CN108469783A - Deep hole deviation from circular from prediction technique based on Bayesian network - Google Patents
Deep hole deviation from circular from prediction technique based on Bayesian network Download PDFInfo
- Publication number
- CN108469783A CN108469783A CN201810467595.XA CN201810467595A CN108469783A CN 108469783 A CN108469783 A CN 108469783A CN 201810467595 A CN201810467595 A CN 201810467595A CN 108469783 A CN108469783 A CN 108469783A
- Authority
- CN
- China
- Prior art keywords
- deep hole
- node
- circular
- variable
- bayesian network
- 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
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
- G05B19/00—Programme-control systems
- G05B19/02—Programme-control systems electric
- G05B19/18—Numerical control [NC], i.e. automatically operating machines, in particular machine tools, e.g. in a manufacturing environment, so as to execute positioning, movement or co-ordinated operations by means of programme data in numerical form
- G05B19/404—Numerical control [NC], i.e. automatically operating machines, in particular machine tools, e.g. in a manufacturing environment, so as to execute positioning, movement or co-ordinated operations by means of programme data in numerical form characterised by control arrangements for compensation, e.g. for backlash, overshoot, tool offset, tool wear, temperature, machine construction errors, load, inertia
-
- 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
- G05B2219/30—Nc systems
- G05B2219/35—Nc in input of data, input till input file format
- G05B2219/35408—Calculate new position data from actual data to compensate for contour error
Landscapes
- Engineering & Computer Science (AREA)
- Human Computer Interaction (AREA)
- Manufacturing & Machinery (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Numerical Control (AREA)
Abstract
The deep hole deviation from circular from prediction technique based on Bayesian network that the invention discloses a kind of, the technical issues of for solving existing deep hole deviation from circular from prediction technique poor practicability.Technical solution is the mapping relations using Bayesian network structure deep hole machining parameter and deep hole deviation from circular from.Using deep hole machining parameter as the input node of Bayesian network model, using deep hole deviation from circular from as the class variable of output, simultaneously using Bayesian network using the axial force of cutter, torque and vibration performance during deep hole machining as the hidden node of network, network parameter is found out by the Bayesian Network Learning algorithm of missing value data, the validity that ensure that the imperfect situation drag of process data improves the applicability of different deep hole machining data.Meanwhile according to the inherent mechanism of action of each factor of deep hole machining, the causality of each network node is built, model tormulation is a kind of interpretable probabilistic model, and parameter optimization is processed using the model when predicting that error is unsatisfactory for tolerance.
Description
Technical field
The present invention relates to a kind of deep hole deviation from circular from prediction technique, more particularly to a kind of deep hole circle based on Bayesian network
Spend error prediction method.
Background technology
Document " the deep hole deviation from circular from prediction technique based on vibration cutting pattern feature, war industry's journal, 2018, Vol39
(2), p364-372 " discloses a kind of deep hole deviation from circular from prediction technique based on vibration cutting pattern feature.This method uses
Wavelet Packet Transform Method extracts the energy feature of vibration cutting during deep hole machining, and is that input is special with the energy feature of extraction
Sign, Fuzzy clustering techniques are introduced into normal linearity algorithm of support vector machine, vibration cutting feature and deep hole machining are constructed
Mapping relations between deviation from circular from so that the fuzzy input space partitioning problem of vibration cutting pattern is converted to initial input sky
Between initial-value problem, solve the problems, such as vibration cutting feature height overlapping.Meanwhile using the output error of identification model as target
Function reversely corrects space overlap coefficient, and realizing in the case where regular number is less still has preferable deviation from circular from prediction essence
Degree and generalization ability.For document the method by extracting the pre- sounding hole deviation from circular from of vibration cutting feature, characteristic variable is single, fits
It is not strong with property;Method needs to use sensor special, and cost is higher, is difficult to accurately obtain vibration cutting in practical deep hole machining
When information or imperfect information, missing value data can not be handled;In addition, this method does not account for deep hole machining parameter and vibration cutting
The relationship of feature, model interpretation is not strong, when predicting deep hole deviation from circular from and being unsatisfactory for tolerance, can not provide corresponding
Machining parameters optimization foundation.
