CN112230628A - Data identification method and system for high-noise industrial process - Google Patents
Data identification method and system for high-noise industrial process Download PDFInfo
- Publication number
- CN112230628A CN112230628A CN202011244175.9A CN202011244175A CN112230628A CN 112230628 A CN112230628 A CN 112230628A CN 202011244175 A CN202011244175 A CN 202011244175A CN 112230628 A CN112230628 A CN 112230628A
- Authority
- CN
- China
- Prior art keywords
- data
- transfer function
- historical time
- frequency domain
- model
- 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
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0218—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults
- G05B23/0243—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults model based detection method, e.g. first-principles knowledge model
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B2219/00—Program-control systems
- G05B2219/20—Pc systems
- G05B2219/24—Pc safety
- G05B2219/24065—Real time diagnostics
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Automation & Control Theory (AREA)
- Feedback Control In General (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention discloses a data identification method and a data identification system for a high-noise industrial process. The method comprises the following steps: acquiring historical time series data of an industrial process; preprocessing the historical time-series data; determining a Butterworth filter based on a noise period in the historical time series data; filtering the preprocessed historical time series data through the Bass Wales filter; constructing a frequency domain transfer function model; training the frequency domain transfer function model through the filtered historical time sequence; and converting the trained frequency domain transfer function model into a time domain to obtain a time domain transfer function model as a system identification model, and identifying the data of the high-noise industrial process by using the system identification model. The invention can carry out online prediction in an industrial control environment and obtain a prediction result in real time, thereby meeting the requirements of an actual industrial process, effectively analyzing the frequency characteristic of the industrial process and explicitly identifying and filtering high-frequency fluctuation noise.
Description
Technical Field
The invention relates to the field of industrial control, in particular to a data identification method and system for a high-noise industrial process.
Background
In recent years, with the rapid development of industrial intelligence and big data technology, more and more intelligent detection and control facilities such as DCS and PLC are integrated into the industrial production process, and the automation level in the industrial field is improved. The method also provides an opportunity for grasping the operation rule of the complex equipment, predicting the complex equipment according to the past data and the existing state and further estimating the yield. However, due to the complexity of the application environment, sensor data in the actual environment is inevitably affected by high-frequency noise such as electromagnetic interference, power grid fluctuation and the like, so that the fluctuation of the data recorded by the system is far higher than the actual situation, and the system is difficult to be reliably modeled. Most parameters of the real industrial environment are slowly changed, the severe fluctuation with the frequency close to the sampling frequency is not concerned in the actual production, and the fitted model needs more trend model characteristics.
The process of system modeling is a process of searching a model which is consistent with the system, so that the output of the model is as close as possible to the real output. At present, many algorithms are used for time series prediction of complex industrial systems, mainly including linear regression, sliding autoregression, state space model, BP neural network, support vector machine, etc. The linear regression method has low complexity and is not suitable for complex industrial processes; the method can eliminate random fluctuation in prediction, but is not easy to express noise with non-0 mean and periodicity. The neural networks such as BPNN have the problems of difficulty in capturing long-term dependence of the system, poor interpretability and robustness and the like. The method such as the support vector machine can reduce the dimension of the high-dimensional feature, but the kernel function or the distance function which is necessary for the method cannot be automatically selected. The industrial process modeling algorithms have two common problems, namely that the industrial process modeling algorithms cannot explicitly identify periodic noise, and when the amplitude of a noise component is large, a system cannot correctly identify a target model; secondly, they can only construct discrete time models for discrete data collected by sensors, and when the system control or sampling frequency changes, the models often need to be retrained and predicted.
Disclosure of Invention
Based on this, the object of the present invention is to provide a data identification method and system for high-noise industrial process, which can effectively analyze the frequency characteristics of the industrial process and explicitly identify and filter the high-frequency fluctuation noise.
