CN106840523A - The synergetic immunity detection method of leakage in a kind of tyre vulcanizer drain valve steam - Google Patents
The synergetic immunity detection method of leakage in a kind of tyre vulcanizer drain valve steam Download PDFInfo
- Publication number
- CN106840523A CN106840523A CN201610965184.4A CN201610965184A CN106840523A CN 106840523 A CN106840523 A CN 106840523A CN 201610965184 A CN201610965184 A CN 201610965184A CN 106840523 A CN106840523 A CN 106840523A
- Authority
- CN
- China
- Prior art keywords
- vulcanizer
- steam
- antibody
- drain valve
- detection method
- 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
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M3/00—Investigating fluid-tightness of structures
Abstract
A kind of synergetic immunity detection method of leakage in tyre vulcanizer drain valve steam, computer based manual system comprises the following steps:A:The initialization of calculating main frame system:(1) by the Dynamic Baseline regression model set up between workshop level steam consumption and device level technological parameter, dynamic risk baseline is produced, sets danger threshold;(2) vulcanizer state parameter and jet chimney state parameter are utilized, the detector for producing leakage vulcanizer is clustered by artificial immune network;B:The synergetic immunity detection method of computer each run:The Dynamic Baseline regression model and the detector are carried out into synergetic immunity detection, synergetic immunity detection method proposed by the present invention can exactly detect vulcanizer with the presence or absence of leakage event in drain valve steam, loss and false drop rate are controlled 2.62% and 5.26% respectively, compared with traditional device level state parameter detection method, loss and false drop rate reduce by 6.67% and 14.28% respectively.
Description
Technical field
The present invention relates to drain valve steam inner hexagon technical field, more particularly to a kind of tyre vulcanizer drain valve steam
The synergetic immunity detection method of interior leakage.
Background technology
In tyre vulcanization workshop, green tire is fixed to regulation shape, high intensity, elastomeric finished product at high temperature under high pressure
Tire, steam is the most frequently used thermal source and pressure medium in vulcanization process, and it is high that substantial amounts of steam consumption result in vulcanization process
Energy consumption cost and environmental pollution cost;And leakage (hereinafter referred to as drain valve in leakage) is common in curing department in drain valve steam
Anomalous event, usually trigger substantial amounts of waste of steam, reduce the energy efficiency of vulcanization;One vulcanization that there is leakage in drain valve
More than 3 times when steam consumption in the machine course of work is normal, the steam consumption of leakage is total often close to curing department's steam
The 15% of consumption.Obviously, find that leakage is beneficial to reduce production cost, potential safety hazard and pollutant emission in drain valve in time.
Modern tire vulcanization process is automatically performed by vulcanizer, the structural principle of vulcanizer as shown in figure 4, during vulcanization,
One green tire 1 is firstly fixed between the vaporium 2 of vulcanizer and capsule 3, and then high pressure supersaturated vapor is filled with steaming respectively
Steam chest and capsule, heat green tire and induce vulcanization reaction.After a period of time, high pressure nitrogen is filled with capsule, there is provided sizing pressure.Dredge
Water valve 4 connects jet chimney and vaporium, the condensed water for being produced in discharge sulfidation in time.But leakage causes steam in drain valve
Discharged in the state of uncooled, so as to cause energy waste.Due to the limitation of technology and cost, few separate unit vulcanizers are matched somebody with somebody
Standby vaporimeter scale, it is difficult to passing through simple observation finds failure.In actual production, workshop generally relies on regularly manual detection
(dismounting or leak detector) finds failure, but time-consuming, and efficiency is low.Leakage does not influence production under mild state in drain valve, perhaps
Multiple enterprises are even without the seriousness for recognizing to be leaked in drain valve.And leakage in drain valve is included pipeline and let out by part researcher suggestion
The scope of leakage, and detected by analyzing the state parameters such as the temperature around potential leakage point, pressure, flow velocity, vibrations and sound wave
Method finds failure, although can in a short time find leakage failure, and relatively accurately positions leakage point, but state parameter
Detection method is only more effective to simple shape, the transmission pipeline system of flow speed stability, is difficult to apply to tube topology complexity, stream
The unstable curing department of speed.
For leakage event in drain valve, the energy consumption measure that being recognized using energy efficiency evaluation needs device level is supported, and
The metering that this exactly most Tire production enterprise is lacked, although existing research is that leakage is asked in vulcanizer drain valve steam
Topic has established theoretical foundation, but develops a kind of effective detection method for the problem and also need to overcome three below difficult point:
Limitation, the accuracy of energy efficiency evaluation, the disturbance factor of vulcanizer temperature and pressure of device level metering;I.e. (1) is because of cost and work
Skill, the steam measurement of device level is had difficulties in curing department, and the resistance vapour effect of particularly steam-flow meter can be influenceed to vulcanization
The precise control of temperature and pressure, usual vulcanizer is not equipped with steam measurement, is only equipped with the temperature and pressure of vaporium and capsule
Sensor, not possessing the change of dependence separate unit vulcanizer steam consumption carries out the condition of leak detection;(2) vulcanization efficiency is produced
The influence of the factors such as speed, technological parameter, environment temperature, the threshold test failure is simply provided to workshop efficiency, can trigger height
False drop rate and loss, and leakage vulcanizer cannot be positioned;(3) disturbance of vapour source pressure and steam flow rate, can cover hydrophobic
The change of vapor (steam) temperature and pressure that leakage triggers in valve, the independent change with vulcanizer temperature and pressure goes to detect that the failure is same
Sample can produce false drop rate and loss higher.Therefore, under existing metered condition, simply by technological parameter and energy efficiency evaluation
The effective detection of leakage in drain valve cannot be realized.
The content of the invention
It is an object of the invention to propose a kind of tire for being combined workshop level steam consumption and device level technological parameter
The synergetic immunity detection method of leakage in vulcanizer drain valve steam, its loss and false drop rate are substantially reduced, can effectively controlled
Loss and false drop rate are respectively 2.62% and 5.26%.
It is that, up to this purpose, the present invention uses following technical scheme:
The synergetic immunity detection method of leakage in a kind of tyre vulcanizer drain valve steam, computer based manual system,
Comprise the following steps:
A:The initialization of calculating main frame system:
(1) structure of danger threshold model:It is dynamic by what is set up between workshop level steam consumption and device level technological parameter
State baseline regression model, produces dynamic risk baseline, sets danger threshold;
(2) structure of detection model:Using vulcanizer state parameter and jet chimney state parameter, by artificial immunity net
Network cluster produces the detector of leakage vulcanizer;
B:The synergetic immunity detection method of computer each run:
The structure of synergetic immunity detection model:The Dynamic Baseline regression model and the detector are carried out into synergetic immunity
Detection.
Further illustrate, the synergetic immunity detection is event will to be leaked in the vulcanizer drain valve as antigen, described
The external pressure of the vulcanizer in device level technological parameter and outer temperature (pVOAnd tVO) used as epitope, the workshop level steam is used
Amount (M) is danger signal source;
Periodic statistics and detection are carried out to the workshop level steam consumption (M), when the workshop level steam consumption (M) is super
Go out danger threshold scope, then send distress signal, trigger the external pressure and outer temperature (p of the vulcanizerVOAnd tVO) with the detector
Matching, if the match is successful, confirm failure occur, triggering alarm;The external pressure of wherein described vulcanizer and outer temperature (pVOWith
tVO) multiple repairing weld detection can be carried out.
Further illustrate, the foundation of the Dynamic Baseline regression model include the foundation of the thermal balance model for vulcanizing energy consumption and
The danger threshold regression analysis of workshop steam consumption.
