CN110096041A - Filter consistency method of quality control based on resonator assembly technology - Google Patents
Filter consistency method of quality control based on resonator assembly technology Download PDFInfo
- Publication number
- CN110096041A CN110096041A CN201910376043.2A CN201910376043A CN110096041A CN 110096041 A CN110096041 A CN 110096041A CN 201910376043 A CN201910376043 A CN 201910376043A CN 110096041 A CN110096041 A CN 110096041A
- Authority
- CN
- China
- Prior art keywords
- resonator
- filter
- standard
- parameter
- quality
- 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
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000005516 engineering process Methods 0.000 title claims abstract description 10
- 238000003908 quality control method Methods 0.000 title claims abstract description 10
- 230000008569 process Effects 0.000 claims abstract description 30
- 238000012545 processing Methods 0.000 claims abstract description 18
- 238000012360 testing method Methods 0.000 claims abstract description 9
- 238000013528 artificial neural network Methods 0.000 claims description 11
- 230000001537 neural effect Effects 0.000 claims description 11
- 238000012549 training Methods 0.000 claims description 5
- 230000004913 activation Effects 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000009467 reduction Effects 0.000 abstract description 5
- 238000004458 analytical method Methods 0.000 abstract description 4
- 230000006870 function Effects 0.000 description 10
- 238000010586 diagram Methods 0.000 description 3
- 241001269238 Data Species 0.000 description 2
- 230000006399 behavior Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 239000013078 crystal Substances 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000011056 performance test Methods 0.000 description 2
- 210000004556 brain Anatomy 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000003780 insertion Methods 0.000 description 1
- 230000037431 insertion Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000000513 principal component analysis Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 230000000153 supplemental effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
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
- G05B19/418—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM]
- G05B19/41875—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM] characterised by quality surveillance of production
-
- 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
- G05B2219/32—Operator till task planning
- G05B2219/32368—Quality control
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Landscapes
- Engineering & Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Manufacturing & Machinery (AREA)
- Quality & Reliability (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Measurement Of Resistance Or Impedance (AREA)
- Control Of Motors That Do Not Use Commutators (AREA)
Abstract
A kind of the step of filter consistency method of quality control use based on resonator assembly technology includes: step 1: test resonator parameter, including frequency, resistance, direct capacitance, dynamic capacity and inductance；Step 2: using processing quality consistency model predictive filter quality, up to standard then to enter in next step, below standard, replaces resonator, repeats step 2；Step 3: assembly resonator；Step 4: test filter performance, it is up to standard then to enter subsequent processing, then return step two not up to standard.The present invention can do the comprehensive analysis of multiple parameters to resonator, be attributed to a few multi-stress by variable of the dimensionality reduction to multiple intricate relationships；The performance data that preceding look-ahead resonator is assembled into filter can be assembled in process；Can before collocation look-ahead arrange in pairs or groups result and fed back.
Description
Technical field
The present invention relates to filter production technical fields, and in particular to a kind of filter one based on resonator assembly technology
Cause property amount control method.
Background technique
Resonator assembly process assembly duration is longer, influences on product quality more.Resonator assembly process is to work at present
The means of skill quality control are the methods inferred by experimental result, substantially with the following method:
1. combination by hand: by resonator is manually assembled into filter, the filter installed being passed through Network Analyzer
The test for carrying out performance, according to the performance image of Self -adaptive come random replacement crystal.
2. the experience of dependence: after labor mix's junction filter, by empiric observation Network Analyzer image waveform
Variation, by repeatedly practicing, finding single crystal may be to the influence of Network Analyzer waveform, and selective exchange is individually brilliant
Body makes the performance of filter reach technique requirement.
3. factorial analysis: since resonator parameter factors are excessive, being judged by the correlation to a few parameters, can be lost
Many information are lost, it is difficult to ensure that the influence size of other factors.
The performance test of filter can not directly by the spectral discrimination of Network Analyzer, which resonator causes four
Index is unqualified, can only be speculated by experience, is not necessarily most suitable matched combined, therefore the height of properties of product is different,
Consistency is poor.The experience formed at this stage is the waveform that 1, No. 2 resonator influences left side, and 5, No. 6 resonators influence right side
Waveform, more dependence artificial experience fail to form effective pass on knowledge.Assembly process assembly time is longer, to product quality
It influences more.Average each filter was assemblied between 20~40 minutes.After being gradually assembled into filter with resonator, most
The resonator for having some residual eventually can not arrange in pairs or groups.