In order to achieve the purpose, the invention provides the following scheme:
a method of data identification of a high noise industrial process, comprising:
acquiring historical time series data of an industrial process;
preprocessing the historical time-series data;
determining a Butterworth filter based on a noise period in the historical time series data;
filtering the preprocessed historical time series data through the Bass Wales filter;
constructing a frequency domain transfer function model;
training the frequency domain transfer function model through the filtered historical time sequence;
and converting the trained frequency domain transfer function model into a time domain to obtain a time domain transfer function model as a system identification model, and identifying the data of the high-noise industrial process by using the system identification model.
Optionally, the preprocessing the historical time-series data specifically includes:
completing, removing a mean value and carrying out differential processing on the historical time series data;
the processed data is converted to the frequency domain using a laplace transform.
Optionally, the determining a basterworth filter based on the noise period in the historical time-series data specifically includes:
determining parameters of a Butterworth filter based on a period of noise in the historical time series data;
and determining the Butterworth filter according to the parameters of the Butterworth filter.
Optionally, the frequency domain transfer function model g(s) is:
where s is a complex frequency domain operator,
in order to respond to the parameters of the function,
for the parameters of the driving function, a (m), b (m) are parameters to be learned by the model, m is the number of model zeros, n is the order of the model, y(s) is the response function, and u(s) is the driving function.
Optionally, the training of the frequency-domain transfer function model by the filtered historical time sequence specifically includes:
inputting the filtered historical time sequence into a frequency domain transfer function model to obtain output data;
judging whether the error between the output data and the actual value is within a threshold value range;
if so, determining the frequency domain transfer function model as a trained frequency domain transfer function model;
if not, adjusting the parameters of the frequency domain transfer function model through a least square method to enable the error between the output data and the actual value to be within a threshold range.
A data identification system for a high noise industrial process, comprising:
the data acquisition module is used for acquiring historical time series data of the industrial process;
the preprocessing module is used for preprocessing the historical time sequence data;
a filter determination module that determines a Butterworth filter based on a noise period in the historical time series data;
the filtering module is used for filtering the preprocessed historical time sequence data through the Bass Walsh filter;
the model construction module is used for constructing a frequency domain transfer function model;
the training module is used for training the frequency domain transfer function model through the filtered historical time sequence;
and the conversion module is used for converting the trained frequency domain transfer function model into a time domain to obtain a time domain transfer function model as a system identification model, and identifying the data of the high-noise industrial process by using the system identification model.
Optionally, the preprocessing module specifically includes:
the processing unit is used for performing completion, mean value removal and difference processing on the historical time series data;
a conversion unit for converting the processed data into a frequency domain using a laplace transform.
Optionally, the filter determining module specifically includes:
a parameter determination unit that determines a parameter of a Butterworth filter based on a noise cycle in the historical time-series data;
and the filter determining unit is used for determining the Butterworth filter according to the parameters of the Butterworth filter.
Optionally, the training module specifically includes:
the input unit is used for inputting the filtered historical time sequence into the frequency domain transfer function model to obtain output data;
the judging unit is used for judging whether the error between the output data and the actual value is within a threshold value range or not;
a result determining unit, configured to determine that the frequency domain transfer function model is a trained frequency domain transfer function model when an error between the output data and an actual value is within a threshold range;
and when the error between the output data and the actual value is in a threshold range, adjusting the parameters of the frequency domain transfer function model by a least square method to ensure that the error between the output data and the actual value is in the threshold range.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
the invention preprocesses the historical time sequence data; determining a Butterworth filter based on a noise period in the historical time series data; filtering the preprocessed historical time series data through the Bass Wales filter; constructing a frequency domain transfer function model; training the frequency domain transfer function model through the filtered historical time sequence; and converting the trained frequency domain transfer function model into a time domain to obtain a time domain transfer function model as a system identification model, and identifying the data of the high-noise industrial process by using the system identification model. The system identification model provided by the invention can carry out online prediction in an industrial control environment and obtain a prediction result in real time, thereby meeting the requirement of an actual industrial process, effectively analyzing the frequency characteristic of the industrial process and explicitly identifying and filtering high-frequency fluctuation noise.