Further illustrate, the thermal balance model of the vulcanization energy consumption is formulated as:Wherein QVIt is tire sulphur in the cycle
Change the heat of consumption;QViIt is the total amount of heat of vulcanization process i consumption;KHIt is the specific heat of green tire in vulcanization process, mTRiGreen tire quality,
ΔtOiIt is tOAnd tAiDifference, tOAnd tAiRespectively curing temperature and environment temperature; KSIt is vulcanizer surface coefficient of heat transfer, AiSulphur
Change machine steam chamber surface is accumulated, △ tSiIt is tSAnd tAiDifference, tSIt is vulcanizer steam chamber surface temperature, tAiIt is environment temperature, τiFor
The vulcanizer available machine time in curing cycle;KIIt is the coefficient after merger;DTRjAnd BTRiRespectively tire outside diameter and width.
Further illustrate, the danger threshold regression analysis of the workshop steam consumption is many based on the thermal balance model
First linear regression prediction assesses the reasonable interval of workshop steam consumption (M), and the interval upper limit is danger threshold, when the workshop of actual measurement
Steam consumption (M) exceeds danger threshold, then send the danger signal of leakage in drain valve.
Further illustrate, the artificial immune network cluster is to leak feature according in vulcanization operating mode feature and drain valve, is carried
Take the Inner temperature t of vulcanizer state parameterVI, internal pressure pVI, and its outer temperature tVO, external pressure pVOWith the external pressure of jet chimney state parameter
Vapor (steam) temperature tO, pressure pODifference, then cluster sample characteristics S be formulated as:S=<pVI,tVI,△pVO,△tVO>, △
pVO=pVO-pO, △ tVO=tVO-tO。
Further illustrate, the artificial immune network cluster includes immune compression and immune cluster;The immune compression mould
Intend the Immune Clone Selection of Immune System and immune taboo, if training sample is SPL, aiNet of the input based on Immune Clone Selection mechanism
Memory network (M is exported after algorithmD), the immune compression includes following algorithm steps:
(1) initiation parameter n, ξ, σs,σd,σf, randomly generate an Antibody Network A;
(2) following iteration is entered
(2.1)MD=Φ
(2.2) for each ag ∈ SPL, into following circulation;
(2.2.1) calculates the distance of ag and ab (ab ∈ A), obtains Distance matrix D;
(2.2.2) selects n with antigen affinity highest antibody from A;
N selected antibody of (2.2.3) clone, is incorporated to A, and affinity antibody cloning number higher is more, if clone's sum is
Nc;
(2.2.4) is according to formula A=A- ψ (A-X) to NcIndividual clonal antibody enters row variation, and in formula, ψ rates of change vector is and anti-
Former affinity is relevant, and affinity antibody variation rate higher is lower;
(2.2.5) calculates antigen ag and NcThe distance of individual antibody variants;
(2.2.6) selection ξ has the antibody of maximum antigen affinity, forms a memory network Mp;
(2.2.7) is from MpMiddle removal is less than threshold value σ with antigen ag distancesdAntibody;
(2.2.8) calculates MpThe distance between middle antibody sij;
(2.2.9) clone inhibition, from MpMiddle removal sij<σdAntibody;
(2.2.10) merges MpTo MD, i.e. MD=[MD;Mp];
(2.3) clone's taboo, if antigen ag will be less than threshold value σ by distancesAntibody capture, if certain
The antigen number of individual antibody capture is less than threshold value σf, then from MDMiddle removing.
(2.4) clone inhibition again, from MDRemove sij<σdAntibody;
(2.5) a new antibody population A is randomly generated, then with MDMerge, i.e. A=[A;MD];
(3) M is calculatedDAverage distance between middle antibody, if cycle-index exceed maximum, or average distance with it is upper
An iteration compares less than certain threshold value, then stop iteration;Otherwise return to step (2) continues iteration.
Further illustrate, the immune cluster is to memory network (MD) cluster, memory network (MD) can be considered number of vertex
It is antibody number, side right value is a complete graph of affinity between antibody;From memory network (MD) one minimum spanning tree of middle generation;
If pruning threshold is λ, side of the side right value more than λ will be wiped out in complete graph, and then complete graph is divided into multiple connected subgraphs,
Each connected subgraph is one and clusters, and each clusters and represents a feature space for operating mode.
Beneficial effects of the present invention:The present invention proposes a kind of workshop level steam consumption and device level technological parameter is combined
Drain valve steam in leakage synergetic immunity detection method, association is established based on workshop level steam consumption and device level state parameter
With immune detection model, device level steam consumption is overcome to lack the detection difficult brought;Car is established from vulcanization energy consumption mechanism
Dynamic Baseline regression model between intercaste steam consumption and device level technological parameter, produces accurate dynamic risk baseline to send out
Existing efficiency declines;Based on vulcanizer state parameter and jet chimney state parameter, leakage is produced using artificial immune network cluster
The detector of vulcanizer, carrys out the interference of disturbance cancelling factor;The synergetic immunity detection method can exactly detect vulcanizer
With the presence or absence of event is leaked in drain valve steam, loss and false drop rate control, 2.62% and 5.26%, to be set with traditional respectively
Standby level state parameter detection method is compared, and loss and false drop rate reduce by 6.67% and 14.28% respectively.
Brief description of the drawings
Fig. 1 is that the synergetic immunity detection method of leakage in the tyre vulcanizer drain valve steam of one embodiment of the invention is
System schematic diagram;
Fig. 2 is the module of the synergetic immunity detection method of leakage in one embodiment of the invention tyre vulcanizer drain valve steam
Figure;
Fig. 3 is the framework mapping graph of the synergetic immunity detection with biologic immunity mechanism of one embodiment of the invention;
Fig. 4 is the structural representation of the tyre vulcanizer of one embodiment of the invention;
Fig. 5 is the cyclic curve of the gentle internal pressures of capsule Inner of one embodiment of the invention;
Fig. 6 is curing department's steam pipe system schematic diagram of one embodiment of the invention;
Fig. 7 is the workshop steam consumption measured value and detection baseline of one embodiment of the invention;
Fig. 8 be one embodiment of the invention beta pruning after immunological network minimum spanning tree distance map;
Wherein:Green tire 1, vaporium 2, capsule 3, drain valve 4.
Specific embodiment
Further illustrate technical scheme below in conjunction with the accompanying drawings and by specific embodiment.
The synergetic immunity detection method of leakage in a kind of tyre vulcanizer drain valve steam, computer based manual system,
Comprise the following steps:
A:The initialization of calculating main frame system:
(1) structure of danger threshold model:It is dynamic by what is set up between workshop level steam consumption and device level technological parameter
State baseline regression model, produces dynamic risk baseline, sets danger threshold;
(2) structure of detection model:Using vulcanizer state parameter and jet chimney state parameter, by artificial immunity net
Network cluster produces the detector of leakage vulcanizer;
B:The synergetic immunity detection method of computer each run:
The structure of synergetic immunity detection model:The Dynamic Baseline regression model and the detector are carried out into synergetic immunity
Detection.
In tyre vulcanization workshop, leakage frequently results in huge waste of steam in vulcanizer drain valve steam, because equipment
The missing of level steam consumption, the disturbance of the factor such as vapour source, pipe flow speed, not timing be hydrophobic in addition, only with steam room temperature and
Pressure change carries out detection and there is loss and false drop rate higher.
