CN104200109B - A kind of city rail vehicle air-conditioning system method for diagnosing faults and device - Google Patents
A kind of city rail vehicle air-conditioning system method for diagnosing faults and device Download PDFInfo
- Publication number
- CN104200109B CN104200109B CN201410455410.5A CN201410455410A CN104200109B CN 104200109 B CN104200109 B CN 104200109B CN 201410455410 A CN201410455410 A CN 201410455410A CN 104200109 B CN104200109 B CN 104200109B
- Authority
- CN
- China
- Prior art keywords
- signal
- degree
- matrix
- fault
- subordinated
- 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.)
- Active
Links
Landscapes
- Air Conditioning Control Device (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The present invention provides a kind of city rail vehicle air-conditioning system method for diagnosing faults and devices, after the fault-signal acquired when getting acquisition module and being monitored to air-conditioning system, the fault-signal is handled using wavelet package transforms algorithm, to obtain the reconstruction signal of each frequency range of the fault-signal, later, fault-signal characteristic information is can be obtained according to the reconstruction signal, at this time, it is calculated using possibility clustering algorithm, it can quickly and accurately determine fault type corresponding with the fault message, accurately and quickly to be handled according to the fault type, to make the fast quick-recovery normal work of the city rail vehicle air-conditioning system, ensure the environment by bus of occupant comfort.
Description
Technical field
The present invention relates to air-conditioning system fault diagnosis technology fields, more particularly to a kind of city rail vehicle air-conditioning system
System method for diagnosing faults and device.
Background technology
Nowadays, in order to improve comfort level of the passenger in city rail vehicle, usually air-conditioning system can be all set on city rail vehicle,
The humidity and temperature of air in a car compartment are adjusted, specifically, the control system passed through in the air-conditioning system controls air-conditioner set pair
After air in compartment and extraneous new gas are handled, by ventilating system will treated air delivery to guest room, to protect
The harmony of the temperature and humidity of the objective room air of card.Wherein, in order to ensure the reliably working of city rail vehicle air-conditioning system, in reality
In the application of border, control system needs the operating condition of monitoring air-conditioner set and ventilating system in real time, to occur in this two parts
Timely processing is carried out when failure.
Currently, existing city rail vehicle air-conditioning system fault diagnosis usually all manually carries out, diagnosis efficiency and accurate
Degree is all relatively low, moreover, because the air-conditioning system is complicated, internal part can not be observed directly in the process of running,
It will be unable to be diagnosed to be fault zone and its fault type in time at this time, and then also just can not fast quick-recovery city rail vehicle air-conditioning system
Normal work.
Invention content
In view of this, the present invention provides a kind of city rail vehicle air-conditioning system method for diagnosing faults and device, solve existing
Have when manually carrying out city rail vehicle air-conditioning system fault diagnosis in technology, because the air-conditioning system structure can not, internal part exists
The technical issues of can not directly being observed in operational process, thus greatly reducing fault diagnosis efficiency and accuracy.
To achieve the above object, the present invention provides the following technical solutions：
A kind of city rail vehicle air-conditioning system method for diagnosing faults, the method includes：
Obtain the fault-signal acquired when acquisition module is monitored air-conditioning system；
Using wavelet package transforms algorithm, the fault-signal is handled, obtains each frequency letter in the fault-signal
Number reconstruction signal；
Using the reconstruction signal, fault-signal feature vector is obtained；
The fault-signal feature vector is calculated using possibility clustering algorithm, is determined and the fault-signal pair
The fault type answered.
Preferably, described to use wavelet package transforms algorithm, the fault-signal is handled, the fault-signal is obtained
In each frequency signal reconstruction signal, including：
WAVELET PACKET DECOMPOSITION is carried out to the fault-signal using preset decomposition scale, obtains each frequency in the fault-signal
Signal；
It extracts the signal characteristic under each preset decomposition scale in each frequency signal and makees each frequency under corresponding preset decomposition scale
The reconstruction signal of rate signal.
Preferably, described to utilize the reconstruction signal, fault-signal feature vector is obtained, including：
Calculate the energy of the reconstruction signal of each frequency signal and all frequency signals under the last one preset decomposition scale
Reconstruction signal gross energy；
The energy of the reconstruction signal is normalized, determines the energy of each reconstruction signal in the gross energy
Shared ratio；
According to the energy of each reconstruction signal ratio shared in the gross energy, structure fault-signal feature to
Amount.
Preferably, described that the fault-signal feature vector is calculated using possibility clustering algorithm, determining and institute
The corresponding fault type of fault-signal is stated, including：
The initial parameter of possibility clustering algorithm is obtained, the initial parameter includes：Cluster centre number, weighted index, just
Beginning cluster centre matrix and initial subordinated-degree matrix；
Calculate each sample for being determined by the fault-signal feature vector to the initial cluster center matrix away from
From；
By in the distance of each described sample to the initial cluster center and the initial subordinated-degree matrix with institute
It states and substitutes into degree of membership iterative formula apart from corresponding degree of membership, calculate the subordinated-degree matrix after iteration；
Judge whether the subordinated-degree matrix after the iteration and the difference of the initial subordinated-degree matrix are less than default maximum
Allowable error；
If it is, using the subordinated-degree matrix after the iteration as target subordinated-degree matrix；
If it is not, then using the subordinated-degree matrix after the iteration as initial subordinated-degree matrix, and judge current iteration time
Whether number reaches default maximum iteration；
If not up to, returned described by the distance of each described sample to the initial cluster center and described
Degree of membership iterative formula is substituted into apart from corresponding degree of membership with described in initial subordinated-degree matrix, calculates the degree of membership after iteration
Matrix step continues to execute；
If reached, using the corresponding subordinated-degree matrix of current iteration number as target subordinated-degree matrix；
By in the fault-signal feature vector each sample and the target subordinated-degree matrix in the sample
This corresponding degree of membership substitutes into cluster centre iterative formula, calculates target cluster centre matrix；
Calculate the approach degree of the data sample and each cluster centre in the target cluster centre matrix of unknown failure type；
According to result of calculation, fault type corresponding with the fault-signal is determined.