Summary of the invention
In view of the deficiencies of the prior art, the present invention proposes a kind of mass data using assembly process history, in conjunction with machine
Learning algorithm constructs resonator assembly process quality conformance Controlling model, guarantees product quality consistency, reduces processing quality
Deviation, advanced optimize technique, improve the assembly efficiency of assembly process, reduce the filter based on resonator assembly technology of duration
Wave device consistency method of quality control.Specific technical solution is as follows:
A kind of filter consistency method of quality control based on resonator assembly technology, specific steps are as follows:
Step 1: test resonator parameter, including frequency, resistance, direct capacitance, dynamic capacity and inductance；
Step 2: use processing quality consistency model predictive filter quality, it is up to standard then enter in next step, it is below standard then
Resonator is replaced, step 2 is repeated；
Step 3: assembly resonator；
Step 4: test filter performance, it is up to standard then to enter subsequent processing, then return step two not up to standard.
To better implement the present invention, further are as follows:
The step 2 specifically:
2.1 collect the historical data in resonator assembling process, according to the different judging characteristic parameter weights of correlation size
The property wanted, gives characteristic parameter respective weights, and selectivity rejects feature of the correlation less than 0.3；
2.2 pairs of characteristic parameters carry out dimension-reduction treatment；
2.3 pairs of characteristic parameters do normalized；
2.4 using the control parameter of assembly process as training sample, and carries out cluster centralized processing to data；
2.5 predict the process results parameter of resonator assembly process using radial base neural net.
The activation primitive of the radial base neural net indicates are as follows:
Therefore the formula of radial base neural net can be calculated are as follows:
Wherein ciFor Basis Function Center point, σ is the width parameter of function, wijFor the weight of neural network, x is neural network
Input, yjFor the output of neural network；
There are three the parameters for needing to solve: Basis Function Center, the weight of width (variance) and hidden layer to output layer:
Basis Function Center can be clustered by k-means and choose h central point；
Width is by equations:cmaxFor the maximum distance between selected central point；
Connection weight asks local derviation to be directly calculated loss function with least square method, calculation formula are as follows:
The invention has the benefit that 1, the comprehensive analysis of multiple parameters can be done to resonator, by dimensionality reduction to multiple
The variable of intricate relationship is attributed to a few multi-stress；It 2, can the look-ahead resonator assembly before process is assembled
At the performance data of filter；Can before collocation look-ahead arrange in pairs or groups result and fed back；It 3, can be according to assembler
The filter results data of sequence quality conformance model prediction find suitable resonator combination；4, to Field Force's know-how
It is required that it is low, without too many artificial experience, form effective pass on knowledge；5, it arranges in pairs or groups, assembles according to prediction matched combined
The average each filter assembly of process needs 2-3 minutes, improves 10 times of working efficiency.
Detailed description of the invention
Fig. 1 is process schematic diagram in the present invention；
Fig. 2 is the method flow diagram in the present invention；
Fig. 3 is the schematic diagram of radial base neural net in the present invention.
Specific embodiment
The preferred embodiments of the present invention will be described in detail with reference to the accompanying drawing, so that advantages and features of the invention energy
It is easier to be readily appreciated by one skilled in the art, so as to make a clearer definition of the protection scope of the present invention.
Explanation of nouns:
ANN: the network for the extensive parallel interconnection that artificial neural network is made of adaptable simple unit, by more
A neuron models connect and compose according to certain rules, are that a kind of neural network for simulating human brain is artificial to can be realized class
The machine learning techniques of intelligence.
Dimensionality reduction: dimensionality reduction refers to using certain mapping method, by the sky of the Mapping of data points in former higher dimensional space to low dimensional
Between in.
Cluster: cluster is a task for being grouped object, and similar object is made to be classified as one kind, dissimilar object
It is classified as inhomogeneity.