The present invention proposes leakage in the drain valve steam that a kind of workshop level steam consumption and device level technological parameter are combined
Synergetic immunity detection method, exploitation at present is overcome with three innovations on leaking the inspection of problem in vulcanizer drain valve steam
Three difficult points of survey method, i.e. (1) establish synergetic immunity detection mould based on workshop level steam consumption and device level state parameter
Type, overcomes device level steam consumption to lack the detection difficult brought;(2) workshop level steam is established from vulcanization energy consumption mechanism to use
Dynamic Baseline regression model between amount and device level technological parameter, produces accurate dynamic risk baseline come under finding efficiency
Drop;(3) based on vulcanizer state parameter and jet chimney state parameter, leakage vulcanizer is produced using artificial immune network cluster
Detector, carry out the interference of disturbance cancelling factor.The synergetic immunity detection method can exactly detect whether vulcanizer is deposited
Event is leaked in drain valve steam, loss and false drop rate are controlled 2.62% and 5.26%, with traditional device level shape respectively
State parameter detection method is compared, and loss and false drop rate reduce by 6.67% and 14.28% respectively.Fig. 1 is that tyre vulcanizer is hydrophobic
The system schematic of the synergetic immunity detection method of leakage in valve steam;Fig. 2 is the collaboration of leakage in tyre vulcanizer drain valve steam
The module map of immunologic detection method.
Further illustrate, the synergetic immunity detection is event will to be leaked in the vulcanizer drain valve as antigen, described
The external pressure of the vulcanizer in device level technological parameter and outer temperature (pVOAnd tVO) used as epitope, the workshop level steam is used
Amount (M) is danger signal source;
Periodic statistics and detection are carried out to the workshop level steam consumption (M), when the workshop level steam consumption (M) is super
Go out danger threshold scope, then send distress signal, trigger the external pressure and outer temperature (p of the vulcanizerVOAnd tVO) with the detector
Matching, if the match is successful, confirm failure occur, triggering alarm;The external pressure of wherein described vulcanizer and outer temperature (pVOWith
tVO) multiple repairing weld detection can be carried out.
Danger theory in Immune System proposes that an immune system is only produced to harmful antigen and is immunized instead
Should, T lymphocytes and bone-marrow-derived lymphocyte must are fulfilled for two conditions and can just be activated, and one is that they are necessary with the affinity of antigen
More than matching threshold, two is that must receive costimulatory signal.When antigen invades human body and triggers cell abnormal growth, one
Danger signal will be sent by T lymphocytes, be then passed to bone-marrow-derived lymphocyte and be made immune response.Bone-marrow-derived lymphocyte is until receiving
Costimulatory signal is just made a response.In Immune System, autologous/non-autologous identification is selected by T lymphocytes by negative
Rule is selected to perform;The identification of harmful/harmless antigens is then performed by damaged cell by danger threshold.Harmful/harmless antigens
Identification ensure that immune system not to harmless antigens produce immune response, so as to restrained effectively the generation of wrong reaction.
Danger theory of the present invention in Immune System, the collaboration to leaking event in the vulcanizer drain valve is exempted from
Epidemic disease detection is developed based on following two facts:(1) leakage can reduce the external pressure (p of failure vulcanizer in drain valveVO) and it is outer
Temperature (tVO), it is considered as immune response;(2) leakage can significantly reduce the energy efficiency of failure vulcanizer in drain valve, so as to trigger workshop
Steam consumption (M) rises, and is considered as danger signal;
The synergetic immunity detection is mapped with the framework of biologic immunity mechanism as shown in figure 3, i.e. usual workshop level steam is used
Amount (M) is, by periodic statistics and monitoring, when beyond danger threshold scope, then to send distress signal, by the external pressure of vulcanizer and
Outer temperature (pVOAnd tVO) working procedure parameter detected and offered, and drain valve inner hexagon is carried out by detector;Therefore sulphur is worked as
The external pressure of change machine and outer temperature (pVOAnd tVO), and failure report is just triggered after workshop level steam consumption (M) changes to certain limit
It is alert, suppress false drop rate to it;And in order to suppress missing inspection, the external pressure of vulcanizer and outer temperature (pVOAnd tVO) can multiple repairing weld carry out
Detection, to reduce the random disturbances of disturbance factor.
Further illustrate, the foundation of the Dynamic Baseline regression model include the foundation of the thermal balance model for vulcanizing energy consumption and
The danger threshold regression analysis of workshop steam consumption.
Further illustrate, the thermal balance model of the vulcanization energy consumption is formulated as:Wherein QVIt is tire sulphur in the cycle
Change the heat of consumption;QViIt is the total amount of heat of vulcanization process i consumption;KHIt is the specific heat of green tire in vulcanization process, mTRiGreen tire quality,
ΔtOiIt is tOAnd tAiDifference, tOAnd tAiRespectively curing temperature and environment temperature;KSIt is vulcanizer surface coefficient of heat transfer, AiVulcanization
Machine steam chamber surface is accumulated, △ tSiIt is tSAnd tAiDifference, tSIt is vulcanizer steam chamber surface temperature, tAiIt is environment temperature, τiIt is sulphur
The vulcanizer available machine time in the change cycle;KIIt is the coefficient after merger;DTRjAnd BTRiRespectively tire outside diameter and width.
The energy consumed in tire vulcanization process is mainly provided by steam and nitrogen, and energy ezpenditure is mainly used in heating and determines
Type green tire, sizing is main to be completed by high pressure nitrogen, and its energy consumption is well below heating energy consumption.Nitrogen does work and sizing energy consumption,
And sulfur vulcanization chemistries heat is ignored, mainly from the energy supply of thermally equilibrated angle analysis steam and heating energy consumption.
For single track vulcanization process, shown in its energy balance such as formula (1):
QVi=QHi+QSi+QIi+QOi (1)
In formula, QViIt is the total amount of heat of vulcanization process i consumption, QHiIt is that heating green tire absorbs heat, QSiIt is from vulcanizer steam
The heat that chamber surface is distributed, QIiIt is capsule discharge of steam heat, QOiIt is vaporium condensed water elimination heat,.
According to heat absorption principle, QHiAs shown in formula (2):
QHi=KHmTRi(tO-tAi)=KHmTRi△tOi (2)
In formula, KHIt is the specific heat of green tire in vulcanization process, mTRiGreen tire quality is (with m in table 1TR),tOAnd tAiRespectively vulcanize
Temperature (with external pressure steam temperature) and environment temperature, Δ tOiIt is both differences.
According to heat transfer theory, QSiDescription is as shown in formula (3):
QSi=αSAi(tS-tAi)τi=KSAi△tSiτi (3)
In formula, αSIt is vulcanizer surface coefficient of heat transfer, AiVulcanizer steam chamber surface area, tSIt is vulcanizer steam chamber surface
Temperature, is approximately constant, tAiIt is environment temperature, △ tSiIt is both differences, τiIt is the vulcanizer available machine time in curing cycle, including
Standby time, KSIt is coefficient, same to αS。
QIiCapsule filled volume is directly proportional during to work, as shown in formula (4):
In formula, qIIt is steam waste heat coefficient when capsule steam is discharged, DTRjAnd BTRiThat is tire outside diameter and width, with table 1
Middle DTRAnd BTRImplication is identical, KIIt is the coefficient after merger.
QOiRefer to the heat that condensed water is discharged from vaporium, its condensed water is the saturation water under internal pressure steam pressure, according to
Condensation principle, condensation water quantity and QHiAnd QSiIt is directly proportional, as shown in formula (5):
QOi=KO(QHi+QSi) (5)
In formula, KOIt is condensation coefficient.
