CN112365458A - Grate cooler snowman identification method and system based on ANN neural network - Google Patents
Grate cooler snowman identification method and system based on ANN neural network Download PDFInfo
- Publication number
- CN112365458A CN112365458A CN202011201201.XA CN202011201201A CN112365458A CN 112365458 A CN112365458 A CN 112365458A CN 202011201201 A CN202011201201 A CN 202011201201A CN 112365458 A CN112365458 A CN 112365458A
- Authority
- CN
- China
- Prior art keywords
- data
- snowman
- grate cooler
- neural network
- ann
- 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
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 34
- 238000000034 method Methods 0.000 title claims abstract description 18
- 238000004140 cleaning Methods 0.000 claims abstract description 22
- 238000004519 manufacturing process Methods 0.000 claims abstract description 17
- 239000004568 cement Substances 0.000 claims abstract description 10
- 210000002569 neuron Anatomy 0.000 claims description 18
- 238000003062 neural network model Methods 0.000 claims description 11
- 238000013459 approach Methods 0.000 claims description 6
- 238000009825 accumulation Methods 0.000 claims description 5
- 238000002372 labelling Methods 0.000 claims description 5
- 230000004913 activation Effects 0.000 claims description 3
- 230000006870 function Effects 0.000 claims description 3
- 230000001537 neural effect Effects 0.000 claims 2
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 238000013135 deep learning Methods 0.000 abstract description 2
- 238000005516 engineering process Methods 0.000 abstract description 2
- 230000035945 sensitivity Effects 0.000 abstract description 2
- 239000000779 smoke Substances 0.000 description 6
- 230000007423 decrease Effects 0.000 description 3
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 239000000428 dust Substances 0.000 description 2
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000001354 calcination Methods 0.000 description 1
- 238000001816 cooling Methods 0.000 description 1
- 238000007599 discharging Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 229910052742 iron Inorganic materials 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000005728 strengthening Methods 0.000 description 1
- 238000009423 ventilation Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F27—FURNACES; KILNS; OVENS; RETORTS
- F27D—DETAILS OR ACCESSORIES OF FURNACES, KILNS, OVENS, OR RETORTS, IN SO FAR AS THEY ARE OF KINDS OCCURRING IN MORE THAN ONE KIND OF FURNACE
- F27D15/00—Handling or treating discharged material; Supports or receiving chambers therefor
- F27D15/02—Cooling
- F27D15/0206—Cooling with means to convey the charge
- F27D15/0213—Cooling with means to convey the charge comprising a cooling grate
- F27D15/022—Cooling with means to convey the charge comprising a cooling grate grate plates
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/048—Activation functions
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Mechanical Engineering (AREA)
- Quality & Reliability (AREA)
- Feedback Control In General (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention relates to the field of deep learning of artificial intelligence technology, in particular to a grate cooler snowman identification method and a grate cooler snowman identification system based on an ANN neural network, wherein the method comprises the following steps: analyzing the performance characteristics of the formed grate cooler snowman, determining a characteristic model of the formed grate cooler snowman, analyzing a business data rule of the formed grate cooler snowman, constructing a data rule engine, acquiring offline business data according to characteristic parameters of the characteristic model, cleaning, marking the cleaned offline business data, constructing and training an ANN artificial neural network model, acquiring relevant parameters of the grate cooler working condition in real time, cleaning and assembling according to a data standard range defined by the data rule engine, inputting the assembled parameters into the trained grate cooler snowman recognition model, and acquiring a recognition result from the grate cooler snowman recognition model. The method has the advantages of high intelligence, sensitivity, accuracy and the like, and can accurately identify the occurrence of the grate cooler snowman phenomenon in the cement production process in time.
Description
Technical Field
The invention relates to the field of deep learning of artificial intelligence technology, in particular to a grate cooler snowman identification method and system based on an ANN neural network.
Background
At present, in the production process of cement in a novel dry kiln, a cooling mode of clinker is basically cooled by a grate cooler. In the actual production process, the grate plate can not push away the discharged clinker in time due to various reasons, so that the discharged clinker at the discharging chute at the turning rear side of the front wall of the grate cooler and the rotary kiln cylinder is higher and higher, and the phenomenon of snowman is caused.