It is as depicted in figs. 1 and 2: a kind of filter consistency method of quality control based on resonator assembly technology, specifically
Step are as follows:
Step 1: test resonator parameter, including frequency, resistance, direct capacitance, dynamic capacity and inductance；
Step 2: use processing quality consistency model predictive filter quality, it is up to standard then enter in next step, it is below standard then
Resonator is replaced, step 2 is repeated；
2.1 collect the history resonator behavior parameter of resonator assembly process, historical process result parameter, by what is be collected into
Data are stored in database with document form, and original sample number is arrived in storage after process historical data is converted to format data
According in library, the resonator behavior supplemental characteristic of assembly process includes room temperature data and high/low temperature data, room temperature data parameters packet
Frequency, resistance, F/T, R/T, direct capacitance, dynamic capacity, inductance, Q are included, high/low temperature data have a plurality of temperature data, and such as -45
It spends, -25 degree, 25 degree and other temperature, totally 8 temperature, parameter has ppm, resistance at each temperature, and technique requires to be that resonator fills
The performance data for the filter being made into, the input data item for compiling rear model is the performance room temperature and height of 12 resonators
Warm performance parameter gives characteristic parameter respective weights, selectivity according to the different judging characteristic parameter importance of correlation size
Reject feature of the correlation less than 0.3；
2.2 after the data processing of characteristic parameter, carries out the structure of the quality conformance Controlling model of resonator assembly process
It builds, dimensionality reduction is carried out to characteristic parameter according to principal component analysis, linear discriminant analysis before modeling, i.e. linear transformation will be original multiple
Characteristic variable is converted into several overall targets, thus simplified model structure；
2.3 due to different evaluation index there is different dimension and dimensional unit to need to eliminate dimension impact to dress
Control parameter with process does normalized, and normalized processing formula is as follows:
Wherein μ is the mean value of all sample datas, and σ is the standard deviation of all sample datas；
2.4 will be used as training sample after data preparation, excessively quasi- when training in order to prevent for more dispersed data
It closes, it is poor in the upper effect of verifying collection, it needs to carry out cluster centralized processing to data, the data with same characteristic features is gathered in one
Group, cluster centre depend on the circumstances, and determine sample at a distance from cluster centre by Euclidean distance, cluster formula is as follows:
Wherein y1It is cluster centre, dist is the Euclidean distance of sample distance center；Progressive alternate is needed in cluster process
Center could finally be converged to by calculating range estimation classification.The formula of every single-step iteration centering is as follows:
Wherein μiIt is cluster ciMean vector, otherwise referred to as mass center, ciIndicate the i-th class；
As shown in Figure 3: 2.5 predict the process results parameter of resonator assembly process using radial base neural net；It is described
The activation primitive of radial base neural net indicates are as follows:
Therefore the formula of radial base neural net can be calculated are as follows:
Wherein ciFor Basis Function Center point, σ is the width parameter of function, wijFor the weight of neural network, x is neural network
Input, yjFor the output of neural network；
There are three the parameters for needing to solve: Basis Function Center, the weight of width (variance) and hidden layer to output layer:
Basis Function Center can be clustered by k-means and choose h central point；
Width is by equations:cmaxFor the maximum distance between selected central point；
Connection weight asks local derviation to be directly calculated loss function with least square method, calculation formula are as follows:
In the present embodiment: using the performance parameter of 12 resonators as the parameter of input layer, 4 performances of filter being referred to
Mark: the output with outer clutter, insertion loss, stopband attenuation, passband fluctuation as output layer, in conjunction with radial base neural net
(RBFNN) algorithm establishes model for resonator assembly process problem, and total number of samples is according to totally 1000, wherein using 900 galley proofs
Notebook data carries out model training, remaining 100 sample data is tested, 264 nodes of input layer, 3653, middle layer sections
Point, 4 nodes of output layer, processing quality consistency control application model automatic acquisition resonator before assembly process is started to work
Performance test parameter passes through processing quality consistency model prediction result parameter；
Step 3: after prediction, carrying out the assembly of resonator, and the control of processing quality consistency is answered after assembly
It obtains process results data automatically with model, and by model predictive error module error in judgement, readjusts processing quality one
The parameter of cause property Controlling model, reaches the ability of self study；
Step 4: test filter performance, it is up to standard then to enter subsequent processing, then return step two not up to standard.
Claims (3)
1. a kind of filter consistency method of quality control based on resonator assembly technology, which is characterized in that the step of use
Include:
Step 1: test resonator parameter, including frequency, resistance, direct capacitance, dynamic capacity and inductance；
Step 2: using processing quality consistency model predictive filter quality, up to standard then to enter in next step, below standard, replaces
Resonator repeats step 2；
Step 3: assembly resonator；
Step 4: test filter performance, it is up to standard then to enter subsequent processing, then return step two not up to standard.
2. the filter consistency method of quality control based on resonator assembly technology, feature exist according to claim 1
In the step 2 specifically:
2.1 collect the historical data in resonator assembling process, important according to the different judging characteristic parameters of correlation size
Property, characteristic parameter respective weights are given, selectivity rejects feature of the correlation less than 0.3；
2.2 pairs of characteristic parameters carry out dimension-reduction treatment；
2.3 pairs of characteristic parameters do normalized；
2.4 using the control parameter of assembly process as training sample, and carries out cluster centralized processing to data；
2.5 predict the process results parameter of resonator assembly process using radial base neural net.