In thermal balance estimation, can be by KOIt is integrated into KHAnd KSIn, it is not necessary to individually list.For a measurement period
Speech, it is assumed that n vulcanization process is completed in the cycle, then the heat Q for being consumed in the cycleVAs shown in formula (6):
The formula of the thermal balance model of the vulcanization energy consumption expresses workshop steam consumption (M) and vulcanization parameters between
Derivation relationship.Under normal circumstances, the workshop steam consumption (M) of actual measurement should be with the Q obtained by reasoningVApproximately equal, QVEstimate
Evaluation can be as the danger threshold of (M), as M and QVWhen differing larger, then workshop may occur trap leaking event.
Formula (6) is converted into formula (7):
QV=KHXH+KSXS+KIXI+K0 (7)
In formula, XH、XI、XSIt is explanatory variable, corresponding with the statistical items in formula (6) respectively, K0It is normal for needed for regression analysis
It is several, close to 0.
Multiple linear regression based on formula (7) predicts to assess the reasonable interval of workshop steam consumption that the interval upper limit is
Danger threshold, multiple linear regression prediction steps are as follows:
(1) a recurrence sample, Q are obtained from historical dataVObserved value be actual measurement in measurement period workshop steam
Consumption M, XH、XI、XSCounted according to vulcanization parameters respectively, if sample is matrix<Q,X>, Q is the observed value vector of M,
X is XH、XI、XSObservation value matrix;
(2) based on sample<Q,X>To KH、KI、KS、K0Parameter carries out least-squares estimation, ifTo estimate parameter vector, then shown in its estimate such as formula (8):
(3) for individual of sample<QVf,XHf,XIf,XSf>, QVfCorresponding point prediction value such as formula (9):
(4) error ε of predicted value and observed value is calculatedf, as shown in formula (10):
(5) sample is calculated<Q,X>Error variance σ2, as shown in formula (11):
In formula, m is sample size, and m-4 is the free degree (4 represent coefficient number).
(6) statistic t such as formulas (12) are constructed:
Then t obeys the free degree for the t of m-4 is distributed.
(7) level of significance α is given, critical value t is obtainedα/2(m-4), the then Q of confidence level 1- αVfForecast interval such as formula (13):
Because leakage can only trigger steam consumption to rise extremely in drain valve steam, abnormal decline will not be triggered, so only needing
Unilateral detection is done, is takenIt is danger threshold, once the M values of actual measurement exceed danger threshold, then
Send the danger signal of leakage in drain valve.
Further illustrate, the danger threshold regression analysis of the workshop steam consumption is many based on the thermal balance model
First linear regression prediction assesses the reasonable interval of workshop steam consumption (M), and the interval upper limit is danger threshold, when the workshop of actual measurement
Steam consumption (M) exceeds danger threshold, then send the danger signal of leakage in drain valve.
Further illustrate, the artificial immune network cluster is to leak feature according in vulcanization operating mode feature and drain valve, is carried
Take the Inner temperature t of vulcanizer state parameterVI, internal pressure pVI, and its outer temperature tVO, external pressure pVOWith the external pressure of jet chimney state parameter
Vapor (steam) temperature tO, pressure pODifference, then cluster sample characteristics S be formulated as:S=<pVI,tVI,△pVO,△tVO>, △
pVO=pVO-pO, △ tVO=tVO-tO。
The Inner temperature t of the vulcanizer state parameterVIWith internal pressure pVIMainly characterize different vulcanization operating modes;Vulcanizer state is joined
Several outer temperature tVO, external pressure pVOWith the external pressure steam temperature t of jet chimney state parameterO, pressure pODifference be mainly used to judge
Leakage event in drain valve.
Supplementary notes:In sulfidation, external pressure steam enters vaporium from jet chimney by choke valve, by expansion
After heat transfer, temperature and pressure can decline.When being leaked in drain valve, steam can be considered abnormal expansion in vaporium, draw
Hair temperature and pressure declines extremely therewith.But, vapour source temperature and pressure, pipeline steam flow velocity and normal hydrophobic action
Fluctuation, can equally trigger steam chamber temperature and pressure fluctuate, therefore, simply in vaporium temperature and pressure setting threshold
Value, can not effectively find leakage in drain valve steam, and analog value after temperature and pressure in vaporium and pipe decompression valve is carried out
Compare, will can more protrude fault signature.
The different operating modes of sulfidation have larger difference to the depletion rate of external pressure steam, can also trigger steam indoor temperature
With the change of pressure.The feature for vulcanizing each operating mode can recognize that its signature analysis is as follows according to capsule temperature and pressure:
(1) die-filling/die sinking:pVIIt is atmospheric pressure, tVIClose to tA;Punching block radiates to external world, and external pressure steam consumption is fast, pVOSurely
It is fixed, tVODecline.
(2) it is incubated the initial stage:pVIClose to pI, tVIInitial stage rapid increase;External pressure steam consumption is fast, pVOStabilization, tVODecline.
(3) it is incubated the later stage:pVIClose to pI, tVIClose to tI;External pressure steam consumption is slow, pVOStabilization, tVOStabilization.
(4) pressurize:pVIClose to pN, tVIDecline;External pressure steam consumption is slow, pVOStabilization, tVOStabilization.
(5) it is standby:pVIIt is atmospheric pressure, tVIClose to tVO;External pressure steam consumption is slow, pVOStabilization, tVOStabilization.
During above-mentioned operating mode, hydrophobic action, p are such as run intoVOAnd tVOThere are decline, but rapid recovery, therefore, hydrophobic is one
Individual pVOAnd tVOInterference source.
By historical process parameter (pVI, tVI, pVO, tVO) each operating mode is clustered, the ginseng under nominal situation can be extracted
Number cluster feature, after event generation is leaked in drain valve, the operating mode feature of actual measurement can deviate normal cluster feature.
Further illustrate, the artificial immune network cluster includes immune compression and immune cluster;The immune compression mould
Intend the Immune Clone Selection of Immune System and immune taboo, if training sample is SPL, aiNet of the input based on Immune Clone Selection mechanism
Memory network (M is exported after algorithmD), the immune compression includes following algorithm steps:
(1) initiation parameter n, ξ, σs,σd,σf, randomly generate an Antibody Network A;
(2) following iteration is entered
(2.1)MD=Φ
(2.2) for each ag ∈ SPL, into following circulation;
(2.2.1) calculates the distance of ag and ab (ab ∈ A), obtains Distance matrix D;
(2.2.2) selects n with antigen affinity highest antibody from A;
N selected antibody of (2.2.3) clone, is incorporated to A, and affinity antibody cloning number higher is more, if clone's sum is
Nc;
(2.2.4) is according to formula A=A- ψ(A-X)To NcIndividual clonal antibody enters row variation, in formula, ψ rates of change vector, with antigen
Affinity is relevant, and affinity antibody variation rate higher is lower;
(2.2.5) calculates antigen ag and NcThe distance of individual antibody variants;
(2.2.6) selection ξ has the antibody of maximum antigen affinity, forms a memory network Mp;
(2.2.7) is from MpMiddle removal is less than threshold value σ with antigen ag distancesdAntibody;
(2.2.8) calculates MpThe distance between middle antibody sij;
(2.2.9) clone inhibition, from MpMiddle removal sij<σdAntibody;
(2.2.10) merges MpTo MD, i.e. MD=[MD;Mp];
(2.3) clone's taboo, if antigen ag will be less than threshold value σ by distancesAntibody capture, if certain antibody capture
Antigen number is less than threshold value σf, then from MDMiddle removing.