Invention content
In order to overcome the shortcomings of that existing deep hole deviation from circular from prediction technique poor practicability, the present invention provide a kind of based on pattra leaves
The deep hole deviation from circular from prediction technique of this network.This method is missed using Bayesian network structure deep hole machining parameter with deep hole circularity
The mapping relations of difference.Include the speed of mainshaft, feed speed and drilling depth as Bayesian network model using deep hole machining parameter
Input node while missing value data can be handled using Bayesian network using deep hole deviation from circular from as the class variable of output
The characteristics of, the axial force of cutter, torque and vibration performance during deep hole machining are passed through into missing value as the hidden node of network
The Bayesian Network Learning algorithm of data finds out network parameter, ensure that the validity of the imperfect situation drag of process data,
To improve applicability of the deep hole deviation from circular from prediction technique for different deep hole machining data.Meanwhile it is each according to deep hole machining
The inherent mechanism of action of factor builds the causality of each network node, and model tormulation is a kind of interpretable probabilistic model, from
And parameter optimization is processed using the model when predicting that error is unsatisfactory for tolerance.
The technical solution adopted by the present invention to solve the technical problems:A kind of deep hole deviation from circular from based on Bayesian network
Prediction technique, its main feature is that including the following steps:
Step 1: determining the node variable of Bayesian network.The node set V of deep hole machining Bayesian network is expressed as
V={ X, H, C } (1)
In formula, X is the set of deep hole machining parametric variable, and H is the set of deep hole machining process variable, and C is deep hole circularity
Error variance.When carrying out node variable screening, the variable of had an impact deep hole deviation from circular from C is concluded first with expertise
Set X and H, determine variable factors complete or collected works, and delete constant value variable therein.Then remaining is acquired by grey relational grade analysis
With the degree of association sequence of deep hole deviation from circular from each variable, the degree of association sequence of deep hole deviation from circular from is deleted therein time by each variable
Variable is wanted, using remaining each variable factors as the node variable of deep hole machining Bayesian network.
Step 2: using the element in the deep hole machining parametric variable set X after screening as input node, deep hole machining mistake
Element in journey variables collection H is as hidden node, class nodes of the deep hole deviation from circular from C as output.On this basis to each change
The being associated property that influences each other between amount is analyzed, and judges to whether there is direct causality and determining causal side between each variable
To by each two, there are causal nodes to be connected with directed arc, constructs directed acyclic graph, obtains deep hole machining Bayesian network
The structure of network.
Step 3: determining the active region of each machined parameters in deep hole machining parametric variable set X according to deep hole machining data
Between, and using wide interval method to each parameter enliven section carry out it is discrete.For hidden in deep hole machining process variable set H
Node need to only give discrete segment number.Deviation from circular from node C is set as two value nodes, and it is qualified and not to be divided by practical tolerance
Qualified two classes.
Step 4: choosing Dirichlet is distributed prior probability distribution as each node of Bayesian network, each node it is general
Rate is distributed as
In formula, i is the serial number of each node in deep hole machining Bayesian network node set V, and j is the father node of each child node
Serial number, k be each node where discrete segment serial number, Pa (Vi) it is V in Bayesian networkiFather node, mijkFor training number
Meet variable V iniTake k-th of state value and Pa (Vi) take the example number of j-th of state value, αijkIt is distributed for Dirichlet
Equivalent samples amount.
Step 5: using the deep hole machining data under different deep hole machining parameter sets X as the training number of Bayesian network
According to, the Bayesian network parameters that missing value data are carried out using desired optimization algorithm are learnt, and the probability distribution of each node is calculated, it is expected that
The iterative formula of optimization algorithm probability calculation is
In formula,Meet node variable V in data after benefit it is expected each iteration step of optimization algorithmiTake k-th of state value
And Pa (Vi) take the sum of the broken power sample weights of j-th of state value, riFor node variable ViAll state value numbers.Work as iteration
When number reaches the maximum iteration of setting, it is expected that optimization algorithm iteration terminates, each node V is obtainediProbability distribution.
Step 6: according to the deep hole machining Bayesian network that study obtains, uncertainty is carried out using Junction tree and is pushed away
Reason.In practical deep hole machining, it is input with deep hole machining parameter sets X, carries out causal reasoning, it is full to calculate deep hole deviation from circular from
The probability P (C | X) of sufficient tolerance, realizes the prediction of deep hole deviation from circular from;If the probability that deviation from circular from meets tolerance is small
In 50%, then need to carry out diagnostic reasoning according to the deep hole machining Bayesian network that study obtains, calculating causes deviation from circular from discontented
The probability P (X | C) of each machined parameters of sufficient tolerance, and the machined parameters value of maximum probability is adjusted.