Preferably, described by the distance of each described sample to the initial cluster center and the initial degree of membership
Degree of membership iterative formula is substituted into apart from corresponding degree of membership with described in matrix, calculates the subordinated-degree matrix after iteration, including：
By in the distance of each described sample to the initial cluster center and the initial subordinated-degree matrix with institute
It states and determines formula apart from corresponding degree of membership substitution adjustment parameter, determine with described apart from corresponding adjustment parameter ηi, the tune
Section parameter determines that formula is：
Wherein, i=1,2 ..., the preset decomposition scale parameter；J=1,2 ..., n, n are indicated under each preset decomposition scale
The dispersion index that each node calculates；uijIndicate that each node corresponds under each preset decomposition scale in initial subordinated-degree matrix
Degree of membership, dijIndicate each sample to the initial cluster center distance；M indicates the weighted index；K=1；
By the distance of each described sample to the initial cluster center and with described apart from corresponding adjustment parameter
Substitute into degree of membership iterative formula, calculate the subordinated-degree matrix u after iteration,ij, the degree of membership iterative formula is：
Preferably, the cluster centre iterative formula is：
Wherein, u,ijIndicate degree of membership corresponding with each node under each preset decomposition scale in target subordinated-degree matrix；I=
1,2 ..., the preset decomposition scale parameter；J=1,2 ..., n；N indicates that each node is computed under each preset decomposition scale
The dispersion index arrived；M indicates the weighted index, xjIndicate j-th of sample in the fault-signal feature vector.
Preferably, in respectively being clustered in the data sample and the target cluster centre matrix of the calculating unknown failure type
The approach degree of the heart, including：
According to formulaCalculate data sample and the institute of unknown failure type
State the approach degree σ (Data, V) of each cluster centre in target cluster centre matrix；
Wherein, min expressions are minimized operation；Max expressions are maximized operation；dataikIndicate unknown failure type
Data sample, v,jkIndicate each cluster centre in the target cluster centre matrix；I=1,2 ..., n；J=1,2 ..., C；n
Indicate the dispersion index that each node calculates under each preset decomposition scale；C indicates the cluster centre number, m
Indicate the weighted index.
A kind of city rail vehicle air-conditioning system trouble-shooter, described device include：
First acquisition module, for obtaining the fault-signal acquired when acquisition module is monitored air-conditioning system；
First processing module handles the fault-signal, obtains the event for using wavelet package transforms algorithm
Hinder the reconstruction signal of each frequency signal in signal；
First determining module obtains fault-signal feature vector for utilizing the reconstruction signal；
First computing module, for being calculated the fault-signal feature vector using possibility clustering algorithm, really
Fixed fault type corresponding with the fault-signal.
Preferably, the first processing module includes：
First resolving cell obtains institute for carrying out WAVELET PACKET DECOMPOSITION to the fault-signal using preset decomposition scale
State each frequency signal in fault-signal；
First extraction unit is made to correspond in advance for extracting the signal characteristic under each preset decomposition scale in each frequency signal
If the reconstruction signal of each frequency signal under decomposition scale.
Preferably, first determining module includes：
First computing unit, the energy for calculating the reconstruction signal of each frequency signal under the last one preset decomposition scale
And the gross energy of the reconstruction signal of all frequency signals；
First processing units are normalized for the energy to the reconstruction signal, determine each reconstruction signal
Energy ratio shared in the gross energy；
First construction unit, for according to the energy of each reconstruction signal ratio shared in the gross energy, structure
Build fault-signal feature vector.
It can be seen via above technical scheme that compared with prior art, the present disclosure provides a kind of city rail vehicle skies
Adjusting system method for diagnosing faults and device, the fault-signal acquired when getting acquisition module and being monitored to air-conditioning system
Afterwards, the fault-signal is handled using wavelet package transforms algorithm, to obtain the reconstruction signal of each frequency range of the fault-signal,
Later, fault-signal characteristic information can be obtained according to the reconstruction signal, at this point, being carried out to it using possibility clustering algorithm
It calculates, you can fault type corresponding with the fault message is quickly and accurately determined, to be carried out according to the fault type
It accurately and quickly handles, to make the fast quick-recovery normal work of the city rail vehicle air-conditioning system, ensures riding for occupant comfort
Environment.
Description of the drawings
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technology description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The embodiment of invention for those of ordinary skill in the art without creative efforts, can also basis
The attached drawing of offer obtains other attached drawings.
Fig. 1 is a kind of flow chart of city rail vehicle air-conditioning system method for diagnosing faults of the present invention；
Fig. 2 is a kind of decomposition texture schematic diagram of fault-signal WAVELET PACKET DECOMPOSITION of the present invention；
Fig. 3 is the flow chart of another city rail vehicle air-conditioning system method for diagnosing faults of the present invention；
Fig. 4 is a kind of structural schematic diagram of city rail vehicle air-conditioning system trouble-shooter of the present invention.