(2.4) clone inhibition again, from MDRemove sij<σdAntibody;
(2.5) a new antibody population A is randomly generated, then with MDMerge, i.e. A=[A;MD];
(3) M is calculatedDAverage distance between middle antibody, if cycle-index exceed maximum, or average distance with it is upper
An iteration compares less than certain threshold value, then stop iteration;Otherwise return to step (2) continues iteration.
In order to recognize the space characteristics of irregular shape, therefore employ the artificial immune network based on Immune Clone Selection mechanism
Algorithm (aiNet), so as to obtain an immunological network for characteristic feature profile as detector, enhancing detection performance.Wherein close
In Algorithm of Artificial Immune Network (aiNet) parameter as shown in following table:
Supplementary notes:The Immune Clone Selection of immune compression simulation Immune System and immune taboo, data quilt to be clustered
It is considered as antigen, antigen is extracted memory network after repetitious stimulation, and memory network forms the spatial image of antigen instrument, but its
Scale is but far smaller than antigen levels.Therefore, memory network is a compression set of antigen collection.Parent between antibody and antigen
And power is represented with the Euler's distance between them, mean that affinity is stronger apart from smaller.Affinity between two antibody
Represented with Euler's distance, mean that affinity is weaker apart from smaller.
Further illustrate, the immune cluster is to memory network (MD) cluster, memory network (MD) can be considered number of vertex
It is antibody number, side right value is a complete graph of affinity between antibody;From memory network (MD) one minimum spanning tree of middle generation;
If pruning threshold is λ, side of the side right value more than λ will be wiped out in complete graph, and then complete graph is divided into multiple connected subgraphs,
Each connected subgraph is one and clusters, and each clusters and represents a feature space for operating mode.
During detection, an antigen by by the antibody capture nearest apart from it, the distance referred to as antigen and Antibody Network away from
From, a recognition threshold is set, when the distance of the antigen and Antibody Network is more than recognition threshold, then the antigen cannot be included into a certain
Cluster, it may be possible to by leaking the abnormal antigen that event is produced in drain valve.
Require supplementation with explanation, the present invention is that isobaric variable temperature sulfuration process more general at present was vulcanized analyzing it
Journey, and two large-scale passenger vehicle tyre enterprises positioned at Pearl River Delta area have been investigated, being determined according to situation about investigating can be automatic
The energy usage of perception, technological parameter and state parameter, and the relational model set up between parameter.
Temperature, pressure and cure time are three key control parameters of sulfuration process, and the above two are by vulcanizer steam chamber
Determine that the latter is then the time that tire carries out vulcanization reaction in vulcanizer with the temperature and pressure in capsule.In order to keep equal
Even to be heated, the steam pressure and temperature parameter that are filled with capsule and vaporium are simultaneously differed, and are supplied by different steam pipe systems
Should.The steam referred to as internal pressure steam of capsule is filled with, vulcanization passes through to vacuumize discharge after terminating, and the temperature and pressure in capsule is usual
Referred to as interior gentle internal pressure;The steam referred to as external pressure steam of vaporium is filled with, its temperature and pressure is less than internal pressure steam, in vaporium
Temperature and pressure is commonly referred to outer gentle external pressure.It is gently outer outward to be pressed in substantially constant during the entire process of vulcanizer work, by outer
Pressure automatic steam control.External pressure steam transmits heat by punching block to green tire, and the condensed water in vaporium is periodically discharged by drain valve.
For capsule, if ignoring mechanically actuated and condensate water discharging process, technical process therein can be divided into insulation and pressurize
Two stages, holding stage is after green tire loads vulcanizer, to be persistently filled with internal pressure steam, capsule is kept steady temperature and pressure
Power;Packing stage is that internal pressure steam stops being filled with, and high pressure nitrogen starts to be filled with pressurize, until punching block is opened, tire is from vulcanizer
Middle taking-up.Internal pressure steam is conducted heat by capsule to green tire in whole process.Fig. 5 shows the cyclic curve of the gentle internal pressures of Inner.From
Temperature curve can be seen that temperature and rise rapidly when being incubated and starting, then held stationary, after packing stage starts, temperature
It is again rapid to raise, then keep near-linear to decline, until terminating.Pressure is then in two stage equal held stationaries.Interior gentle internal pressure
Change curve clearly reflect the different phase of vulcanization.
When state parameter is perceived, used as key control parameter, deployment is sensed in vulcanizer for outer temperature, external pressure, Inner temperature, internal pressure
Device, and automatic data collection and monitoring in the way of near real-time by Process Control System (PCS).The retrievable retrievable work of vulcanizer
Skill parameter and product parameters are as shown in table 1:
The retrievable technological parameter of the vulcanizer of table 1 and product parameters
Curing department's steam pipe system schematic diagram is as shown in Figure 6.Steam enters curing department from main pipeline, and experience is twice
Enter vulcanizer after conversion.Source steam is mixed and is converted into internal pressure steam by 1# and 2# steam converter valves with deaerated water, and 1# and 2# steams
Part internal pressure steam is then converted into external pressure steam by vapour pressure-reducing valve.Internal pressure and the double laterals of external pressure are for raising can after main pipeline
The redundancy pipeline that the property depend on is designed.For reduces cost, resistance vapour influence is reduced, the steam stream of main pipeline is only installed in pipe-line system
Gauge.Meanwhile, steam converter valve and pressure-reducing valve have the function of measurement temperature and pressure.Recovery channel after condensate drain is in this hair
It doesn't matter for bright research, negligible.According to actual conditions, the retrievable actual parameter in workshop is as shown in table 2:
Table 2 workshop gets parms
The present invention has carried out case verification:
The data sample of 30 days, including the workshop shown in table 1 are extracted from some tire enterprise curing department for cooperating
Vulcanizer parameter shown in parameter and table 2.The sampling in every 5 minutes of workshop level parameter once, altogether obtains 4320 records, wherein workshop
Steam consumption (M) is counted once per hour, and 720 records are obtained altogether;Workshop has vulcanizer 95, and vulcanizer parameter is per operation
Extract once, 122744 records are extracted altogether, wherein comprising product parameters such as diameter of tyres, broadband, quality, when insulation and pressurize
Between be relatively fixed for separate unit vulcanizer, it doesn't matter with the specific moment specifically taken in operation, interior temperature, internal pressure, outer temperature,
External pressure belongs to dynamic change parameter, the grab sample from vulcanizer PCS.
Confirm that during sampling, the drain valve for having 4 vulcanizers has interior leakage failure, through technical staff through record of examination
After transferring interior temperature, internal pressure, outer temperature, the external pressure parameter curve trace analysis of failure vulcanizer, determine substantially 4 failures respectively from
14th, 15,15,16 days start, and are repaired in the 18th day.Count accordingly, there are 105 failure notes in workshop steam consumption record
, there are 1089 failure loggings in vulcanizer reference record in record.Below test MATLAB R2013Ra program on the basis of,
Carried out in ordinary PC.
1. workshop steam consumption danger threshold detection test
1.1. the accuracy test of danger threshold regression model
The accuracy of danger threshold directly influences detection performance, pass of the danger threshold of the invention according to formula (6)
It is that model is produced by regression analysis.In enterprise practical, through the steam consumption frequently with unit mass tire as sulfidation energy
Effect index, the present invention sets up a simple quality regression model, is contrasted with formula (6), as shown in formula (18) accordingly:
In formula, KMIt is quality coefficient.