The beneficial effects of the invention are as follows:This method is using Bayesian network structure deep hole machining parameter and deep hole deviation from circular from
Mapping relations.Include the speed of mainshaft, feed speed and drilling depth as Bayesian network model using deep hole machining parameter
Input node using deep hole deviation from circular from as the class variable of output, while can handle missing value data using Bayesian network
Feature passes through missing value number using the axial force of cutter, torque and vibration performance during deep hole machining as the hidden node of network
According to Bayesian Network Learning algorithm find out network parameter, ensure that the validity of the imperfect situation drag of process data, from
And improve applicability of the deep hole deviation from circular from prediction technique for different deep hole machining data.Meanwhile according to deep hole machining it is each because
The inherent mechanism of action of element builds the causality of each network node, and model tormulation is a kind of interpretable probabilistic model, to
When predicting that error is unsatisfactory for tolerance parameter optimization is processed using the model.
It elaborates with reference to the accompanying drawings and detailed description to the present invention.
Description of the drawings
Fig. 1 is the flow chart of the deep hole deviation from circular from prediction technique the present invention is based on Bayesian network.
Fig. 2 is the deep hole machining bayesian network structure figure of the method for the present invention structure.
Specific implementation mode
Referring to Fig.1-2.The present invention is based on the deep hole deviation from circular from prediction techniques of Bayesian network to be as follows:
Step 1: determining the node of deep hole machining Bayesian network.
According to deep hole machining knowwhy, deep hole deviation from circular from formation mechenism is analyzed, determines Bayesian network on this basis
The node variable of network.The node set V of deep hole machining Bayesian network is expressed as
V={ X, H, C } (1)
In formula, X is the set of deep hole machining parametric variable, and H is the set of deep hole machining process variable, and C is deep hole circularity
Error variance.When carrying out node variable screening, the variable of had an impact deep hole deviation from circular from C is concluded first with expertise
Set X and H, determine variable factors complete or collected works, and delete constant value variable therein.Then remaining is acquired by grey relational grade analysis
With the degree of association sequence of deep hole deviation from circular from each variable, the degree of association sequence of deep hole deviation from circular from is deleted therein time by each variable
Variable is wanted, using remaining each variable factors as the node variable of deep hole machining Bayesian network.In given machine tool, drilling in this example
In the case of cutter, workpiece material and cooling condition, the node set V of the deep hole machining Bayesian network filtered out includes
X={ S, F, D }
H={ Thru, Tor, Vibra }
In formula, S, F, D indicate that the speed of mainshaft, feed speed and drilling depth, Thru, Tor, Vibra are indicated respectively respectively
Feed axial force, torque and drilling rod vibration.
Step 2: determining the structure of deep hole machining Bayesian network.
By the element S in the deep hole machining parametric variable set X after screening, F, D are as input node, deep hole machining process
Element T hru, Tor, Vibra in variables collection H is as hidden node, class nodes of the deep hole deviation from circular from set C as output.
The being associated property that influences each other between each variable on this basis is analyzed, and judges to whether there is direct causality simultaneously between each variable
Determine causal direction, there are causal nodes to be connected with directed arc by each two, constructs directed acyclic graph, obtains
The model structure of deep hole machining Bayesian network.
Step 3: determining the range and discrete segment of each node variable of deep hole machining Bayesian network.
Section is enlivened according to what deep hole machining data determined each machined parameters in deep hole machining parametric variable X, use is wide
Interval method to each parameter enliven section carry out it is discrete.Speed of mainshaft S is divided into 500~1500rpm in this example, 1500~
Tri- discrete segments of 2500rpm, 2500~3500rpm;Feed speed F is divided into 50~150mm/min, 150~250mm/
Tri- discrete segments of min, 250~350mm/min;Drilling depth D is divided into 20~40mm, 40~60mm, 60~80mm tri-
A discrete segment.For the hidden node in deep hole machining process variable set H, it is 3 to give discrete segment number, deviation from circular from node
C is set as two value nodes.The a diameter of φ 4mm of deep hole in this example, roundness tolerance is 5 μm, by this tolerance by deep hole each
The deviation from circular from of depth is divided into qualified and unqualified two class.
Step 4: determining the prior probability distribution of each node of deep hole machining Bayesian network.
Choose prior probability distribution of the Dirichlet distributions as each node of Bayesian network, the probability distribution of each node
For
In formula, i is the serial number of each node in deep hole machining Bayesian network node set V, and j is the father node of each child node
Serial number, k be each node where discrete segment serial number, Pa (Vi) it is V in Bayesian networkiFather node, mijkFor training number
Meet variable V iniTake k-th of state value and Pa (Vi) take the example number of j-th of state value, αijkIt is distributed for Dirichlet
Equivalent samples amount.