Specific implementation mode
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation describes, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
In practical applications, it in order to provide comfortably environment by bus to passenger, usually can be all equipped in city rail vehicle
Air-conditioning system, to adjust the humidity and temperature of the air in compartment.Wherein, air-conditioning system is mainly by control system, ventage
System and air-conditioner set three parts composition, and the control system by PLC (Programmable Logic Controller, may be programmed
Logic controller), temperature expansion module and information display module composition, will be obtained from contact net by auxiliary power supply system
1500V direct currents are rectified into 110V direct currents to power, and the core by PLC as the control system is realized and transported to air-conditioner set
The real-time monitoring of market condition, and real-time display is carried out by information display module；Air-conditioner set mainly by air conditioner unit and
Compressor condenser unit is constituted, when the fresh air and indoor return air that the centrifugal blower received in ventilating system sucks outside vehicle
Afterwards, a series of processing of air will be carried out, and gas sends guest room back to by treated by main air duct.Wherein, air-conditioner set is installed
At the both ends of every section car body top, a main air duct is shared, air supply duct is longitudinally arranged along car body top, from compartment both ends to vehicle
Air-supply in the middle part of compartment, and it is equipped with fresh wind port at air-conditioner set both ends, front end is equipped with air outlet, and roof air-supply passage is blown with air conditioner
Mouth is connected, and ensures that hot and cold air exchanges in guest room, and a pleasant environment is created for guest room.
It has been investigated that for complicated air-conditioning system, same failure often causes there are many different reasons,
And failure symptom caused by the failure of different components is also identical sometimes, staff can not diagnose fault according to failure symptom
Type, still more, in air-conditioning system operational process, since its internal part is directly to observe, thus, if staff
Check inside air-conditioning system whether failure, it is necessary to the air-conditioning system is out of service, it is clear that the inspection and processing procedure are inevitable
It can influence the environment by bus of passenger.
In order to solve manually to carry out the problems in city rail vehicle air-conditioning system fault diagnosis, the embodiment of the invention discloses one
Kind city rail vehicle air-conditioning system method for diagnosing faults and device, are adopted when getting acquisition module and being monitored to air-conditioning system
After the fault-signal of collection, the fault-signal is handled using wavelet package transforms algorithm, to obtain each frequency of the fault-signal
The reconstruction signal of section obtains fault-signal characteristic information according to the reconstruction signal later, at this point, possibility cluster is used to calculate again
Method calculates it, you can and fault type corresponding with the fault message is automatically determined out, fault diagnosis is carried out without artificial,
Speed and the accuracy for improving fault diagnosis ensure that make the fast quick-recovery normal work of the city rail vehicle air-conditioning system
The environment by bus of occupant comfort.
Embodiment one：
As shown in Figure 1, for a kind of flow chart of city rail vehicle air-conditioning system method for diagnosing faults of the present invention, this method can be with
Including：
Step S101：Obtain the fault-signal acquired when acquisition module is monitored air-conditioning system.
In the practical application of the embodiment of the present invention, city rail vehicle air-conditioning system can in real time be supervised by acquisition module
It surveys, to acquire the fault-signal of the air-conditioning system in time.Wherein, which is specifically as follows and each component of air-conditioning system
Connected sensor, the type of the sensor can according to coupled component it needs to be determined that, then the fault-signal specifically may be used
Think voltage or current signal.
Step S102：Using wavelet package transforms algorithm, which is handled, obtains each frequency in the fault-signal
The reconstruction signal of rate signal.
In embodiments of the present invention, after obtaining the fault-signal of city rail vehicle, WAVELET PACKET DECOMPOSITION can be carried out to it, to
Obtain the approximate part of low-frequency range and the detail section of high band.Specifically, using preset decomposition scale to acquired failure
Signal is decomposed, and each frequency signal in the fault-signal is obtained, and is extracted under each preset decomposition scale in each frequency band signals
Signal characteristic makees the reconstruction signal of each frequency signal under corresponding preset decomposition scale.Wherein, preset decomposition scale refers to failure
The WAVELET PACKET DECOMPOSITION number of plies that signal carries out
For example, if preset decomposition scale is 3 layers of WAVELET PACKET DECOMPOSITION, when the fault-signal for getting air-conditioning system
Afterwards, 3 layers of WAVELET PACKET DECOMPOSITION will be carried out to the fault-signal, decomposition texture is as shown in Fig. 2, (i, j) indicates i-th layer of jth in figure
A node (i=0,1,2,3；J=0,1,2,3,4,5,6,7), every layer has 2iA node, each node can represent centainly
Signal characteristic.Wherein, (0,0) node on behalf original signal S, the first layer low frequency system of (1,0) node on behalf WAVELET PACKET DECOMPOSITION
Number, the first layer high frequency coefficient of (1,1) node on behalf WAVELET PACKET DECOMPOSITION, the 0th node system of (2,0) node on behalf second layer,
Other and so on.
Later, all WAVELET PACKET DECOMPOSITION coefficients (i.e. each frequency signal) is reconstructed, obtains the reconstruction signal of every layer of each frequency,
Still by taking Fig. 2 as an example, then the reconstruction signal of i-th layer of j-th of node is represented by Sij, such as S32Indicate the reconstruct letter of (3,2) node
Number, then, the reconstruction signal of original signal and the relationship of the reconstruction signal of the 3rd node layer can be expressed as：
S=S30+S31+S32+S33+S34+S35+S36+S37 (1)
Step S103：Using the reconstruction signal, fault-signal feature vector is obtained.
In embodiments of the present invention, the reconstruction signal of each frequency signal under the last one preset decomposition scale need to only be calculated
Energy (is denoted as Eij)And under the preset decomposition scale reconstruction signal of all frequency signals gross energy (being denoted as E), lead to
The normalized to the energy of the reconstruction signal is crossed, determines the energy of each reconstruction signal ratio shared in the gross energy,
And it constructs using each ratio of gained as the fault-signal feature vector of vector element, at this point, in order to intuitively find out energy
Variation can also draw energy spectrum scalogram according to gained ratio.