In order to distinguish, the regression model shown in formula (6) and formula (18) is referred to as comprehensive regression model and quality returns mould
Type, QVObserved value be a hour workshop steam consumption (M), the observed value of each explanatory variable in formula the right according to vulcanizer parameter according to
Hour statistical computation is obtained, and obtains the sample S that capacity is 720M, 240 are taken from normal sample as recurrence sample SMr,
Remaining 480 used as test sample SMt。
After regression analysis determines each coefficient, the order of accuarcy of regression model is with returning returning for sample and test sample
Return variances sigma2 εWith the amendment coefficient of determinationTo assess.The amendment coefficient of determinationAs shown in formula (19):
Test result is as shown in table 3:
The workshop steam consumption regression test result of table 3
Regression variance σ2 εThe accuracy of regression model is assessed from the departure degree between regression forecasting value and observed value, is worth
Smaller expression regression model is more accurate.From table 3 it can be seen that either test sample still returns sample, comprehensive regression model
Regression variance will significantly less than quality regression model, thus can determine whether comprehensive regression model prediction accuracy on it is excellent
In quality regression model.
The amendment coefficient of determinationRegression model is measured from the linear fit degree between regression forecasting value and observed value
Accuracy, explanatory variable is bigger to the combined effect degree of explained variable in being worth bigger expression model.From table 4, it can be seen that
Either test sample still returns sample, and the amendment coefficient of determination of comprehensive regression model is noticeably greater than quality regression model, by
This can determine whether that comprehensive regression model combines many energy consumption factors such as tire quality, diameter, width, cure time, with list
One tire quality is compared, and can more accurately explain steam consumption, thus can obtain more excellent predicted value.
1.2. danger threshold detection test
By SMAs test sample, using each coefficient obtained in 1.1 and regression variance is returned, be respectively adopted comprehensive
Regression model and quality regression model are closed, according to formulaDangerous baseline is produced, and to test
Sample carries out fault detect.The danger threshold line that comprehensive regression model is produced is referred to as comprehensive baseline, quality regression model is produced
Danger threshold line turn into quality baseline.
Detection performance is evaluated using false drop rate, loss and error rate, during flase drop is testing result, confirming as normally
Record be judged as it is abnormal;Missing inspection is to be judged as normally abnormal record is confirmed as;False drop rate be flase drop number with it is normal
Several ratios, loss is the ratio of missing inspection number and abnormal number, and error rate is flase drop number and missing inspection number sum and population sample number
Ratio.
Because tα/2Related to confidence level 1- α, in this test, confidence level is attempted between 90%~99%, then takes one
Optimal confidence level parameter, carrys out the detection performance of the comprehensive baseline of comparing and quality baseline, and the test result for obtaining is as shown in table 4.
The workshop steam consumption danger threshold testing result of table 4
From table 4, it can be seen that false drop rate and loss are a pair conflicting indexs, produced for same regression model
Danger threshold for, confidence level increase, then false drop rate decline, loss rise.With reference to three indexs, comprehensive baseline is in confidence
Spend for 97% when error rate be optimal level, respectively 1.30% and 5.71%;Quality baseline is wrong when confidence level is 99%
Rate is optimal level, respectively 5.85% and 18.10% by mistake.Therefore, from the point of view of testing result, the detection performance of comprehensive baseline
It is significantly better than quality baseline.
With the time as abscissa, steam consumption is abscissa, interception part-time section workshop steam consumption measured value and
Detection baseline shows as shown in Figure 7.
Crest and trough are constantly present interior on the same day by M values, because environment temperature meeting cyclically-varying in a day,
And tire heat transfer and steam chamber surface radiating are related to environment temperature, if the product specification of Workshop Production and start number of units do not have
There is too big change, then cause a hour principal element for steam consumption change to be environment temperature, thus M values are interior with ring on the same day
Border temperature change and fluctuate.
It is very sensitive to observation value changes by dangerous baseline it is observed that comprehensive baseline is fitted relatively by force with observation line,
Also obtain good detection performance.Quality baseline because have ignored environment temperature, the influence of product specification parameter, depending mainly on product
The tire quality that goes out and fluctuate, and this parameter is fuctuation within a narrow range in normal production, quality baseline and observation line be fitted compared with
It is weak, it is insensitive to observation value changes, also generate false drop rate and loss higher.Further analysis finds, its flase drop master
Occur in the standby more long or environment temperature relatively low period, missing inspection occurs mainly in the environment temperature period higher, only
Leakage than it is more serious when could find failure.
In order to suppress false drop rate, quality baseline is scheduled on confidence level level higher, but this result in the rising of loss again,
It can be seen from figure 7 that ought be in only one and two vulcanizer drain valves in the case of leakage, quality regression model danger baseline
Failure is not found, and until leakage occurs in three drain valves simultaneously, workshop steam consumption significantly rises, and just detects the event
Barrier.
To sum up analyze, comprehensive baseline can suppress the influence to vulcanization efficiency, specific mass base because of factors such as standby, product specifications
Line can more efficiently occur leakage event in drain valve.
2. vulcanizer temperature and pressure parameter immune detection
4 parameters such as interior temperature, internal pressure, outer temperature, external pressure are extracted from vulcanizer parameter sample, then from workshop parameter
In extract the temperature and pressure of external pressure steam in jet chimney, then by the outer temperature of vulcanizer, external pressure and external pressure steam temperature and pressure
Power is subtracted each other, and obtains the outer temperature difference and external differential, during calculating, according to the vulcanizer parameter extraction time, finds the nearest car of time therewith
Between parameter, between the two parameter correspondence subtracts each other.So, a vulcanization machine testing parameter sample S as shown in formula (14) is obtainedV,
Sample size is 122744, and failure subsample size is 1089.If normal subsample is SVn, failure subsample is SVa。
2.1. immunological network generation test
From SVnIn random extract 10% and constitute training sample SVt, for producing immunological network.AiNet is set for convenience
Empirical parameter in algorithm, sample SVrIn each dimension data carried out standardization processing, i.e., the interval of each dimension data
Compress or be stretched to interval [0,1].In aiNet algorithms, n, ξ, σfAnd σdConvergence rate is only influenceed, convergence quality is not interfered with,
Rule of thumb it is set to 4,10,5 and 1.
Although pruning threshold λ is the key parameter for influenceing immunological network, it is not a sensitive parameter, and its value can
With heuristic, determine after carrying out visual observation according to immunological network minimum spanning tree distance map.One immunological network most your pupil
Into tree distance map as shown in figure 8, abscissa represents the bark mark of immunological network minimum spanning tree in figure, ordinate is the distance on side.
It can be seen that there are 4 height of post (i.e. distance) to be significantly higher than other adjacent edges, illustrate that this is between difference clusters
Connection side, it should to wipe out.As long as pruning threshold is set to maximum universal between back gauge and Smallest connection back gauge, then algorithm is just
The correct beta pruning of energy, immunological network minimum spanning tree is to be split into 5 stalk trees after beta pruning, and the node of 5 stalk trees constitutes 5 and clusters,
Correct each operating mode for having distinguished vulcanization.From the point of view of Fig. 8, the suitable interval of λ is [0.12,0.20], in subsequent experimental, will be set
It is 0.15.
Compression threshold σsIt is another key parameter for influenceing immunological network, it is excessive then to influence correctly to cluster, it is too small
Then immunological network is excessive, influences compression ratio.Here, using heuristic, σ is allowedsProgressively successively decrease from higher value, first generate immune net
Network, then organizes a test sample to be detected again, observation detection performance, finds out the maximum that can keep detecting that performance is optimal
σs。
2.2. immune detection test
From SVn-SVrIn at random extract 5000 samples, with SVaMerge composition test sample SVd.By compression threshold σs
Change between 0.11 to 0.01,0.02 is reduced every time, based on training sample SVtProduce immunological network, antigen and the knowledge between clustering
Other distance is set to 2 σs, followed by immunological network to SVdDetected, observed whether it can correctly distinguish normal sample and failure
Sample.Testing result is as shown in table 5:
The difference of table 5 σsUnder immune detection result
As can be seen from Table 5, σ is worked assAfter dropping to 0.05, false drop rate and loss no longer change, but immune
Network size is with σsDecline and increase, detection efficiency can be influenceed.Therefore, σsIt is more suitable to be set to 0.05, that is, can guarantee that inspection
Efficiency is surveyed, and can keep preferably detecting performance.From the point of view of testing result, false drop rate and loss respectively reach 9.39% He
20.50%, the disturbance factors such as vapour source parameter, steam consumption, not timing be hydrophobic are primarily due to, disturb normal and failure
Feature recognition, detection performance is unsatisfactory.
