interpolation supplement is carried out on the removed data abnormal points to ensure that the data lengths are consistent;
step S34: and carrying out normalization processing on the data subjected to interpolation processing.
3. The ultrasonic water meter flow data calibration method based on deep learning of claim 1, wherein: in step S5, 70% of the preprocessed data is used as a training set, and 30% is used as a test set.
4. The ultrasonic water meter flow data calibration method based on deep learning of claim 1, wherein: the flow data calibration model in step S4 uses the Relu function as an activation function.
5. The ultrasonic water meter flow data calibration method based on deep learning of claim 1, wherein: the flow data calibration model in step S4 is output in a probabilistic manner using a normalized exponential function.
6. The ultrasonic water meter flow data calibration method based on deep learning of claim 1, wherein: the step S5 specifically includes:
step S51: establishing a flow data calibration model based on a one-dimensional convolutional neural network, and setting training parameters and a time threshold of the flow data calibration model; the training parameters at least comprise a period, a batch size, iteration times and a learning rate;
step S52: training the flow data calibration model by using training set data;
step S53: judging whether the identification accuracy meets an early stop method every time the time threshold value is passed, if so, entering step S54; otherwise, go to step S52;
step S54: and sending the test set data into a flow data calibration model for testing, verifying whether the accuracy requirement is met, if so, entering a step S6, and otherwise, skipping to S52.
7. The ultrasonic water meter flow data calibration method based on deep learning of claim 1, wherein: the flow data calibration model in step S5 optimizes the training parameters by using an adaptive matrix estimation algorithm during the training process.
8. The ultrasonic water meter flow data calibration method based on deep learning of claim 1, wherein: the step S6 specifically includes: and after normalizing the obtained ultrasonic flight time difference of the flow to be calibrated, the water temperature and the difference between the maximum amplitudes of the upstream and the downstream of the transducer, inputting a flow data calibration model for data calibration.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210131336.6A CN114166318A (en) | 2022-02-14 | 2022-02-14 | Ultrasonic water meter flow data calibration method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210131336.6A CN114166318A (en) | 2022-02-14 | 2022-02-14 | Ultrasonic water meter flow data calibration method based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114166318A true CN114166318A (en) | 2022-03-11 |
Family
ID=80489932
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210131336.6A Pending CN114166318A (en) | 2022-02-14 | 2022-02-14 | Ultrasonic water meter flow data calibration method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114166318A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114485865A (en) * | 2022-04-15 | 2022-05-13 | 青岛鼎信通讯股份有限公司 | Ultrasonic water meter flow calibration method based on Shannon entropy |
| CN116124234A (en) * | 2023-02-24 | 2023-05-16 | 宁波力擎超声科技有限公司 | Ultrasonic flowmeter for gas |
| CN116147724A (en) * | 2023-02-20 | 2023-05-23 | 青岛鼎信通讯科技有限公司 | Metering method suitable for ultrasonic water meter |
| CN116878599A (en) * | 2023-09-06 | 2023-10-13 | 青岛鼎信通讯科技有限公司 | Flow metering method of ultrasonic water meter |
| CN117131926A (en) * | 2023-10-24 | 2023-11-28 | 山东计保电气有限公司 | Electric energy metering error analysis method based on data storage |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104964719A (en) * | 2015-06-30 | 2015-10-07 | 安徽工业大学 | Household electronic water meter flow metering method based on BP neural network |
| CN109724657A (en) * | 2018-12-29 | 2019-05-07 | 杭州莱宸科技有限公司 | Watermeter flowing rate metering method and system based on modified Delphi approach |
| CN109764924A (en) * | 2019-02-25 | 2019-05-17 | 上海碳索能源环境服务有限公司 | Wide working condition high precision intelligent flow meter based on neural network model |
| CN111595401A (en) * | 2020-04-23 | 2020-08-28 | 广州能源检测研究院 | Pipeline gas measurement system, method and storage medium |
| CN113551744A (en) * | 2021-08-10 | 2021-10-26 | 中国石油大学(北京) | Ultrasonic flowmeter performance online monitoring method and system |
| CN113916329A (en) * | 2021-11-03 | 2022-01-11 | 国家石油天然气管网集团有限公司 | Natural gas flowmeter calibrating device and method based on neural network |
-
2022
- 2022-02-14 CN CN202210131336.6A patent/CN114166318A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104964719A (en) * | 2015-06-30 | 2015-10-07 | 安徽工业大学 | Household electronic water meter flow metering method based on BP neural network |