The snowman is an abnormal phenomenon which often occurs in the grate cooler, and the damage of the snowman is very serious. After the snowman is formed, the ventilation of the system, the secondary air quantity entering the kiln and the air temperature can be influenced, the heat balance of the kiln and a preheater system is damaged, the calcination in the kiln is poor, the yield and the quality of clinker are reduced, the positive pressure of the kiln head and the material leakage at the tail of the kiln are caused in serious cases, and the iron protection at the mouth of the kiln is abraded seriously.
In the existing cement production process control, a method of carrying out inspection before a grate cooler by strengthening manpower on site is generally adopted so as to find the occurrence of the phenomenon of snowman accumulation of the grate cooler in time and carry out treatment in time. The method is limited by environment and subjective factors of people, so that a worker cannot be guaranteed to watch in front of the grate cooler at any time and any place, the phenomenon of snowman accumulation can be found in time, and once the phenomenon of snowman accumulation occurs but the phenomenon is found out not timely enough, the snowman can be piled too much and collapsed, so that the grate bed is pressed to be dead, and the accident of the press bed is caused.
Disclosure of Invention
In order to solve the technical problems in the prior art, the invention provides a grate cooler snowman identification method and system based on an ANN neural network, and the specific technical scheme is as follows.
A grate cooler snowman recognition system based on an ANN neural network comprises a main device, a feature set definition device, a training set device, an ANN artificial neural network training device, a data rule engine, a data cleaning device and a grate cooler snowman recognition device, wherein the feature set definition device, the training set device, the ANN artificial neural network training device, the data rule engine, the data cleaning device and the grate cooler snowman recognition device are arranged on the main device;
the characteristic set definition device defines factors influencing the grate cooler snowman, determines the characteristic set of the grate cooler snowman, and inputs the characteristic set as an input variable of the ANN artificial neural network to the training set device;
the training set device collects the off-line business historical data of the kiln condition in the production process according to the feature set, covers the parameters defined in the feature set, cleans the off-line business historical data, completes data labeling and inputs the data to the ANN neural network model training device;
the ANN neural network model training device is used for performing semi-supervised learning on a large amount of input grate cooler snowman offline service historical data;
the data rule engine device records the rule characteristics in the device in a form of a table;
the data cleaning device acquires relevant characteristic parameter data of the grate cooler snowman in the production process in real time, and performs data cleaning and assembling on the characteristic parameter data according to rules preset by a data rule engine;
the grate cooler snowman identification device identifies and classifies real-time data related to grate cooler input in the production process, and the classification comprises two normal categories of grate cooler snowman accumulation and grate cooler.
Further, the data labeling is to label the data sample meeting the snowman standard as the snowman, and the data sample not meeting the standard is normal.
Furthermore, the ANN neural network model training device is provided with a neuron structure, the neuron structure comprises an input layer, a first hidden layer, a second hidden layer, a third hidden layer and an output layer, and the semi-supervised learning means that the actual output value of the ANN neural network model device approaches to the expected output by giving a training set of the snowman expected to output and continuously adjusting the weight of each neuron in each hidden layer.
Further, the rule characteristics include: the method comprises the following steps of data acquisition frequency, data acquisition range, data standard range, data difference value threshold, acquisition type and algorithm operation frequency, wherein the acquisition type is a mean value or interpolation.
Further, when the difference value of the related characteristic parameter data and the preset standard exceeds a preset deviation threshold value, the data is removed and recorded into a dirty data recording table in a table form.