3. the filter consistency method of quality control based on resonator assembly technology, feature exist according to claim 2
In: the activation primitive of the radial base neural net indicates are as follows:
Therefore the formula that can calculate radial base neural net is
Wherein ciFor Basis Function Center point, σ is the width parameter of function, wijFor the weight of neural network, x is the defeated of neural network
Enter, yjFor the output of neural network；
There are three the parameters for needing to solve: Basis Function Center, the weight of width (variance) and hidden layer to output layer:
Basis Function Center can be clustered by k-means and choose h central point；
Width is by equations:cmaxFor the maximum distance between selected central point；
Connection weight asks local derviation to be directly calculated loss function with least square method, calculation formula are as follows:
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910376043.2A CN110096041A (en) | 2019-05-07 | 2019-05-07 | Filter consistency method of quality control based on resonator assembly technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910376043.2A CN110096041A (en) | 2019-05-07 | 2019-05-07 | Filter consistency method of quality control based on resonator assembly technology |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110096041A true CN110096041A (en) | 2019-08-06 |
Family
ID=67447000
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910376043.2A Pending CN110096041A (en) | 2019-05-07 | 2019-05-07 | Filter consistency method of quality control based on resonator assembly technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110096041A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115495995A (en) * | 2022-10-25 | 2022-12-20 | 深圳飞骧科技股份有限公司 | Simulation test fitting process parameter method, system, equipment and storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108415393A (en) * | 2018-04-19 | 2018-08-17 | 中江联合(北京)科技有限公司 | A kind of GaAs product quality consistency control method and system |
| CN108594776A (en) * | 2018-04-19 | 2018-09-28 | 中江联合(北京)科技有限公司 | A kind of GaAs quality conformance control method and system based on critical process |
-
2019
- 2019-05-07 CN CN201910376043.2A patent/CN110096041A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108415393A (en) * | 2018-04-19 | 2018-08-17 | 中江联合(北京)科技有限公司 | A kind of GaAs product quality consistency control method and system |
| CN108594776A (en) * | 2018-04-19 | 2018-09-28 | 中江联合(北京)科技有限公司 | A kind of GaAs quality conformance control method and system based on critical process |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115495995A (en) * | 2022-10-25 | 2022-12-20 | 深圳飞骧科技股份有限公司 | Simulation test fitting process parameter method, system, equipment and storage medium |
| CN115495995B (en) * | 2022-10-25 | 2023-07-07 | 深圳飞骧科技股份有限公司 | Method, system, equipment and storage medium for fitting process parameters through simulation test |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110674752B (en) | Tool wear state identification and prediction method based on hidden Markov model | |
| CN106021364B (en) | Foundation, image searching method and the device of picture searching dependency prediction model | |
| CN101634605B (en) | Intelligent gearbox fault diagnosis method based on mixed inference and neural network | |
| US7155423B1 (en) | Multi-criterial decision making system and method | |
| CN110595765A (en) | Wind turbine generator gearbox fault diagnosis method based on VMD and FA _ PNN | |
| CN104751842B (en) | The optimization method and system of deep neural network | |
| US20170330078A1 (en) | Method and system for automated model building | |
| CN109325691A (en) | Abnormal behaviour analysis method, electronic equipment and computer program product | |
| CN107122790A (en) | Non-intrusion type load recognizer based on hybrid neural networks and integrated study | |
| CN109142946A (en) | Transformer fault detection method based on ant group algorithm optimization random forest | |
| CN106628097A (en) | Ship equipment fault diagnosis method based on improved radial basis function neutral network | |
| CN106557816A (en) | Possesses the production equipment of rote learning device and assembling exerciser | |
| CN103577676A (en) | Grey weighting method for sewage treatment process comprehensive evaluation | |
| CN110166289A (en) | A kind of method and device identifying target information assets | |
| CN107169566A (en) | Dynamic neural network model training method and device | |
| Leauprasert et al. | Intelligent machine learning techniques for condition assessment of power transformers | |
| CN110096041A (en) | Filter consistency method of quality control based on resonator assembly technology | |
| CN106911512A (en) | Link Forecasting Methodology and system based on game in commutative figure | |
| CN108363738B (en) | Recommendation method for industrial equipment data analysis algorithm | |
| Bui et al. | An ant system algorithm for graph bisection | |
| CN115879412A (en) | Layout level circuit diagram size parameter optimization method based on transfer learning | |
| Jagielska | Linguistic rule extraction from neural networks for descriptive data mining | |
| Henri et al. | A deep transfer learning model for the identification of bird songs: A case study for Mauritius | |
| CN115471043A (en) | Wax mould casting heat treatment equipment health state evaluation method based on integrated SAE-SOM | |
| CN115169521A (en) | Graph neural network interpretation method for keeping prediction sequence and structure dependency relationship |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190806 |