| CN109724657A (en) * | 2018-12-29 | 2019-05-07 | 杭州莱宸科技有限公司 | Watermeter flowing rate metering method and system based on modified Delphi approach |
| CN109764924A (en) * | 2019-02-25 | 2019-05-17 | 上海碳索能源环境服务有限公司 | Wide working condition high precision intelligent flow meter based on neural network model |
| CN111595401A (en) * | 2020-04-23 | 2020-08-28 | 广州能源检测研究院 | Pipeline gas measurement system, method and storage medium |
| CN113551744A (en) * | 2021-08-10 | 2021-10-26 | 中国石油大学(北京) | Ultrasonic flowmeter performance online monitoring method and system |
| CN113916329A (en) * | 2021-11-03 | 2022-01-11 | 国家石油天然气管网集团有限公司 | Natural gas flowmeter calibrating device and method based on neural network |
Non-Patent Citations (4)
| Title |
|---|
| 刘扬等: "BP神经网络流量计量的户用电子水表设计", 《仪表技术与传感器》 * |
| 姜万录等: "基于神经网络软测量的动态流量测量方法研究", 《流体传动与控制》 * |
| 陈平等: "神经网络在小管径超声波流量测试中的应用", 《计算机工程与应用》 * |
| 黄挚雄等: "基于神经网络的超声波流速测量模型", 《中南大学学报(自然科学版)》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114485865A (en) * | 2022-04-15 | 2022-05-13 | 青岛鼎信通讯股份有限公司 | Ultrasonic water meter flow calibration method based on Shannon entropy |
| CN114485865B (en) * | 2022-04-15 | 2022-09-27 | 青岛鼎信通讯股份有限公司 | Ultrasonic water meter flow calibration method based on Shannon entropy |
| CN116147724A (en) * | 2023-02-20 | 2023-05-23 | 青岛鼎信通讯科技有限公司 | Metering method suitable for ultrasonic water meter |
| CN116147724B (en) * | 2023-02-20 | 2024-01-19 | 青岛鼎信通讯科技有限公司 | Metering method suitable for ultrasonic water meter |
| CN116124234A (en) * | 2023-02-24 | 2023-05-16 | 宁波力擎超声科技有限公司 | Ultrasonic flowmeter for gas |
| CN116878599A (en) * | 2023-09-06 | 2023-10-13 | 青岛鼎信通讯科技有限公司 | Flow metering method of ultrasonic water meter |
| CN116878599B (en) * | 2023-09-06 | 2024-01-09 | 青岛鼎信通讯科技有限公司 | Flow metering method of ultrasonic water meter |
| CN117131926A (en) * | 2023-10-24 | 2023-11-28 | 山东计保电气有限公司 | Electric energy metering error analysis method based on data storage |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114166318A (en) | Ultrasonic water meter flow data calibration method based on deep learning | |
| CN109117951B (en) | BP neural network-based probability load flow online calculation method | |
| CN108304623B (en) | Probability load flow online calculation method based on stack noise reduction automatic encoder | |
| CN110443724B (en) | Electric power system rapid state estimation method based on deep learning | |
| CN112926265A (en) | Atmospheric porous probe measurement calibration method based on genetic algorithm optimization neural network | |
| CN110866448A (en) | Flutter signal analysis method based on convolutional neural network and short-time Fourier transform | |
| CN104112062B (en) | The acquisition methods of wind-resources distribution based on interpolation method | |
| CN114826543B (en) | AIOT-based steam ejector pump parameter transmission system and method | |
| WO2024087129A1 (en) | Generative adversarial multi-head attention neural network self-learning method for aero-engine data reconstruction | |
| CN109460621A (en) | Based on the gas turbine blades fault detection method for improving neural network | |
| CN113537788B (en) | Urban traffic jam recognition method based on virus propagation theory | |
| Liao et al. | Data-driven missing data imputation for wind farms using context encoder | |
| CN112395737A (en) | Method for reconstructing measurement data loss of high-speed train based on generation countermeasure network | |
| CN112861429A (en) | Wind turbine engine room transfer function calculation method | |
| CN106503861A (en) | Wind power forecasting method based on many meteorological sources wind speed fusions of probability statistics and particle group optimizing | |
| CN116992779A (en) | Simulation method and system of photovoltaic energy storage system based on digital twin model | |
| CN114330915A (en) | Short-term wind power combination model prediction method | |
| CN111222229B (en) | Method for constructing instantaneous flow measurement model in gas-liquid two-phase flow dynamic flow process | |
| CN113111573B (en) | Landslide displacement prediction method based on GRU | |
| CN114330486A (en) | Power system bad data identification method based on improved Wasserstein GAN | |
| CN107577896B (en) | Wind power plant multi-machine aggregation equivalent method based on hybrid Copula theory | |
| CN113705888A (en) | Industrial steam generation amount prediction method and system based on Pearson correlation and neural network | |
| CN117313516A (en) | Fermentation product prediction method based on space-time diagram embedding | |
| CN115098563B (en) | Time sequence abnormity detection method and system based on GCN and attention VAE | |
| CN116147724A (en) | Metering method suitable for ultrasonic water meter |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20220311 |
|
| WD01 | Invention patent application deemed withdrawn after publication |