Step 5: the Bayesian network parameters for carrying out missing value data according to deep hole machining data learn.
Each deep hole machining parameter is divided into Three factors-levels by its respective discrete segment in this example, deep hole is carried out and adds
The total factor of work is tested, and using obtain 27 groups of data as the training data of Bayesian network, does not acquire deep hole machining in experiment
Axial force, torque and drilling rod vibration data are fed in the process.Using the Bayes for it is expected optimization (EM) algorithm progress missing value data
Network parameter learns, and calculates the probability distribution of each node, the iterative formula of EM algorithm probability calculations is
In formula,To meet node variable V in data after the benefit of each iteration step of EM algorithmsiTake k-th of state value and Pa
(Vi) take the sum of the sample weights of j-th of state value, riFor node variable ViAll state value numbers, deviation from circular from C in this example
State value be 2, the state value of remaining each node variable is 3.The maximum iterations of EM algorithms is set as 5, when iteration time
When number is up to maximum iteration, iteration terminates, and obtains each node ViProbability distribution.
Step 6: carrying out deep hole deviation from circular from prediction according to the Bayesian network after study.
According to the deep hole machining Bayesian network after study, uncertain inference is carried out using Junction tree, with deep hole
Machined parameters set X is input, calculates the probability that deep hole deviation from circular from meets tolerance, realizes the pre- of deep hole deviation from circular from
It surveys.It is obtained in speed of mainshaft S=2000rpm, feed speed F using the deep hole machining Bayesian Network Inference after study in this example
Under conditions of=300mm/min, drilling depth D=50mm, the probability that deep hole deviation from circular from meets tolerance is 33.6%, into
Row diagnostic reasoning, which obtains, causes the parameter of deviation from circular from maximum probability to be feed speed F, sets feed speed F to 200mm/
After min, it is 70.4% to recalculate deep hole deviation from circular to meet the probability of tolerance.It realizes and is carried out using Bayesian network
Deviation from circular from is predicted and the target of parameter optimization.
Claims (1)
1. a kind of deep hole deviation from circular from prediction technique based on Bayesian network, it is characterised in that include the following steps:
Step 1: determining the node variable of Bayesian network;The node set V of deep hole machining Bayesian network is expressed as
V={ X, H, C } (1)
In formula, X is the set of deep hole machining parametric variable, and H is the set of deep hole machining process variable, and C is deep hole deviation from circular from
Variable;When carrying out node variable screening, the variables collection of had an impact deep hole deviation from circular from C is concluded first with expertise
X and H determines variable factors complete or collected works, and deletes constant value variable therein;Then remaining each change is acquired by grey relational grade analysis
With the degree of association sequence of deep hole deviation from circular from amount, secondary change therein is deleted to the degree of association sequence of deep hole deviation from circular from by each variable
Amount, using remaining each variable factors as the node variable of deep hole machining Bayesian network;
Step 2: using the element in the deep hole machining parametric variable set X after screening as input node, deep hole machining process becomes
Element in duration set H is as hidden node, class nodes of the deep hole deviation from circular from C as output;On this basis between each variable
The being associated property that influences each other is analyzed, and judges to whether there is direct causality and determining causal direction between each variable,
By each two, there are causal nodes to be connected with directed arc, constructs directed acyclic graph, obtains deep hole machining Bayesian network
Structure;
Step 3: the section of enlivening of each machined parameters in deep hole machining parametric variable set X is determined according to deep hole machining data, and
Using wide interval method to each parameter enliven section carry out it is discrete;For the hidden node in deep hole machining process variable set H
Discrete segment number only need to be given;Deviation from circular from node C is set as two value nodes, and it is qualified and unqualified to be divided by practical tolerance
Two classes;
Step 4: choosing prior probability distribution of the Dirichlet distributions as each node of Bayesian network, the probability point of each node
Cloth is
In formula, i is the serial number of each node in deep hole machining Bayesian network node set V, and j is the sequence of the father node of each child node
Number, k is the discrete segment serial number where each node, Pa (Vi) it is V in Bayesian networkiFather node, mijkFor in training data
Meet variable ViTake k-th of state value and Pa (Vi) take the example number of j-th of state value, αijkFor Dirichlet distribution etc.