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 data identification of a high noise industrial process according to an embodiment of the present invention;
FIG. 2 is a graph of the amplitude, frequency, phase and frequency of a fifth order Butterworth low pass filter according to an embodiment of the present invention;
FIG. 3 is a block diagram of a data identification system for a high noise industrial process according to an embodiment of 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 data identification method and a data identification system for a high-noise industrial process, which can effectively analyze the frequency characteristic of the industrial process and explicitly identify and filter high-frequency fluctuation noise.
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, a data recognition method of a high noise industrial process includes the steps of:
step 101: historical time series data of an industrial process is obtained.
Step 102: and preprocessing the historical time-series data.
Step 103: a basteworth filter is determined based on a period of noise in the historical time series data.
Step 104: and filtering the preprocessed historical time sequence data through the Bass Waters filter.
Step 105: and constructing a frequency domain transfer function model.
Step 106: and training the frequency domain transfer function model through the filtered historical time sequence.
Step 107: and converting the trained frequency domain transfer function model into a time domain to obtain a time domain transfer function model as a system identification model, and identifying the data of the high-noise industrial process by using the system identification model.
Wherein, step 102: and preprocessing the historical time-series data. The method specifically comprises the following steps:
completing, removing a mean value and carrying out differential processing on the historical time series data; the processed data is converted to the frequency domain using a laplace transform.
Step 103: a basteworth filter is determined based on a period of noise in the historical time series data. The method specifically comprises the following steps:
determining parameters of a Butterworth filter based on a period of noise in the historical time series data; and determining the Butterworth filter according to the parameters of the Butterworth filter.
Step 105: and constructing a frequency domain transfer function model. The method specifically comprises the following steps:
the frequency domain transfer function model g(s) is:
where s is a complex frequency domain operator,
in order to respond to the parameters of the function,
for the parameters of the driving function, a (m), b (m) are parameters to be learned by the model, m is the number of model zeros, n is the order of the model, y(s) is the response function, and u(s) is the driving function.
Step 106: and training the system recognition model through the filtered historical time sequence. The method specifically comprises the following steps:
step 1061: inputting the filtered historical time sequence into the system identification model to obtain output data;
step 1062: judging whether the error between the output data and the actual value is within a threshold value range;
step 1063: if so, determining the system identification model as a trained system identification model;
step 1064: if not, adjusting the parameters of the system identification model through a least square method to enable the error between the output data and the actual value to be within a threshold range.
The invention is described in detail below by means of specific examples:
step 101: historical time series data of an industrial process is obtained.
Assuming an industrial scene, the system external factor input quantity at the time t obtained from the sensor is utThe unknown system noise is e and the system output to be solved is ytK is the sliding window size of the expectation at model prediction,
in the present invention, the model is expressed by a function f (·), and the system is expressed in the time domain as:
step 102: and preprocessing the historical time-series data.
In order to avoid the influence of the characteristics of data with large bias and non-stationary time series on model identification, the data after passing through the filter needs to be subjected to mean value removal and trend removal. The mean value removal requires directly and respectively averaging the parameters of the data set, and then explicitly subtracting the interference from each data set, as shown in the following formula, ymAnd urWhich represents the original data of the image data,
and
the filtered data is represented by the filtered data,
and
to average:
trending requires differencing the data as shown in the following equation:
when data acquired by a sensor is missing at random time, in order to ensure the prediction effect of the model, a linear interpolation method is used for complementing the missing data, and linear interpolation or spline interpolation is generally selected according to the characteristics of the data.
By performing laplacian transform on each input/output of the processed system, the system can be represented in the frequency domain as:
where s is the laplacian operator, and f (·) is divided into two parts, including the filter function l(s) and the system identification model g(s), which are described in further detail below.
Firstly, in order to ensure that the input-output relationship of the principle ideal system is not changed, the invention must use a uniform filter to filter all the input and output of the system, and use G to filter0(s) represents an ideal fit system, assuming that the r input m output system to be fitted satisfies:
wherein
Is the input vector of the system and is,
is the output vector of the system and is,
is a zero mean random interference noise vector.
Step 103: a basteworth filter is determined based on a period of noise in the historical time series data.