2.3. cooperation detection is immunized
According to the logic flow of synergetic immunity detection method, first detected for workshop steam consumption per hour, if
Beyond danger threshold, then send distress signal, each vulcanizer technological parameter in this hour is extracted successively and is detected.Because vulcanization
Machine technological parameter is that every operation is extracted once, and vulcanizer has multiple working procedure parameter samples, immunological network inspection in one hour
During survey, as long as the corresponding working procedure parameter sample of vulcanizer detects one for abnormal, that is, think there is leakage thing in drain valve
Part, is verified by multiple samples and suppresses loss.
During detection, 1.2 detection sample and confidence level is continued to use in danger threshold detection, and its testing result and 1.2 is identical,
Therefore retest is not done.Vulcanizer temperature and pressure parameter immune detection uses sample SV, continue to use σ in 4.2.1sWhen=0.05
The immunological network of generation, (also referred to as working procedure parameter is only to have counted vulcanizer temperature and pressure parameter immune detection simultaneously in test
Vertical detection) testing result.Test result is as shown in table 6:
The synergetic immunity testing result of table 6
From testing result as can be seen that synergetic immunity detection loss and false drop rate be substantially less than vulcanizer temperature and
Pressure parameter individually detects that particularly loss is greatly improved.Synergetic immunity detects that the main cause that loss declines is
Leakage directly results in workshop steam consumption lifting in vulcanizer drain valve, reacts than vulcanizer temperature and pressure sensitive, therefore
Loss is inhibited.And cooperation detection is compared with working procedure parameter independent detection, the multiple working procedure parameters sampling to the corresponding period
Detection, eliminates the random error of Subsampling, and loss is inhibited.From the point of view of final relatively low verification and measurement ratio and false drop rate,
The detection method has substantially met application requirement, even if even have missing inspection, can also be corrected in next round detection.
Know-why of the invention is described above in association with specific embodiment.These descriptions are intended merely to explain of the invention
Principle, and can not by any way be construed to limiting the scope of the invention.Based on explanation herein, the technology of this area
Personnel associate other specific embodiments of the invention by would not require any inventive effort, these modes fall within
Within protection scope of the present invention.
Claims (8)
1. the synergetic immunity detection method leaked in a kind of tyre vulcanizer drain valve steam, computer based manual system, its
It is characterised by:Comprise the following steps:
A:The initialization of calculating main frame system:
(1) structure of danger threshold model:By the dynamic base set up between workshop level steam consumption and device level technological parameter
Line regression model, produces dynamic risk baseline, sets danger threshold;
(2) structure of detection model:It is poly- by artificial immune network using vulcanizer state parameter and jet chimney state parameter
Class produces the detector of leakage vulcanizer;
B:The synergetic immunity detection method of computer each run:
The structure of synergetic immunity detection model:The Dynamic Baseline regression model and the detector are carried out into synergetic immunity inspection
Survey.
2. the synergetic immunity detection method leaked in a kind of tyre vulcanizer drain valve steam according to claim 1, it is special
Levy and be:The synergetic immunity detection is event will to be leaked in the vulcanizer drain valve as antigen, the device level technique ginseng
The external pressure of the vulcanizer in number and outer temperature (pVOAnd tVO) used as epitope, the workshop level steam consumption (M) is believed for dangerous
Number source;
Periodic statistics and detection are carried out to the workshop level steam consumption (M), when the workshop level steam consumption (M) is beyond danger
Dangerous threshold range, then send distress signal, and triggers the external pressure and outer temperature (p of the vulcanizerVOAnd tVO) with the detector
Match somebody with somebody, if the match is successful, confirm that failure occurs, triggering alarm;The external pressure of wherein described vulcanizer and outer temperature (pVOAnd tVO) can
Carry out multiple repairing weld detection.
3. the synergetic immunity detection method leaked in a kind of tyre vulcanizer drain valve steam according to claim 1, it is special
Levy and be:The foundation of the Dynamic Baseline regression model includes foundation and the workshop steam consumption of the thermal balance model for vulcanizing energy consumption
Danger threshold regression analysis.
4. the synergetic immunity detection method leaked in a kind of tyre vulcanizer drain valve steam according to claim 3, it is special
Levy and be:The thermal balance model of the vulcanization energy consumption is formulated as:Wherein QVIt is tire sulphur in the cycle
Change the heat of consumption;QViIt is the total amount of heat of vulcanization process i consumption;KHIt is the specific heat of green tire in vulcanization process, mTRiGreen tire quality,
ΔtOiIt is tOAnd tAiDifference, tOAnd tAiRespectively curing temperature and environment temperature;KSIt is vulcanizer surface coefficient of heat transfer, AiVulcanization
Machine steam chamber surface is accumulated, △ tSiIt is tSAnd tAiDifference, tSIt is vulcanizer steam chamber surface temperature, tAiIt is environment temperature, τiIt is sulphur
The vulcanizer available machine time in the change cycle;KIIt is the coefficient after merger;DTRjAnd BTRiRespectively tire outside diameter and width.
5. the synergetic immunity detection method leaked in a kind of tyre vulcanizer drain valve steam according to claim 4, it is special
Levy and be:The danger threshold regression analysis of the workshop steam consumption is that the multiple linear regression based on the thermal balance model is pre-
Survey to assess the reasonable interval of workshop steam consumption (M), the interval upper limit is danger threshold, when the workshop steam consumption (M) of actual measurement
Beyond danger threshold, then the danger signal of leakage in drain valve is sent.
6. the synergetic immunity detection method leaked in a kind of tyre vulcanizer drain valve steam according to claim 1, it is special
Levy and be:Artificial immune network cluster is according to leaking feature in vulcanization operating mode feature and drain valve, extract vulcanizer state
The Inner temperature t of parameterVI, internal pressure pVI, and its outer temperature tVO, external pressure pVOWith the external pressure steam temperature t of jet chimney state parameterO, pressure
Power pODifference, then cluster sample characteristics S be formulated as:S=<pVI,tVI,△pVO,△tVO>, △ pVO=pVO-pO, △
tVO=tVO-tO。
7. the synergetic immunity detection method leaked in a kind of tyre vulcanizer drain valve steam according to claim 1, it is special
Levy and be:The artificial immune network cluster includes immune compression and immune cluster;The immune compression simulation biological immune system
The Immune Clone Selection of system and immune taboo, if training sample is SPL, input is exported after the aiNet algorithms based on Immune Clone Selection mechanism to be remembered
Recall network (MD), the immune compression includes following algorithm steps:
(1) initiation parameter n, ξ, σs,σd,σf, randomly generate an Antibody Network A;
(2) following iteration is entered
(2.1)MD=Φ
(2.2) for each ag ∈ SPL, into following circulation;
(2.2.1) calculates the distance of ag and ab (ab ∈ A), obtains Distance matrix D;
(2.2.2) selects n with antigen affinity highest antibody from A;
N selected antibody of (2.2.3) clone, is incorporated to A, and affinity antibody cloning number higher is more, if clone's sum is Nc;
(2.2.4) is according to formula A=A- ψ (A-X) to NcIndividual clonal antibody enters row variation, and in formula, ψ rates of change vector is affine with antigen
Power is relevant, and affinity antibody variation rate higher is lower;
(2.2.5) calculates antigen ag and NcThe distance of individual antibody variants;
(2.2.6) selection ξ has the antibody of maximum antigen affinity, forms a memory network Mp;
(2.2.7) is from MpMiddle removal is less than threshold value σ with antigen ag distancesdAntibody;
(2.2.8) calculates MpThe distance between middle antibody sij;
(2.2.9) clone inhibition, from MpMiddle removal sij<σdAntibody;
(2.2.10) merges MpTo MD, i.e. MD=[MD;Mp];
(2.3) clone's taboo, if antigen ag will be less than threshold value σ by distancesAntibody capture, if the antigen of certain antibody capture
Number is less than threshold value σf, then from MDMiddle removing.