Valence sample size;
Step 5: using the deep hole machining data under different deep hole machining parameter sets X as the training data of Bayesian network, adopt
The Bayesian network parameters study that missing value data are carried out with desired optimization algorithm, calculates the probability distribution of each node, it is expected that optimizing
The iterative formula of algorithm probability calculation is
In formula,Meet node variable V in data after benefit it is expected each iteration step of optimization algorithmiTake k-th of state value and Pa
(Vi) take the sum of the broken power sample weights of j-th of state value, riFor node variable ViAll state value numbers;Work as iterations
When reaching the maximum iteration of setting, it is expected that optimization algorithm iteration terminates, each node V is obtainediProbability distribution;
Step 6: according to the deep hole machining Bayesian network that study obtains, uncertain inference is carried out using Junction tree;
In practical deep hole machining, it is input with deep hole machining parameter sets X, carries out causal reasoning, it is public calculates deep hole deviation from circular from satisfaction
The probability P (C | X) that difference requires, realizes the prediction of deep hole deviation from circular from;If the probability that deviation from circular from meets tolerance is less than
50%, then it needs to carry out diagnostic reasoning according to the deep hole machining Bayesian network that study obtains, calculating causes deviation from circular to be unsatisfactory for
The probability P (X | C) of each machined parameters of tolerance, and the machined parameters value of maximum probability is adjusted.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810467595.XA CN108469783B (en) | 2018-05-14 | 2018-05-14 | Deep hole roundness error prediction method based on Bayesian network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810467595.XA CN108469783B (en) | 2018-05-14 | 2018-05-14 | Deep hole roundness error prediction method based on Bayesian network |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108469783A true CN108469783A (en) | 2018-08-31 |
CN108469783B CN108469783B (en) | 2021-02-02 |
Family
ID=63261245
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810467595.XA Active CN108469783B (en) | 2018-05-14 | 2018-05-14 | Deep hole roundness error prediction method based on Bayesian network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108469783B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110567421A (en) * | 2019-09-05 | 2019-12-13 | 上海应用技术大学 | dynamic roundness uncertainty evaluation method based on Bayesian principle |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6553548B1 (en) * | 1999-12-14 | 2003-04-22 | International Business Machines Corporation | System and method for recovering from design errors in integrated circuits |
CN101436057A (en) * | 2008-12-18 | 2009-05-20 | 浙江大学 | Numerical control machining tool heat error Bayes network compensation method |
CN103514486A (en) * | 2012-06-15 | 2014-01-15 | 上海宝信软件股份有限公司 | Blast furnace gas incoming flow prediction method based on factor analysis |
CN104950808A (en) * | 2015-07-20 | 2015-09-30 | 攀枝花学院 | Machine tool thermal error compensation method based on augmented naive Bayes network |
CN105229633A (en) * | 2013-03-13 | 2016-01-06 | 萨勒斯福斯通讯有限公司 | For realizing system, method and apparatus disclosed in data upload, process and predicted query API |
WO2016091728A1 (en) * | 2014-12-10 | 2016-06-16 | The Provost, Fellows, Foundation Scholars, And The Other Members Of Board, Of The College Of The Holy And Undivided Trinity Of Queen Elizabeth, Near Dublin | Computer modelling of atmospheric re-entry of objects such as satellites |
CN106022954A (en) * | 2016-05-16 | 2016-10-12 | 四川大学 | Multiple BP neural network load prediction method based on grey correlation degree |
CN106217128A (en) * | 2016-07-06 | 2016-12-14 | 陕西柴油机重工有限公司 | Variable working condition bottom tool state of wear Forecasting Methodology based on big data mining |
CN107248003A (en) * | 2017-08-03 | 2017-10-13 | 浙江大学 | Based on the adaptive soft-sensor Forecasting Methodology with sliding window Bayesian network |
CN107991870A (en) * | 2017-12-05 | 2018-05-04 | 暨南大学 | A kind of fault pre-alarming and life-span prediction method of Escalator equipment |
-
2018
- 2018-05-14 CN CN201810467595.XA patent/CN108469783B/en active Active
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6553548B1 (en) * | 1999-12-14 | 2003-04-22 | International Business Machines Corporation | System and method for recovering from design errors in integrated circuits |