The filter is a Butterworth low-pass filter with three parameters, namely the order of the filter and the cut-off frequency F reduced by 3dBcSampling frequency Fs. Wherein, in order to control the response speed and ringing effect, the fifth order is selected in this embodiment, and the sampling frequency F of the system is assumedc=200Hz, in which there is F produced by an inverter motor1Interference of =80Hz, and F caused by AC2Interference of =50Hz, and a cutoff frequency F is set for suppressing the interferencec=min(F1,F2) =50Hz, the amplitude-frequency-phase-frequency diagram of the fifth-order butterworth low-pass filter is shown in fig. 2.
When the system does not have the condition of using a hardware filter, the filter can be converted into a digital filter form. The system expression of the fifth-order Butterworth low-pass filter in the s domain is as follows:
(ii) a The molecular coefficient matrix of the direct-coupled I-type digital filter converted into the s-domain by using a bilinear transform method is
The denominator coefficient matrix is
The algorithm can calculate and deduce the required digital filter parameters in real time only by a sin/cos table, so the algorithm is also suitable for an embedded system needing online adjustment of the filter parameters.
Step 104: and filtering the preprocessed historical time sequence data through the Bass Waters filter.
In this embodiment, all data is input into MATLAB, and filtered data can be obtained by using filter (node, de, data) (node is a numerator of a transfer function, de is a denominator, and data is data to be filtered).
When the hardware filter is used in a production environment, the filter with the adjusted parameters is placed in front of the DCS data collection system, and then the filtered data can be directly obtained.
Step 105: and constructing a frequency domain transfer function model.
The model g(s) is a frequency domain transfer function model defined as the ratio of the laplacian transform of the output quantities (response function) of the system to the laplacian transform of the input quantities (drive function) under the assumption that all initial conditions are zero. The model defines two parameter matrices, response function parameters b (m) and drive function parameters a (n), where m represents the number of zeros in the model (where m is a common expression independent of the number of system outputs in the foregoing), n represents the number of poles in the model, and n is also referred to as the order of the system. The time domain model of a certain input to a certain output of the system is:
and (3) taking pull type transformation from two ends of the formula, wherein the transformation is in a single-input single-output form of the model:
the r-input m-output system in this example is defined according to the following transfer matrix g(s):
wherein, gij(s) is the transfer function between the jth input to the ith output, in the same form as G(s) in a single-input single-output system.
Step 106: and training the system recognition model through the filtered historical time sequence.
The number of generation fitting parameters in the model is m + n. The number of the system data which can be obtained is far larger than the number of the parameters, so the initial state y (t) of the system can be estimated by directly using least square0) A response function parameter B (m), and a drive function parameter A (n). The estimation process needs to minimize the error of the identification system and the real data, i.e. the following:
wherein
Is formed by the parameter y (t) to be solved0) B (m), A (n) form a vector set,
is the data set of the training set, N is the length of the time series in the data set,
when the formula is calculated, the formula is changed into a form of multiplying an orthogonal matrix by a triangular matrix by using a QR decomposition method, namely, the formula is firstly matched
The sea castle matrix is formed through orthogonal similarity change, and then the matrix is decomposed into a regular orthogonal matrix Q and an upper triangular matrix R, so that generation solving parameters can be obtained.
Model error VNThe Mean Square Error (MSE) calculation is used, and the calculation method comprises the following steps:
wherein
Is the predicted output of the recognition system, yiIs the true output in the dataset, | x | | | represents the L2 norm of x.
As the MSE is related to the extreme value range, variance and other factors of the original data, technicians cannot directly know the fitting accuracy of the model, the introduction of the R-squared fitting value facilitates the manual measurement of the model effect, and the fitting value
The higher the approximation of the system under test to the real system, the better:
。
and calculating the MSE and Fit values of the system according to the result after the parameter optimization, if the accuracy of the system is not in the threshold range, adjusting the order of the transfer function and the number of the numerators and the denominators, and continuously using the least square method for optimization.
Step 107: and converting the trained frequency domain transfer function model into a time domain state space model to serve as a system identification model of the production environment, and identifying the data of the high-noise industrial process by using the system identification model.