An ANN neural network-based identification method of a grate cooler snowman identification system comprises the following steps:
step S01: analyzing the performance characteristics of the formed grate cooler snowman, and determining a characteristic model of the grate cooler snowman;
step S02: analyzing the data rule characteristics formed by the grate cooler snowman and constructing a data rule engine according to the data rule characteristics, wherein the rule characteristics of the rule engine are stored in a form of a table;
step S03: acquiring offline service data according to the characteristic parameters of the characteristic model, cleaning the offline service data according to the characteristic range of the data, removing data outside the characteristic range of the data, and maintaining an offline service data record table by the identification system;
step S04: marking the off-line service data cleaned in the step S03 according to the characteristic model, wherein the marking is carried out in a general combination mode;
step S05: constructing an ANN artificial neural network model;
step S06: training an ANN artificial neural network model;
step S07: acquiring relevant parameter data of the working condition of the grate cooler in the cement production process in real time, cleaning and assembling according to a data standard range defined by a data rule engine, and then storing in a table form;
step S08: taking out the relevant parameters of the assembled grate cooler under the working condition, and inputting the trained ANN artificial neural network model according to the defined algorithm operation frequency;
step S09: and obtaining a model output value from the recognition model of the grate cooler snowman for recognition.
Further, the step S06 specifically includes:
s6.1, acquiring data input from the training device, and adjusting according to the weight;
s6.2, each layer uses a PReLU activation function, formula f (x) = max (0, x), and each neuron is controlled to output through nonlinearity, so that the generalization capability of the model is improved;
s6.3, calculating a reverse propagation error rate c, obtaining a prediction output value h of each neuron, wherein the expected value of a training set is n, and the reverse propagation error rate c = h-n;
s6.4, adjusting through an error weighted derivative formula according to the obtained error range;
s6.5, the training step length is 15, the training precision is 15-8, the steps S6.1-S6.3 are repeated for not less than 20000 times until the training is finished, and the output value of the model approaches to the expected value of the training set.
Further, the step S07 specifically includes:
s7.1, cleaning the acquired data according to a data standard range defined by a data rule engine;
s7.2, calculating the average value of 24 moments at the current moment and every 5 minutes before the current moment by the data type of the average value initially;
s7.3, the mean value of each moment of S7.2 is assembled into parameters and stored in a table form.
The identification method and the identification system provided by the invention have the advantages of high intellectualization, sensitivity, accuracy and the like, and can be used for accurately identifying the occurrence of the grate cooler snowman phenomenon in the cement production process in time.
Drawings
FIG. 1 is a schematic structural diagram of a grate cooler snowman identification model of the invention;
FIG. 2 is a schematic flow chart of an embodiment of identifying a grate cooler snowman according to the present invention;
FIG. 3 is a schematic diagram of a system device for identifying a grate cooler snowman according to the invention.
Detailed Description
In order to make the objects, technical solutions and technical effects of the present invention more apparent, the present invention is further described in detail below with reference to the accompanying drawings and the embodiments.
As shown in fig. 3, the system comprises a main device, and a feature set definition device, a training set device, an ANN artificial neural network training device, a data rule engine device, a data cleaning device and a grate cooler snowman recognition device which are arranged on the main device.
The feature set definition device is used for defining factors influencing the grate cooler snowman, determining the factors as the feature set of the grate cooler snowman, taking the feature set as an input variable of the ANN artificial neural network system, and inputting the input variable into the training set device.
The training set device is used for collecting off-line business historical data of kiln conditions in the cement production process according to the feature set, covering parameters defined in the feature set, cleaning the off-line data and completing data marking, wherein the data marking labels the data samples meeting the standard of the snowman as the snowman according to expert experience, and the data samples not meeting the standard are normal.
The ANN neural network model training device is characterized in that a neuron structure of the ANN neural network model training device is designed by five layers and comprises an input layer, a first hidden layer, a second hidden layer, a third hidden layer and an output layer, input offline data of a large number of grate cooler snowman are subjected to semi-supervised learning, the semi-supervised learning refers to that the actual output value of the ANN neural network model training device approaches to expected output by giving a training set of the snowman expected to be output and continuously adjusting the weight of each neuron in each hidden layer.
The data rule engine device is used for recording rule characteristics such as data acquisition frequency v, a data acquisition range f, a data standard range b, a data difference threshold value y, an acquisition type t and the like in the device in a table form, wherein the data acquisition frequency v < =5 (minutes), and v > =1 minute; the data acquisition range f is 1 mean value of every 5 minutes in nearly 20 minutes, and 4 time values are obtained in total; the acquisition types t are divided into two types, and t is a mean value or interpolation value.