It is still illustrated with the example above, it is assumed that the reconstruction signal S of the 3rd layer of j-th of node3j(j=0,1 ..., 7), it is right
The energy answered is E3j(j=0,1 ..., 7), then the energy of the reconstruction signal of each frequency signal is：
In formula, xjkIndicate the amplitude of the discrete point of the reconstruction signal of each frequency signal；K=1,2 ..., n；N indicates the 3rd layer
J-th of node be computed after dispersion index, i.e. the sampling number of fault-signal.
Then the gross energy of the reconstruction signal of all frequency signals is：
Later, the energy of the reconstruction signal of each frequency signal of gained is normalized, that is, calculates each reconstruct letter
Number energy ratio shared in gross energy, wherein normalized formula is：
At this point, the embodiment of the present invention is then with normalized result P30, P31..., P37For vector element, structure failure letter
Number feature vector T, i.e. T=[P30,P31,P32,P33,P34,P35,P36,P37]。
Step S104：The fault-signal feature vector is calculated using possibility clustering algorithm, is determined and the event
Hinder the corresponding fault type of signal.
In embodiments of the present invention, when determining fault type, which can also include：Display is reported
The fault type, to remind staff to repair the air-conditioning system for the fault type.
Certainly, after executing the step S14, which can also include：It obtains
Processing routine corresponding with the fault type is taken, and executes the processing routine, to exclude the failure of the air-conditioning system, nothing
Troubleshooting need to be manually carried out, has saved labour, and improve work efficiency.
It learns from the above analysis, in embodiments of the present invention, when getting acquisition module and being monitored to air-conditioning system
After the fault-signal acquired, the fault-signal is handled using wavelet package transforms algorithm, to obtain the fault-signal
In the reconstruction signal of each frequency signal can be obtained fault-signal feature vector using the reconstruction signal later, and using may
Property clustering algorithm calculates the fault-signal feature vector, you can quickly and accurately determines corresponding with the fault message
Fault type, accurately and quickly to be handled the failure of the city rail vehicle air-conditioning system according to the fault type,
To make the fast quick-recovery normal work of the air-conditioning system, ensure the environment by bus of occupant comfort.
Embodiment two：
As shown in figure 3, for the flow chart of another city rail vehicle air-conditioning system method for diagnosing faults of the present invention, this method can
To include：
Step S301：Obtain the fault-signal acquired when acquisition module is monitored air-conditioning system.
Step S302：WAVELET PACKET DECOMPOSITION is carried out to the fault-signal using preset decomposition scale, is obtained in the fault-signal
Each frequency signal.
In embodiments of the present invention, due to realizing the WAVELET PACKET DECOMPOSITION to fault-signal using top and bottom process, thus, this is default
Decomposition scale refers to carrying out the WAVELET PACKET DECOMPOSITION number of plies to fault-signal.
Step S303：It extracts the signal characteristic under each preset decomposition scale in each frequency signal and makees corresponding preset decomposition ruler
The reconstruction signal of each frequency signal under degree.
Example as shown in Figure 2, preset decomposition scale refer to carrying out 3 layers of WAVELET PACKET DECOMPOSITION to fault-signal, of the invention real
It applies in example, when extracting each layer of signal characteristic, can be extracted from each frequency signal according to sequence from low to high
Signal characteristic, and as the reconstruction signal of respective frequencies signal.
Step S304：It calculates the energy of the reconstruction signal of each frequency signal under the last one preset decomposition scale and owns
The gross energy of the reconstruction signal of frequency signal.
Still by taking Fig. 2 as an example, the embodiment of the present invention can only calculate the reconstruct letter of each frequency signal of the 3rd layer (i.e. the bottom)
Number ENERGY E3jThe gross energy E of (j=0,1,2,3,4,5,6,7) and all reconstruction signals of third layer, then E=E30+E31+E32
+E33+E34+E35E36+E37。
Step S305：The energy of the reconstruction signal is normalized, determines the energy of each reconstruction signal described
Shared ratio in gross energy.
Wherein, normalized formula can be above-mentioned formula (4).
Step S306：According to the energy of each reconstruction signal ratio shared in gross energy, structure fault-signal feature to
Amount.
The embodiment of the present invention is using each ratio obtained by step S34 as vector element, structure fault-signal feature vector T.
Step S307：Default possibility clustering algorithm is initialized.
Step S308：Obtain the initial parameter of the possibility clustering algorithm.
Wherein, which may include：Cluster centre number C, weighted index m, initial cluster center matrix V and initial
Subordinated-degree matrix U.In practical applications, operator can be set according to actual needs the initial parameter, wherein the cluster centre
Number C can be the quantity of the known fault type of the city rail vehicle air-conditioning system；Weighted index m dimensionless, value range are usual
For [1.1,5]；Initial cluster center matrix V can select every class sample (to be made of the fault-signal of known fault type
Fault feature vector) sample average as its initial value；Initial subordinated-degree matrix U can be the sample pair of each known fault type
The degree of membership of initial cluster center, and element value range therein is [0,1].
Step S309：Calculate each sample for being determined by fault-signal feature vector to initial cluster center matrix away from
From.