After a satisfactory identification result is obtained, the identification model can be converted into a time domain and then deployed to an industrial field for online system prediction. The conversion method generally comprises the steps of firstly solving a signal block diagram and a signal flow diagram of a transfer function by utilizing a Meisen formula, then selecting the output of an integrator as a state variable, and obtaining a state equation and an output equation so as to obtain a state space model. The tf2ss function of the MATLAB itself can also be used to convert the differential equations in the transfer function model into a state space model.
The state space model is defined according to the following formula:
the following table shows the simulation result of whether the pre-filtering module is added in this example, and it can be seen that the fitting effect is significantly improved by the added filtering module:
as shown in fig. 3, the present invention also provides a data recognition system for a high noise industrial process, comprising:
a data acquisition module 301 for acquiring historical time series data of the industrial process.
A preprocessing module 302, configured to preprocess the historical time-series data.
The preprocessing module 302 specifically includes:
the processing unit is used for performing completion, mean value removal and difference processing on the historical time series data;
a conversion unit for converting the processed data into a frequency domain using a laplace transform.
A filter determination module 303 for determining a Butterworth filter based on the noise periods in the historical time series data.
The filter determining module 303 specifically includes:
a parameter determination unit that determines a parameter of a Butterworth filter based on a noise cycle in the historical time-series data;
and the filter determining unit is used for determining the Butterworth filter according to the parameters of the Butterworth filter.
A filtering module 304, configured to filter the pre-processed historical time series data through the basews filter.
A model construction module 305 for constructing a frequency domain transfer function model.
A training module 306, configured to train the frequency-domain transfer function model through the filtered historical time sequence.
The training module 306 specifically includes:
the input unit is used for inputting the filtered historical time sequence into the frequency domain transfer function model to obtain output data;
the judging unit is used for judging whether the error between the output data and the actual value is within a threshold value range or not;
a result determining unit, configured to determine that the frequency domain transfer function model is a trained frequency domain transfer function model when an error between the output data and an actual value is within a threshold range;
and when the error between the output data and the actual value is in a threshold range, adjusting the parameters of the frequency domain transfer function model by a least square method to ensure that the error between the output data and the actual value is in the threshold range.
And the identifying module 307 is configured to convert the trained frequency domain transfer function model into a time domain, obtain a time domain transfer function model as a system identification model, and identify data of the high-noise industrial process by using the system identification model.
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 principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.
Claims (9)
1. A method of data identification of a high noise industrial process, comprising:
acquiring historical time series data of an industrial process;
preprocessing the historical time-series data;
determining a Butterworth filter based on a noise period in the historical time series data;
filtering the preprocessed historical time series data through the Bass Wales filter;
constructing a frequency domain transfer function model;
training the frequency domain transfer function model through the filtered historical time sequence;
and converting the trained frequency domain transfer function model into a time domain to obtain a time domain transfer function model as a system identification model, and identifying the data of the high-noise industrial process by using the system identification model.
2. The method for data recognition of a high-noise industrial process according to claim 1, wherein the pre-processing of the historical time-series data specifically comprises:
completing, removing a mean value and carrying out differential processing on the historical time series data;
the processed data is converted to the frequency domain using a laplace transform.
3. The method for data identification of a high noise industrial process according to claim 1, wherein the determining a Butterworth filter based on the noise periods in the historical time series data specifically comprises:
determining parameters of a Butterworth filter based on a period of noise in the historical time series data;
and determining the Butterworth filter according to the parameters of the Butterworth filter.
4. The method for data identification of a high noise industrial process according to claim 1, wherein the frequency domain transfer function model g(s) is:
where s is a complex frequency domain operator,
in order to respond to the parameters of the function,
for the parameters of the driving function, a (m), b (m) are parameters to be learned by the model, m is the number of model zeros, n is the order of the model, y(s) is the response function, and u(s) is the driving function.