The data cleaning device is used for obtaining relevant characteristic parameters of the grate cooler snowman in the cement production process in real time, cleaning the collected characteristic parameter data of the grate cooler snowman according to rules preset by a data rule engine, removing the data when the difference value of the data and a preset standard exceeds a preset deviation threshold value, and recording the data into a dirty data recording table in a table form.
The grate cooler snowman recognition device is used for classifying the real-time data related to the grate cooler in the cement production process, and the classification comprises two normal categories of grate cooler snowman stacking and grate cooler.
As shown in fig. 1 and fig. 2, the system and method for identifying the snowman of the grate cooler based on the ANN neural network comprises the following steps:
step S01: analyzing the performance characteristics of the formed grate cooler snowman, and determining a characteristic model X of the grate cooler snowman, wherein the characteristic model X comprises 9 characteristic sets and a data characteristic range: grate cooler F2 air chamber v1, secondary air temperature v2, kiln tail smoke chamber gas NOx concentration v3, tertiary air temperature v4, kiln current v5, kiln door cover differential pressure v6, kiln tail smoke chamber gas analyzer CO concentration v7, electric dust collection inlet temperature II v8, discharge grate cooler clinker temperature v9, characteristic model X = { v1|7000< = v1< =9000, v2| -300< = v2< =0, v3|0< = v3< =1500, v3| 800< = v3< =1200, v3| 400< = v3< =1000, v3| -300< = 3< =0, v3| 3< = v < = 3< = 3650 };
step S02: and constructing a data rule engine, and defining the data acquisition frequency st, the data acquisition range sf, the data standard range ss, the data type sl and the algorithm operation frequency. The rule characteristics of the rule engine are stored in a table form;
step S03: acquiring offline service data according to the characteristic parameters of the characteristic model, cleaning the offline service data according to the characteristic range of the data, removing data outside the characteristic range of the data, and maintaining an offline service data record table by the identification system;
step S04: according to the feature model in the step 0S1, the offline service data cleaned in the step S03 is labeled, the labeling is performed by adopting a method of universally combining 4 feature sets, 4 feature sets in the X model in the step S1 are selected, including a secondary air temperature v2, a kiln tail smoke chamber gas NOx concentration v3, a tertiary air temperature v4 and a kiln current v5, to form a large set M = { v2, v3, v4 and v5 }, wherein the number of the sets is n, the combination mode is M, and more than 2 features in the 4 features meet the requirement, namely, the features are labeled as a grate cooler snowman. The characteristic expression mode is that the secondary air temperature v2 gradually decreases in nearly 60 minutes, the concentration v3 of the gas NOx in the kiln tail smoke chamber gradually decreases, the tertiary air temperature v4 gradually increases, and the kiln current v5 is large in early-stage fluctuation but gradually decreases in the middle and later stages. The 4 feature sets in M are combined as follows: c _4^2+ C _4^3+ C _4^4+ =11, data samples of 11 conditions are marked as a grate cooler snowman, the rest data samples are marked as normal, and the training data samples are not less than 14000 groups;
step S05: and constructing an ANN artificial neural network structure, wherein the neuron structure of the ANN artificial neural network structure consists of an input layer, a first hidden layer, a second hidden layer, a third hidden layer and an output layer. The 9 feature sets determined in step S01 are used as parameters of the input layer, that is, the number of neurons of the input layer is defined as 9; the first hidden layer consists of 18 neurons; the second hidden layer y2 consists of 10 neurons; the third hidden layer y3 consists of 4 neurons; the output layer consists of 1 neuron;
step S06: training the ANN artificial neural network constructed in the step S05, wherein the training step specifically comprises the following steps:
s06.1, acquiring data input from the training set, and adjusting according to the weight;
s06.2, each layer uses a PReLU activation function, and each neuron is controlled to output through nonlinearity according to a formula f (x) = max (0, x), so that the generalization capability of the model is improved;
s06.3, calculating a reverse propagation error rate c, obtaining a prediction output value h of each neuron, wherein the expected value of the training set is n, and the reverse propagation error rate c = h-n;
s06.4, adjusting by an error weighted derivative formula according to the obtained error range;
s06.5, the training step length is 15, the training precision is 15-8, the steps S06.1-S06.3 are repeated for no less than 20000 times until the training is finished, and the output value of the model approaches to the expected value of the training set;