In embodiments of the present invention, the element in the fault-signal feature vector of above-mentioned steps S306 structures can be utilized,
Determine the sample X in feature spacej(j=0,1 ..., n), wherein each sample XjIncluding P3i(i=1,2,3,4,5,6,7)
Eight components, it is seen then that when sample difference, including component be also different, but the formula for obtaining each component is identical.If
With with vi(i=0,1 ..., C) indicates each element (i.e. cluster centre) in initial cluster center matrix, then according to pre-determined distance
Calculation formula, calculate each sample to corresponding cluster centre distance dij, wherein the distance calculation formula is：
dij=| | xj-vj||2 (5)
Step S310：By in the distance of each sample to the initial cluster center and initial subordinated-degree matrix with should be away from
Degree of membership iterative formula is substituted into from corresponding degree of membership, calculates the subordinated-degree matrix after iteration.
In embodiments of the present invention, according to above-mentioned formula (5) calculate each sample to corresponding cluster centre distance
dijAfterwards, can by itself and initial subordinated-degree matrix center with this apart from corresponding degree of membership uijFollowing adjustment parameters are substituted into determine
Formula, calculate with this apart from corresponding adjustment parameter ηi, which determines that formula is：
Wherein, the K in the formula is a positive integer, the constant coefficient only calculated as adjustment parameter, no practical significance,
Usually desirable 1.
Later, by the distance d of a sample to corresponding cluster centreijAnd with this apart from corresponding adjustment parameter ηiIt substitutes into
Degree of membership iterative formula, calculate iteration after subordinated-degree matrix U, which can be：
Step S311：Judge whether the subordinated-degree matrix after iteration and the difference of initial subordinated-degree matrix are less than default maximum
Allowable error, if so, thening follow the steps S312；If not, thening follow the steps S313.
If the subordinated-degree matrix after l iteration is denoted as U (l) in embodiments of the present invention, the degree of membership before iteration
Matrix, that is, initial subordinated-degree matrix is U (l-1), wherein l=1,2 ..., z, z are default iterations.After iteration each time,
Whether the difference that the subordinated-degree matrix U (l) and the subordinated-degree matrix U (l-1) before iteration after iteration will be solved, judge the difference
Less than default limits of error ε, i.e. ‖ U (l)-U (l-1) ‖<Whether ε is true, if so, then follow the steps S312；If not at
It is vertical, then follow the steps S313.
Step S312：Using the subordinated-degree matrix after iteration as target subordinated-degree matrix.
After determining target subordinated-degree matrix, step S315 can be directly executed, with the corresponding target cluster centre square of determination
Battle array.
Step S313：Using the subordinated-degree matrix after iteration as initial subordinated-degree matrix, and judge that current iteration number is
It is no to reach default maximum iteration；If reached, S314 is thened follow the steps；If not up to, return to step S310 continues
It executes.
It should be noted that during determining target subordinated-degree matrix, iterations be it is limited, can therefore
Before barrier diagnosis starts, maximum iteration is pre-set.In practical applications, if after default iterations, can not still expire
Sufficient formula ‖ U (l)-U (l-1) ‖<ε, it will the subordinated-degree matrix that last time iteration is obtained is as target subordinated-degree matrix.
Step S314：Using the corresponding subordinated-degree matrix of current iteration number as target subordinated-degree matrix.
Step S315：By in each sample and the target subordinated-degree matrix in fault-signal feature vector with the sample
Corresponding degree of membership substitutes into cluster centre iterative formula, calculates target cluster centre matrix.
Wherein, which can be：
Wherein, u、 ijIndicate degree of membership corresponding with each node under each preset decomposition scale in target subordinated-degree matrix；I=
1,2 ..., preset decomposition scale parameter；J=1,2 ..., n, n indicate that each node under each preset decomposition scale calculates
Dispersion index；M indicates the weighted index, xjIndicate j-th of sample in the fault-signal feature vector.
Step S316：Calculate the patch of the data sample and each cluster centre in target cluster centre matrix of unknown failure type
Recency.
In embodiments of the present invention, approach degree calculation formula used in step S316 can be：
Wherein, min expressions are minimized operation；Max expressions are maximized operation；dataikIndicate unknown failure type
Data sample, v、 jkIndicate each cluster centre in the target cluster centre matrix；I=1,2 ..., n；J=1,2 ..., C；n
Indicate the dispersion index that each node calculates under each preset decomposition scale；C indicates the cluster centre number, m
Indicate the weighted index.
Step S317：According to the result of calculation, fault type corresponding with fault-signal is determined.
In embodiments of the present invention, the corresponding fault type of maximum approach value being calculated is that acquired failure is believed
Number corresponding fault type, at this point, operator can correspondingly be handled air-conditioning system according to the fault type determined, with
The air-conditioning system is set to be restored to normal work as early as possible.
Based on above-mentioned analysis it is found that in embodiments of the present invention, being monitored to air-conditioning system when getting acquisition module
When the fault-signal that is acquired after, the fault-signal is handled using wavelet package transforms algorithm, to obtain failure letter
The reconstruction signal of each frequency signal in number can be obtained fault-signal feature vector, and using can using the reconstruction signal later
Energy property clustering algorithm calculates the fault-signal feature vector, you can quickly and accurately determines and the fault message pair
The fault type answered, accurately and quickly to be located to the failure of the city rail vehicle air-conditioning system according to the fault type
Reason ensures the environment by bus of occupant comfort to make the fast quick-recovery normal work of the air-conditioning system.
Embodiment three：
As shown in figure 4, for a kind of structural schematic diagram of city rail vehicle air-conditioning system trouble-shooter of the present invention, the device
May include：
First acquisition module S401：For obtaining the failure acquired when acquisition module is monitored air-conditioning system letter
Number.
First processing module S402：For using wavelet package transforms algorithm, fault-signal is handled, the failure is obtained
The reconstruction signal of each frequency signal in signal.