5. The method for data recognition of a high-noise industrial process according to claim 1, wherein the training of the frequency-domain transfer function model by the filtered historical time series specifically comprises:
inputting the filtered historical time sequence into a frequency domain transfer function model to obtain output data;
judging whether the error between the output data and the actual value is within a threshold value range;
if so, determining the frequency domain transfer function model as a trained frequency domain transfer function model;
if not, adjusting the parameters of the frequency domain transfer function model through a least square method to enable the error between the output data and the actual value to be within a threshold range.
6. A system for data recognition of a high noise industrial process, comprising:
the data acquisition module is used for acquiring historical time series data of the industrial process;
the preprocessing module is used for preprocessing the historical time sequence data;
a filter determination module that determines a Butterworth filter based on a noise period in the historical time series data;
the filtering module is used for filtering the preprocessed historical time sequence data through the Bass Walsh filter;
the model construction module is used for constructing a frequency domain transfer function model;
the training module is used for training the frequency domain transfer function model through the filtered historical time sequence;
and the conversion module is used for converting the trained frequency domain transfer function model into a time domain to obtain a time domain transfer function model as a system identification model, and identifying the data of the high-noise industrial process by using the system identification model.
7. The system for data recognition of a high-noise industrial process according to claim 6, wherein the preprocessing module comprises:
the processing unit is used for performing completion, mean value removal and difference processing on the historical time series data;
a conversion unit for converting the processed data into a frequency domain using a laplace transform.
8. The data identification system of a high-noise industrial process according to claim 6, wherein the filter determination module specifically comprises:
a parameter determination unit that determines a parameter of a Butterworth filter based on a noise cycle in the historical time-series data;
and the filter determining unit is used for determining the Butterworth filter according to the parameters of the Butterworth filter.
9. The method of claim 6, wherein the training module comprises:
the input unit is used for inputting the filtered historical time sequence into the frequency domain transfer function model to obtain output data;
the judging unit is used for judging whether the error between the output data and the actual value is within a threshold value range or not;
a result determining unit, configured to determine that the frequency domain transfer function model is a trained frequency domain transfer function model when an error between the output data and an actual value is within a threshold range;
and when the error between the output data and the actual value is in a threshold range, adjusting the parameters of the frequency domain transfer function model by a least square method to ensure that the error between the output data and the actual value is in the threshold range.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011244175.9A CN112230628B (en) | 2020-11-10 | 2020-11-10 | Data identification method and system for high-noise industrial process |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011244175.9A CN112230628B (en) | 2020-11-10 | 2020-11-10 | Data identification method and system for high-noise industrial process |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112230628A true CN112230628A (en) | 2021-01-15 |
| CN112230628B CN112230628B (en) | 2021-11-16 |
Family
ID=74122273
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011244175.9A Active CN112230628B (en) | 2020-11-10 | 2020-11-10 | Data identification method and system for high-noise industrial process |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112230628B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113128072A (en) * | 2021-05-13 | 2021-07-16 | 清鸾科技(成都)有限公司 | High-precision transfer function simulation method and device, storage medium and electronic equipment |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4571563A (en) * | 1983-11-18 | 1986-02-18 | Agence Spatiale Europeenne | Integrated microwave filter and method of constructing same |
| US7613311B2 (en) * | 2004-12-15 | 2009-11-03 | Cirrus Logic, Inc | Digital implementation of a fourth order Linkwitz-Riley network with a low cutoff frequency |