step S07: acquiring the relevant parameter data of the grate cooler working condition in the cement production process in real time, wherein the data comprises the characteristic set in the characteristic model defined in the step S1: an air chamber v1 of a grate cooler F2, a secondary air temperature v2, a kiln tail smoke chamber gas NOx concentration v3, a tertiary air temperature v4, kiln current v5, a kiln door cover differential pressure v6, a kiln tail smoke chamber gas analyzer CO concentration v7, an electric dust collection inlet temperature II v8 and a clinker temperature v9 of a grate cooler. According to the data rule engine constructed in step S2, the following steps are completed:
s07.1, cleaning the collected data according to the standard range of the data defined by the data rule engine;
s07.2 initially calculating a mean value of the current time and 24 times in total 5 minutes, 10 minutes, 15 minutes, 20 minutes, 25 minutes, 30 minutes, 35 minutes, 40 minutes, 45 minutes, 50 minutes, 55 minutes, 60 minutes, 65 minutes, 70 minutes, 75 minutes, 80 minutes, 85 minutes, 90 minutes, 95 minutes, 100 minutes, 105 minutes, 110 minutes, 115 minutes before the current time in terms of the data type of the mean value;
s07.3, assembling the mean value of S07.2 at each moment into parameters, and storing the parameters in a table form;
step S08: taking out the parameters assembled in the step S07, and inputting the trained grate cooler snowman recognition model according to the algorithm running frequency defined in the step S02;
step S09: and obtaining a model output value from the identification model of the grate cooler snowman for identification, wherein the output value is 0 or 1, 0 represents the grate cooler snowman, and 1 represents the normal working condition of the grate cooler.
Claims (8)
1. A grate cooler snowman recognition system based on an ANN neural network is characterized by comprising a main device, a feature set definition device, a training set device, an ANN artificial neural network training device, a data rule engine, a data cleaning device and a grate cooler snowman recognition device, wherein the feature set definition device, the training set device, the ANN artificial neural network training device, the data rule engine, the data cleaning device and the grate cooler snowman recognition device are arranged on the main device;
the characteristic set definition device defines factors influencing the grate cooler snowman, determines the characteristic set of the grate cooler snowman, and inputs the characteristic set as an input variable of the ANN artificial neural network to the training set device;
the training set device collects the off-line business historical data of the kiln condition in the production process according to the feature set, covers the parameters defined in the feature set, cleans the off-line business historical data, completes data labeling and inputs the data to the ANN neural network model training device;
the ANN neural network model training device is used for performing semi-supervised learning on a large amount of input grate cooler snowman offline service historical data;
the data rule engine device records the rule characteristics in the device in a form of a table;
the data cleaning device acquires relevant characteristic parameter data of the grate cooler snowman in the production process in real time, and performs data cleaning and assembling on the characteristic parameter data according to rules preset by a data rule engine;
the grate cooler snowman identification device identifies and classifies real-time data related to grate cooler input in the production process, and the classification comprises two normal categories of grate cooler snowman accumulation and grate cooler.
2. The ANN neural network-based grate cooler snowman identification system is characterized in that the data labeling labels the data samples meeting the snowman standard as snowmen, and the data samples not meeting the standard are normal.
3. The ANN neural network-based grate cooler snowman recognition system is characterized in that a neural network model training device is provided with a neural structure, the neural structure comprises an input layer, a first hidden layer, a second hidden layer, a third hidden layer and an output layer, and the semi-supervised learning means that given a training set of the snowman expected to be output, the actual output value of the ANN neural network model device approaches to the expected output by continuously adjusting the weight of each neuron in each hidden layer.
4. The ANN neural network-based grate cooler snowman identification system of claim 1, wherein the regular characteristics comprise: the method comprises the following steps of data acquisition frequency, data acquisition range, data standard range, data difference value threshold, acquisition type and algorithm operation frequency, wherein the acquisition type is a mean value or interpolation.