Wherein, in embodiments of the present invention, first processing module S402 may include：
First resolving cell obtains the failure for carrying out WAVELET PACKET DECOMPOSITION to fault-signal using preset decomposition scale
Each frequency signal in signal.
First extraction unit is made to correspond in advance for extracting the signal characteristic under each preset decomposition scale in each frequency signal
If the reconstruction signal of each frequency signal under decomposition scale.
First determining module S403：For utilizing the reconstruction signal, fault-signal feature vector is obtained.
Preferably, first determining module S403 may include：
First computing unit, the energy for calculating the reconstruction signal of each frequency signal under the last one preset decomposition scale
And the gross energy of the reconstruction signal of all frequency signals.
First processing units are normalized for the energy to each reconstruction signal, determine the energy of each reconstruction signal
Measure shared ratio in gross energy.
First construction unit, for according to the energy of each reconstruction signal ratio shared in gross energy, structure failure letter
Number feature vector.
First computing module S404：For being calculated the fault-signal feature vector using possibility clustering algorithm,
Determine fault type corresponding with fault-signal.
In embodiments of the present invention, first computing module S404 may include：
First acquisition unit：Initial parameter for obtaining possibility clustering algorithm.
Wherein, which includes：Cluster centre number C, weighted index m, initial cluster center matrix V and initially it is subordinate to
Spend matrix U.
Second computing unit, for calculating each sample determined by fault-signal feature vector to initial cluster center
The distance of matrix.
In embodiments of the present invention, each sample in fault-signal feature vector can be calculated according to above-mentioned formula (5)
To the distance in initial cluster center matrix.
First iteration unit is used for the distance of each sample to initial cluster center and initial subordinated-degree matrix
In with this apart from corresponding degree of membership substitute into degree of membership iterative formula, calculate the subordinated-degree matrix after iteration.
Wherein it is possible to calculate the subordinated-degree matrix after iteration according to above-mentioned formula (6) and (7).
First judging unit：For judging whether the difference of subordinated-degree matrix and initial subordinated-degree matrix after iteration is less than
The default limits of error.
First determination unit：It is when being, by the subordinated-degree matrix after iteration for the judging result in the first judging unit
As target subordinated-degree matrix.
Second judgment unit：For the judging result in the first judging unit be it is no when, by the subordinated-degree matrix after iteration
As initial subordinated-degree matrix, and judge whether current iteration number reaches default maximum iteration.
First trigger element：For the judging result in second judgment unit be it is no when, trigger the first iteration unit.
Second determination unit：For the judging result in second judgment unit be when, current iteration number is corresponding
Subordinated-degree matrix is as target subordinated-degree matrix.
Secondary iteration unit：For in fault-signal feature vector each sample and target subordinated-degree matrix in
The corresponding degree of membership of the sample substitutes into cluster centre iterative formula, calculates target cluster centre matrix.
Specifically, in embodiments of the present invention, target cluster centre matrix can be calculated according to above-mentioned formula (8).
Third computing unit：Data sample for calculating unknown failure type is respectively clustered with target cluster centre matrix
The approach degree at center.
Wherein it is possible to calculate the data sample and target cluster centre matrix of unknown failure type according to above-mentioned formula (9)
In each cluster centre approach degree.
Third determination unit determines failure corresponding with fault-signal for the result of calculation according to third computing unit
Type.
In embodiments of the present invention, it is adopted when the first acquisition module is got when acquisition module is monitored air-conditioning system
After the fault-signal of collection, the fault-signal is handled using wavelet package transforms algorithm by first processing module, to obtain
The reconstruction signal of each frequency signal in the fault-signal, and determine the use of the reconstruction signal by first and can be obtained fault-signal spy
Sign vector, later, the first computing module can be used possibility clustering algorithm and calculate the fault-signal feature vector, from
And quickly and accurately determine fault type corresponding with the fault message, so as to according to the fault type to the city rail vehicle
The failure of air-conditioning system is accurately and quickly handled, and is made the fast quick-recovery normal work of the air-conditioning system, is ensured occupant comfort
By bus environment.
It should be noted that the present invention can realize the event of the city rail vehicle air-conditioning system described in above-described embodiment by computer
Hinder diagnostic method, moreover, each module that the city rail vehicle air-conditioning system trouble-shooter described in above-described embodiment is included or
Unit is the name carried out according to its function, can also be by above-mentioned city rail vehicle air-conditioning system failure in addition to above-mentioned fractionation mode
The specific steps of diagnostic method are split again, and the life of device comprising modules or unit is carried out according to the function of every part of fractionation
Name, the invention is not limited in a kind of above-mentioned dividing modes.
In addition, in city rail vehicle air-conditioning system trouble-shooter described in the above embodiment of the present invention, in addition to above-mentioned
Module or unit can also include：Display module for showing determining fault type, or for reporting determining failure
The voice module of type, and the interface unit etc. for connecting each module or unit, the present invention will be arranged no longer one by one herein
It lifts, only if it were not for those skilled in the art are by creating labour determination, belongs to the scope of the present invention.
Each embodiment is described by the way of progressive in this specification, the highlights of each of the examples are with other
The difference of embodiment, just to refer each other for identical similar portion between each embodiment.For device disclosed in embodiment
For, since it is corresponded to the methods disclosed in the examples, so description is fairly simple, related place is said referring to method part
It is bright.
The foregoing description of the disclosed embodiments enables those skilled in the art to implement or use the present invention.