| CN106227964A (en) * | 2016-07-28 | 2016-12-14 | 华侨大学 | Nonlinear systems with hysteresis parameter identification method based on extended state observer |
| CN110008784A (en) * | 2018-01-04 | 2019-07-12 | 西安交通大学 | Milling Force recognition methods and identifying system based on conjugate gradient least-squares algorithm |
| CN110766134A (en) * | 2019-09-25 | 2020-02-07 | 福州大学 | Photovoltaic power station short-term power prediction method based on cyclic neural network |
-
2020
- 2020-11-10 CN CN202011244175.9A patent/CN112230628B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4571563A (en) * | 1983-11-18 | 1986-02-18 | Agence Spatiale Europeenne | Integrated microwave filter and method of constructing same |
| US7613311B2 (en) * | 2004-12-15 | 2009-11-03 | Cirrus Logic, Inc | Digital implementation of a fourth order Linkwitz-Riley network with a low cutoff frequency |
| CN106227964A (en) * | 2016-07-28 | 2016-12-14 | 华侨大学 | Nonlinear systems with hysteresis parameter identification method based on extended state observer |
| CN110008784A (en) * | 2018-01-04 | 2019-07-12 | 西安交通大学 | Milling Force recognition methods and identifying system based on conjugate gradient least-squares algorithm |
| CN110766134A (en) * | 2019-09-25 | 2020-02-07 | 福州大学 | Photovoltaic power station short-term power prediction method based on cyclic neural network |
Non-Patent Citations (3)
| Title |
|---|
| KUBANEK, DAVID: "Transfer Functions of Fractional-Order Band-Pass Filter with Arbitrary Magnitude Slope in Stopband", 《2019 42ND INTERNATIONAL CONFERENCE ON TELECOMMUNICATIONS AND SIGNAL PROCESSING (TSP)》 * |
| MAHATA, SHIBENDU: "Optimal approximation of asymmetric type fractional-order bandpass Butterworth filter using decomposition technique", 《INTERNATIONAL JOURNAL OF CIRCUIT THEORY AND APPLICATIONS》 * |
| 罗抟翼: "《信号、***与自动控制原理》", 31 October 2000, 机械工业出版社 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113128072A (en) * | 2021-05-13 | 2021-07-16 | 清鸾科技(成都)有限公司 | High-precision transfer function simulation method and device, storage medium and electronic equipment |
| CN113128072B (en) * | 2021-05-13 | 2024-01-19 | 清鸾科技(成都)有限公司 | Transfer function high-precision simulation method and device, storage medium and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112230628B (en) | 2021-11-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113609955B (en) | Three-phase inverter parameter identification method based on deep learning and digital twin | |
| CN103559540B (en) | Based on the wind speed ultra-short term on-line prediction method of Adaptive Neuro-fuzzy Inference | |
| CN111241744B (en) | Low-pressure casting machine time sequence data abnormity detection method based on bidirectional LSTM | |
| CN109033513B (en) | Power transformer fault diagnosis method and power transformer fault diagnosis device | |
| CN112067925B (en) | Real-time weighted fault detection method for boost converter circuit | |
| CN111427266A (en) | Nonlinear system identification method aiming at disturbance | |
| CN112230628B (en) | Data identification method and system for high-noise industrial process | |
| CN113029559B (en) | Gear box fault diagnosis method and system | |
| CN115484102A (en) | Industrial control system-oriented anomaly detection system and method | |
| CN117104377B (en) | Intelligent management system and method for electric bicycle | |
| CN117784710A (en) | Remote state monitoring system and method for numerical control machine tool | |
| CN117458955A (en) | Operation control method and system of motor | |
| CN114755010A (en) | Rotary machine vibration fault diagnosis method and system | |
| CN113159088B (en) | Fault monitoring and diagnosis method based on multi-feature fusion and width learning | |
| CN113961613A (en) | Linear prediction method based on periodic filtering | |
| CN115936473A (en) | Unsupervised KPI (Key performance indicator) abnormity detection method combining prediction and reconstruction | |
| CN114818813A (en) | Cutter wear value prediction algorithm based on self-attention convolution neural network | |
| CN115130500A (en) | Motor fault diagnosis system and method based on SDAE algorithm | |
| Liu et al. | Self-tuning information fusion Wiener filter for ARMA signals and its convergence | |
| CN111048203B (en) | Brain blood flow regulator evaluation device | |
| Bettayeb et al. | Reduced order H/sub/spl infin//filtering | |
| CN117404853B (en) | External circulating water cooling system and method for tunnel boring machine | |
| CN117886241B (en) | Tower crane self-checking system based on vibration analysis | |
| CN117388637B (en) | AI-based converter station direct current system abnormal signal identification and auxiliary decision-making method | |
| CN117131465B (en) | Single-pipe tower damage identification method and device, electronic equipment and readable storage medium |
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 |