5. The ANN neural network-based grate cooler snowman recognition system as claimed in claim 1, wherein when the difference value of the relevant characteristic parameter data and the preset standard exceeds a preset deviation threshold value, the data is removed and recorded in a dirty data recording table in a table form.
6. An identification method of the ANN neural network based grate cooler snowman identification system of claims 1-5, comprising the following steps:
step S01: analyzing the performance characteristics of the formed grate cooler snowman, and determining a characteristic model of the grate cooler snowman;
step S02: analyzing the data rule characteristics formed by the grate cooler snowman and constructing a data rule engine according to the data rule characteristics, wherein the rule characteristics of the rule engine are stored in a form of a table;
step S03: acquiring offline service data according to the characteristic parameters of the characteristic model, cleaning the offline service data according to the characteristic range of the data, removing data outside the characteristic range of the data, and maintaining an offline service data record table by the identification system;
step S04: marking the off-line service data cleaned in the step S03 according to the characteristic model, wherein the marking is carried out in a general combination mode;
step S05: constructing an ANN artificial neural network model;
step S06: training an ANN artificial neural network model;
step S07: acquiring relevant parameter data of the working condition of the grate cooler in the cement production process in real time, cleaning and assembling according to a data standard range defined by a data rule engine, and then storing in a table form;
step S08: taking out the relevant parameters of the assembled grate cooler under the working condition, and inputting the trained ANN artificial neural network model according to the defined algorithm operation frequency;
step S09: and obtaining a model output value from the recognition model of the grate cooler snowman for recognition.
7. The grate cooler snowman identification method based on the ANN neural network as claimed in claim 6, wherein the step S06 specifically comprises:
s6.1, acquiring data input from the training device, and adjusting according to the weight;
s6.2, each layer uses a PReLU activation function, formula f (x) = max (0, x), and each neuron is controlled to output through nonlinearity, so that the generalization capability of the model is improved;
s6.3, calculating a reverse propagation error rate c, obtaining a prediction output value h of each neuron, wherein the expected value of a training set is n, and the reverse propagation error rate c = h-n;
s6.4, adjusting through an error weighted derivative formula according to the obtained error range;
s6.5, the training step length is 15, the training precision is 15-8, the steps S6.1-S6.3 are repeated for not less than 20000 times until the training is finished, and the output value of the model approaches to the expected value of the training set.
8. The grate cooler snowman identification method based on the ANN neural network as claimed in claim 6, wherein the step S07 specifically comprises:
s7.1, cleaning the acquired data according to a data standard range defined by a data rule engine;
s7.2, calculating the average value of 24 moments at the current moment and every 5 minutes before the current moment by the data type of the average value initially;
s7.3, the mean value of each moment of S7.2 is assembled into parameters and stored in a table form.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011201201.XA CN112365458A (en) | 2020-11-02 | 2020-11-02 | Grate cooler snowman identification method and system based on ANN neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011201201.XA CN112365458A (en) | 2020-11-02 | 2020-11-02 | Grate cooler snowman identification method and system based on ANN neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112365458A true CN112365458A (en) | 2021-02-12 |
Family
ID=74513256
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011201201.XA Pending CN112365458A (en) | 2020-11-02 | 2020-11-02 | Grate cooler snowman identification method and system based on ANN neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112365458A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2008031177A1 (en) * | 2006-09-11 | 2008-03-20 | Gerdau Açominas S/A | Process using artificial neural network for predictive control in sinter machine |
| US20140365195A1 (en) * | 2013-06-07 | 2014-12-11 | Scientific Design Company, Inc. | System and method for monitoring a process |