(2.4) clone inhibition again, from MDRemove sij<σdAntibody;
(2.5) a new antibody population A is randomly generated, then with MDMerge, i.e. A=[A;MD];
(3) M is calculatedDAverage distance between middle antibody, if cycle-index exceedes maximum, or average distance and last time
Iteration compares less than certain threshold value, then stop iteration;Otherwise return to step (2) continues iteration.
8. the synergetic immunity detection method leaked in a kind of tyre vulcanizer drain valve steam according to claim 1, it is special
Levy and be:The immune cluster is to memory network (MD) cluster, memory network (MD) number of vertex is can be considered for antibody number, side
Weights are a complete graph of affinity between antibody;From memory network (MD) one minimum spanning tree of middle generation;If pruning threshold
It is λ, side of the side right value more than λ will be wiped out in complete graph, and then complete graph is divided into multiple connected subgraphs, each connection
Subgraph is one and clusters, and each clusters and represents a feature space for operating mode.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610965184.4A CN106840523B (en) | 2016-11-04 | 2016-11-04 | The synergetic immunity detection method of leakage in a kind of tyre vulcanizer drain valve steam |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610965184.4A CN106840523B (en) | 2016-11-04 | 2016-11-04 | The synergetic immunity detection method of leakage in a kind of tyre vulcanizer drain valve steam |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106840523A true CN106840523A (en) | 2017-06-13 |
CN106840523B CN106840523B (en) | 2018-05-29 |
Family
ID=59146005
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610965184.4A Active CN106840523B (en) | 2016-11-04 | 2016-11-04 | The synergetic immunity detection method of leakage in a kind of tyre vulcanizer drain valve steam |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106840523B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111814923A (en) * | 2020-09-10 | 2020-10-23 | 上海云从汇临人工智能科技有限公司 | Image clustering method, system, device and medium |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101008992A (en) * | 2006-12-30 | 2007-08-01 | 北京市劳动保护科学研究所 | Method for detecting leakage of pipeline based on artificial neural network |
CN102111307A (en) * | 2009-12-29 | 2011-06-29 | 亿阳信通股份有限公司 | Method and device for monitoring and controlling network risks |
US20110196251A1 (en) * | 2010-02-10 | 2011-08-11 | Nellcor Puritan Bennett Llc | Leak determination in a breathing assistance system |
JP2012185109A (en) * | 2011-03-08 | 2012-09-27 | Panasonic Corp | Gas leak detection system for fuel cell system |
-
2016
- 2016-11-04 CN CN201610965184.4A patent/CN106840523B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101008992A (en) * | 2006-12-30 | 2007-08-01 | 北京市劳动保护科学研究所 | Method for detecting leakage of pipeline based on artificial neural network |
CN102111307A (en) * | 2009-12-29 | 2011-06-29 | 亿阳信通股份有限公司 | Method and device for monitoring and controlling network risks |
US20110196251A1 (en) * | 2010-02-10 | 2011-08-11 | Nellcor Puritan Bennett Llc | Leak determination in a breathing assistance system |
JP2012185109A (en) * | 2011-03-08 | 2012-09-27 | Panasonic Corp | Gas leak detection system for fuel cell system |
Non-Patent Citations (1)
Title |
---|
闫利军 等: "基于免疫算法的概念设计协同决策方法研究", 《第十五届全国机械设计年会论文集》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111814923A (en) * | 2020-09-10 | 2020-10-23 | 上海云从汇临人工智能科技有限公司 | Image clustering method, system, device and medium |
CN111814923B (en) * | 2020-09-10 | 2020-12-25 | 上海云从汇临人工智能科技有限公司 | Image clustering method, system, device and medium |
Also Published As
Publication number | Publication date |
---|---|
CN106840523B (en) | 2018-05-29 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112200433B (en) | Power plant thermal performance analysis and optimization system | |
CN106872657B (en) | A kind of multivariable water quality parameter time series data accident detection method | |
CN110441065B (en) | Gas turbine on-line detection method and device based on LSTM | |
CN107949812B (en) | Method for detecting anomalies in a water distribution system | |
CN104994535B (en) | Sensor data stream method for detecting abnormality based on Multidimensional Data Model | |
CN102246110B (en) | Standardization of data used for monitoring an aircraft engine | |
CN109442221B (en) | A kind of water supply network booster method for detecting extracted based on pressure disturbance | |
CN104899327B (en) | A kind of time series method for detecting abnormality of no class label | |
CN105930629B (en) | A kind of on-line fault diagnosis method based on magnanimity service data | |
CN107562696B (en) | On-line detection and control method for tire product quality | |
CN104573850A (en) | Method for evaluating state of thermal power plant equipment | |
CN106447210A (en) | Distribution network equipment health degree dynamic diagnosis method involving credibility evaluation | |
CN103939749A (en) | Oil pipeline network leakage intelligent self-adaptation monitoring system and method based on big data | |
CN108092623B (en) | A kind of photovoltaic array multisensor fault detecting and positioning method | |
CN106796157A (en) | Abnormal method in detection distribution network, particularly drinking water distribution network | |
CN112836758B (en) | Independent metering area pipe network leakage detection method based on double-layer circulating neural network | |
CN105717888B (en) | A kind of intelligent early-warning system | |
CN112208506B (en) | Intelligent fault detection method for air brake device of railway wagon | |
CN109798935A (en) | A kind of Multiple Source Sensor synergic monitoring method for early warning and system for curtain wall | |
CN110826855A (en) | Method and system for testing network access performance of intelligent power distribution room state monitoring sensor | |
CN108764305A (en) | A kind of improved colony intelligence machine learning fault diagnosis system | |
CN106840523B (en) | The synergetic immunity detection method of leakage in a kind of tyre vulcanizer drain valve steam | |
CN114997539A (en) | Gas pressure regulator fault prediction method and system | |
CN115047839A (en) | Fault monitoring method and system for industrial process of preparing olefin from methanol | |
CN107222169A (en) | A kind of photovoltaic array based on outlier detection is monitored and fault detection method in real time |
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 | ||
TR01 | Transfer of patent right |
Effective date of registration: 20230731 Address after: 164300 feihui import and export processing base to Russia, Aihui District, Heihe City, Heilongjiang Province Patentee after: HEILONGJIANG RED VALLEY AUTOMOTIVE TEST CO.,LTD. Address before: 293 Zhongshan Avenue West, Tianhe District, Guangzhou, Guangdong 510000 Patentee before: GUANGDONG POLYTECHNIC NORMAL University |
|
TR01 | Transfer of patent right |