Various modifications to these embodiments will be apparent to those skilled in the art, as defined herein
General Principle can be realized in other embodiments without departing from the spirit or scope of the present invention.Therefore, of the invention
It is not intended to be limited to the embodiments shown herein, and is to fit to and the principles and novel features disclosed herein phase one
The widest range caused.
Claims (7)
1. a kind of city rail vehicle air-conditioning system method for diagnosing faults, which is characterized in that the method includes：
Obtain the fault-signal that acquisition module is acquired when being monitored to air-conditioning system, the acquisition module is specially and air-conditioning
The connected sensor of various parts, the type of the sensor determine that the fault-signal is specially electricity according to connected components
Press signal or current signal；
WAVELET PACKET DECOMPOSITION is carried out to the fault-signal using preset decomposition scale, obtains each frequency letter in the fault-signal
Number；
It extracts the signal characteristic under each preset decomposition scale in each frequency signal and makees each frequency letter under corresponding preset decomposition scale
Number reconstruction signal；
Using the reconstruction signal, fault-signal feature vector is obtained；The initial parameter of possibility clustering algorithm is obtained, it is described first
Beginning parameter includes：Cluster centre number C, weighted index, initial cluster center matrix and initial subordinated-degree matrix；Wherein, initial poly-
Class center matrix is made of C cluster centre；
It calculates each poly- in each sample to the initial cluster center matrix determined by the fault-signal feature vector
The distance at class center；
By the distance of each cluster centre in each described sample to the initial cluster center matrix and the initial person in servitude
Degree of membership iterative formula is substituted into apart from corresponding degree of membership with described in category degree matrix, calculates the subordinated-degree matrix after iteration；
Judge whether the difference of the subordinated-degree matrix after the iteration and the initial subordinated-degree matrix is maximum allowable less than default
Error；
If it is, using the subordinated-degree matrix after the iteration as target subordinated-degree matrix；
If it is not, then using the subordinated-degree matrix after the iteration as initial subordinated-degree matrix, and judge that current iteration number is
It is no to reach default maximum iteration；
If not up to, returned described by each cluster centre in described each sample to the initial cluster center matrix
Distance and the initial subordinated-degree matrix in it is described apart from corresponding degree of membership substitute into degree of membership iterative formula, calculate
Subordinated-degree matrix step after iteration continues to execute；
If reached, using the corresponding subordinated-degree matrix of current iteration number as target subordinated-degree matrix；
By in the fault-signal feature vector each sample and the target subordinated-degree matrix in the sample pair
The degree of membership answered substitutes into cluster centre iterative formula, calculates target cluster centre matrix；
Calculate the approach degree of the data sample and each cluster centre in the target cluster centre matrix of unknown failure type；
According to result of calculation, fault type corresponding with the fault-signal is determined；
Processing routine corresponding with the fault type is obtained, and executes the processing routine, to exclude the air-conditioning system
Failure.
2. according to the method described in claim 1, it is characterized in that, the utilization reconstruction signal, obtains fault-signal spy
Sign vector, including：
Calculate the weight of the energy of the reconstruction signal of each frequency signal and all frequency signals under the last one preset decomposition scale
The gross energy of structure signal；
The energy of the reconstruction signal is normalized, determines that the energy of each reconstruction signal is shared in the gross energy
Ratio；
According to the energy of each reconstruction signal ratio shared in the gross energy, fault-signal feature vector is built.
3. according to the method described in claim 2, it is characterized in that, described will be in described each sample to the initial clustering
It is substituted into apart from corresponding degree of membership with described in the distance of each cluster centre and the initial subordinated-degree matrix in heart matrix
Degree of membership iterative formula calculates the subordinated-degree matrix after iteration, including：
By the distance of each cluster centre in each described sample to the initial cluster center matrix and the initial person in servitude
Formula is determined apart from corresponding degree of membership substitution adjustment parameter, determine with described apart from corresponding adjusting in category degree matrix with described
Parameter ηi, the adjustment parameter determines that formula is：
Wherein, i=1,2 ..., the preset decomposition scale parameter；J=1,2 ..., n, n indicate each under each preset decomposition scale
The dispersion index that node calculates；uijIndicate in initial subordinated-degree matrix the corresponding person in servitude of each node under each preset decomposition scale
Category degree, dijIndicate each sample to the initial cluster center distance；M indicates the weighted index；K=1；
By the distance of each cluster centre in each described sample to the initial cluster center matrix and with the distance
Corresponding adjustment parameter substitutes into degree of membership iterative formula, calculates the subordinated-degree matrix U after iteration、, the degree of membership iteration public affairs
Formula is：
Wherein, u、 ijIndicate degree of membership corresponding with each node under each preset decomposition scale in the subordinated-degree matrix after iteration.
4. according to the method described in claim 2, it is characterized in that, the cluster centre iterative formula is：
Wherein, u、 ijIndicate degree of membership corresponding with each node under each preset decomposition scale in the subordinated-degree matrix after iteration；I=
1,2 ..., the preset decomposition scale parameter；J=1,2 ..., n；N indicates that each node is computed under each preset decomposition scale
The dispersion index arrived；M indicates the weighted index, xjIndicate j-th of sample in the fault-signal feature vector.
5. according to the method described in claim 2, it is characterized in that, it is described calculate unknown failure type data sample with it is described
The approach degree of each cluster centre in target cluster centre matrix, including：
According to formulaCalculate the data sample of unknown failure type and the mesh
Mark the approach degree σ (Data, V) of each cluster centre in cluster centre matrix；
Wherein, min expressions are minimized operation；Max expressions are maximized operation；dataikIndicate the data of unknown failure type
Sample, v、 jkIndicate each cluster centre in the target cluster centre matrix；I=1,2 ..., n；J=1,2 ..., C；N is indicated
Indicate the dispersion index that each node calculates under each preset decomposition scale；C indicates that the cluster centre number, m indicate
The weighted index.