CN101436057A (en) * | 2008-12-18 | 2009-05-20 | 浙江大学 | Numerical control machining tool heat error Bayes network compensation method |
CN103514486A (en) * | 2012-06-15 | 2014-01-15 | 上海宝信软件股份有限公司 | Blast furnace gas incoming flow prediction method based on factor analysis |
CN105229633A (en) * | 2013-03-13 | 2016-01-06 | 萨勒斯福斯通讯有限公司 | For realizing system, method and apparatus disclosed in data upload, process and predicted query API |
WO2016091728A1 (en) * | 2014-12-10 | 2016-06-16 | The Provost, Fellows, Foundation Scholars, And The Other Members Of Board, Of The College Of The Holy And Undivided Trinity Of Queen Elizabeth, Near Dublin | Computer modelling of atmospheric re-entry of objects such as satellites |
CN104950808A (en) * | 2015-07-20 | 2015-09-30 | 攀枝花学院 | Machine tool thermal error compensation method based on augmented naive Bayes network |
CN106022954A (en) * | 2016-05-16 | 2016-10-12 | 四川大学 | Multiple BP neural network load prediction method based on grey correlation degree |
CN106217128A (en) * | 2016-07-06 | 2016-12-14 | 陕西柴油机重工有限公司 | Variable working condition bottom tool state of wear Forecasting Methodology based on big data mining |
CN107248003A (en) * | 2017-08-03 | 2017-10-13 | 浙江大学 | Based on the adaptive soft-sensor Forecasting Methodology with sliding window Bayesian network |
CN107991870A (en) * | 2017-12-05 | 2018-05-04 | 暨南大学 | A kind of fault pre-alarming and life-span prediction method of Escalator equipment |
Non-Patent Citations (3)
Title |
---|
李超 等: "基于刀具振动模式特征的深孔圆度误差预测方法", 《兵工学报》 * |
王明微 等: "基于贝叶斯网络的切削加工表面粗糙度在线监测方法", 《计算机集成制造系统》 * |
邢康林 等: "中小孔精密磨削监控反馈系统构建与应用", 《组合机床与自动化加工技术》 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110567421A (en) * | 2019-09-05 | 2019-12-13 | 上海应用技术大学 | dynamic roundness uncertainty evaluation method based on Bayesian principle |
Also Published As
Publication number | Publication date |
---|---|
CN108469783B (en) | 2021-02-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Bull et al. | Probabilistic active learning: an online framework for structural health monitoring | |
CN109829236A (en) | A kind of Compressor Fault Diagnosis method based on XGBoost feature extraction | |
CN101984340B (en) | Method for identifying flow type of soft grain two-phase turbulence based on artificial intelligence | |
CN108346293B (en) | Real-time traffic flow short-time prediction method | |
US20020038307A1 (en) | Systems and methods for knowledge discovery in spatial data | |
CN105760673A (en) | Fluvial facies reservoir earthquake sensitive parameter template analysis method | |
CN112529341A (en) | Drilling well leakage probability prediction method based on naive Bayesian algorithm | |
Pan et al. | UAV anomaly detection using active learning and improved S3VM model | |
CN114548482A (en) | Creep type landslide kinetic energy change rate face-slip early warning method | |
Li et al. | The method of reagent control based on time series distribution of bubble size in a gold‐antimony flotation process | |
CN108469783A (en) | Deep hole deviation from circular from prediction technique based on Bayesian network | |
CN112766321B (en) | Geological feature detection and recognition method and system based on deep learning | |
CN110163442A (en) | A kind of gas well plug-ging prediction technique based on integrated study | |
Verma et al. | Quantification of sand fraction from seismic attributes using Neuro-Fuzzy approach | |
CN117093922A (en) | Improved SVM-based complex fluid identification method for unbalanced sample oil reservoir | |
Pramoedyo et al. | Comparison of Random Forest and Naïve Bayes Methods for Classifying and Forecasting Soil Texture in the Area Around Das Kalikonto, East Java | |
Hossain et al. | A rough set based rule induction approach to geoscience data | |
US20240060419A1 (en) | Method and system for automated rock recognition | |
CN111985782B (en) | Automatic driving tramcar running risk assessment method based on environment awareness | |
CN114627333A (en) | Zinc flotation froth image classification algorithm and system for improving deep active learning | |
Syaharuddin et al. | Best Architecture Recommendations of ANN Backpropagation Based on Combination of Learning Rate, Momentum, and Number of Hidden Layers | |
CN109614584B (en) | Reconstruction method of resource logging data | |
CN102654865A (en) | Method and system for digital object classification | |
Wen | Construction project risk evaluation based on rough sets and artificial neural networks | |
CN116822971B (en) | Well wall risk level prediction method |
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 |