| CN107146004A (en) * | 2017-04-20 | 2017-09-08 | 浙江大学 | A kind of slag milling system health status identifying system and method based on data mining |
| CN109933881A (en) * | 2019-03-06 | 2019-06-25 | 武汉大学 | A kind of Fault Diagnosis of Power Electronic Circuits method based on optimization deepness belief network |
| CN111158336A (en) * | 2019-11-22 | 2020-05-15 | 万洲电气股份有限公司 | Industrial intelligent optimization energy-saving system based on cement kiln fault diagnosis |
-
2020
- 2020-11-02 CN CN202011201201.XA patent/CN112365458A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2008031177A1 (en) * | 2006-09-11 | 2008-03-20 | Gerdau Açominas S/A | Process using artificial neural network for predictive control in sinter machine |
| US20140365195A1 (en) * | 2013-06-07 | 2014-12-11 | Scientific Design Company, Inc. | System and method for monitoring a process |
| CN107146004A (en) * | 2017-04-20 | 2017-09-08 | 浙江大学 | A kind of slag milling system health status identifying system and method based on data mining |
| CN109933881A (en) * | 2019-03-06 | 2019-06-25 | 武汉大学 | A kind of Fault Diagnosis of Power Electronic Circuits method based on optimization deepness belief network |
| CN111158336A (en) * | 2019-11-22 | 2020-05-15 | 万洲电气股份有限公司 | Industrial intelligent optimization energy-saving system based on cement kiln fault diagnosis |
Non-Patent Citations (4)
| Title |
|---|
| 李勇等: "《复杂情感分析方法及其应用》", 30 April 2020, 冶金工业出版社 * |
| 李英等: "基于数据挖掘的电网系统故障智能诊断系统研究", 《自动化与仪器仪表》 * |
| 范佳卿: "基于实时特征值的风机振动状态监测与数据挖掘的故障诊断研究", 《中国优秀博硕士学位论文全文数据库(硕士) 工程科技Ⅱ辑》 * |
| 谷雨等: "水泥回转窑故障诊断方法研究", 《自动化仪表》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110502991B (en) | Internal combustion engine health monitoring method and system based on random convolutional neural network structure | |
| WO2021184630A1 (en) | Method for locating pollutant discharge object on basis of knowledge graph, and related device | |
| CN110285969B (en) | Rolling bearing fault migration diagnosis method with polynomial nuclear implantation characteristic distribution adaptation | |
| Lieber et al. | Quality prediction in interlinked manufacturing processes based on supervised & unsupervised machine learning | |
| CN109343344A (en) | Cigarette machine operating parameter optimization method | |
| CN104331521A (en) | Automatic power transformation equipment abnormity identification method based on image processing | |
| CN107977686B (en) | Industrial X-ray image classification method | |
| CN108304567B (en) | Method and system for identifying working condition mode and classifying data of high-voltage transformer | |
| CN112231971B (en) | Blast furnace fault diagnosis method based on relative integral trend diffusion fault sample generation | |
| CN116700172A (en) | Industrial data integrated processing method and system combined with industrial Internet | |
| CN116468160A (en) | Aluminum alloy die casting quality prediction method based on production big data | |
| CN116522156A (en) | Equipment state data analysis system and method based on energy management platform | |
| CN116340746A (en) | Feature selection method based on random forest improvement | |
| CN110728377B (en) | Intelligent fault diagnosis method and system for electromechanical equipment | |
| CN112365458A (en) | Grate cooler snowman identification method and system based on ANN neural network | |
| CN111783544A (en) | Method for building diamond milling head state monitoring system for machining ceramic mobile phone back plate | |
| CN113077423B (en) | Laser selective melting pool image analysis system based on convolutional neural network | |
| CN110852373A (en) | Defect-free sample deep learning network training method based on vision | |
| CN110826600B (en) | Engine surge prediction method based on adaptive resonance network online incremental learning | |
| CN117034197A (en) | Enterprise power consumption typical mode analysis method based on multidimensional Isolate-detection multi-point detection | |
| CN115146675B (en) | Rotary machine migration diagnosis method under variable working condition of depth multi-feature dynamic countermeasure | |
| CN113505639B (en) | Rotary machine multi-parameter health state assessment method based on TPE-XGBoost | |
| CN113963291A (en) | Coal-fired working condition identification method based on video image quality and model modeling method thereof | |
| CN113435794A (en) | Intelligent monitoring method for post-processing of nozzle casting based on image processing | |
| Lööv | Comparison of undersampling methods for prediction of casting defects based on process parameters |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210212 |