6. a kind of city rail vehicle air-conditioning system trouble-shooter, which is characterized in that described device includes：
First acquisition module, it is described to adopt for obtaining the fault-signal acquired when acquisition module is monitored air-conditioning system
Collection module is specially the sensor being connected with each component of air-conditioning system, and the type of the sensor is determined according to connected components, institute
It is specially voltage signal or current signal to state fault-signal；
First processing module handles the fault-signal, obtains the failure letter for using wavelet package transforms algorithm
The reconstruction signal of each frequency signal in number；
First determining module obtains fault-signal feature vector for utilizing the reconstruction signal；First computing module, is used for
The fault-signal feature vector is calculated using possibility clustering algorithm, determines failure corresponding with the fault-signal
Type, and processing routine corresponding with the fault type is obtained, the processing routine is executed, to exclude the air-conditioning system
Failure；
Wherein, the first processing module includes：
First resolving cell obtains the event for carrying out WAVELET PACKET DECOMPOSITION to the fault-signal using preset decomposition scale
Hinder each frequency signal in signal；
First extraction unit makees corresponding default point for extracting the signal characteristic under each preset decomposition scale in each frequency signal
Solve the reconstruction signal of each frequency signal under scale；
First computing module includes：
First acquisition unit：Initial parameter for obtaining possibility clustering algorithm, the initial parameter include：Cluster centre number
C, weighted index, initial cluster center matrix and initial subordinated-degree matrix；Wherein, initial cluster center matrix includes poly- by C
Class center forms；
Second computing unit, for calculating each sample determined by the fault-signal feature vector to initial cluster center
The distance of each cluster centre in matrix；
First iteration unit, for by each described sample to the initial cluster center matrix each cluster centre away from
From and the initial subordinated-degree matrix in it is described with a distance from corresponding degree of membership substitute into degree of membership iterative formula, calculate iteration
Subordinated-degree matrix afterwards；
First judging unit, for judge the subordinated-degree matrix after the iteration and the initial subordinated-degree matrix difference whether
Less than the default limits of error；
First determination unit is when being, by the degree of membership after the iteration for the judging result in first judging unit
Matrix is as target subordinated-degree matrix；
Second judgment unit is used for when the judging result of first judging unit is no, by the degree of membership after the iteration
Matrix judges whether current iteration number reaches default maximum iteration as initial subordinated-degree matrix；
First trigger element, for when the judging result of the second judgment unit is no, triggering the first iteration unit；
Second determination unit, for the judging result in the second judgment unit be when, current iteration number is corresponding
Subordinated-degree matrix is as target subordinated-degree matrix；
Secondary iteration unit, for by the fault-signal feature vector each sample and the target degree of membership square
Degree of membership corresponding with the sample substitutes into cluster centre iterative formula in battle array, calculates target cluster centre matrix；
Third computing unit, the data sample for calculating unknown failure type are respectively clustered with the target cluster centre matrix
The approach degree at center；
Third determination unit, for the result of calculation according to the third computing unit, determination is corresponding with the fault-signal
Fault type.
7. device according to claim 6, which is characterized in that first determining module includes：
First computing unit, for calculate under the last one preset decomposition scale the energy of the reconstruction signal of each frequency signal and
The gross energy of the reconstruction signal of all frequency signals；
First processing units are normalized for the energy to the reconstruction signal, determine the energy of each reconstruction signal
The shared ratio in the gross energy；
First construction unit, for according to the energy of each reconstruction signal ratio shared in the gross energy, structure event
Hinder signal characteristic vector.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410455410.5A CN104200109B (en) | 2014-09-09 | 2014-09-09 | A kind of city rail vehicle air-conditioning system method for diagnosing faults and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410455410.5A CN104200109B (en) | 2014-09-09 | 2014-09-09 | A kind of city rail vehicle air-conditioning system method for diagnosing faults and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104200109A CN104200109A (en) | 2014-12-10 |
| CN104200109B true CN104200109B (en) | 2018-09-28 |
Family
ID=52085402
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410455410.5A Active CN104200109B (en) | 2014-09-09 | 2014-09-09 | A kind of city rail vehicle air-conditioning system method for diagnosing faults and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104200109B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105718733B (en) * | 2016-01-21 | 2018-04-13 | 福建师范大学 | Fault prediction method based on fuzzy nearness and particle filter |
| CN109345658A (en) * | 2018-10-29 | 2019-02-15 | 百度在线网络技术（北京）有限公司 | Restorative procedure, device, equipment, medium and the vehicle of Vehicular system failure |
| CN110059833B (en) * | 2019-02-27 | 2021-06-18 | 山东开创云计算有限公司 | Real-time maintenance method, device and system for shared automobile |
| CN111539478A (en) * | 2020-04-26 | 2020-08-14 | 成都见田科技有限公司 | Intelligent diagnosis method, system and diagnosis equipment for elevator faults |
| CN113466566B (en) * | 2021-05-26 | 2024-02-09 | 山东英信计算机技术有限公司 | Trigger type signal acquisition method and device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102680017A (en) * | 2012-05-31 | 2012-09-19 | 潍柴动力股份有限公司 | Fault diagnostic method and diagnostic device of sensor |
| CN102944418A (en) * | 2012-12-11 | 2013-02-27 | 东南大学 | Wind turbine generator group blade fault diagnosis method |
| CN103886316A (en) * | 2014-02-20 | 2014-06-25 | 东南大学 | Combustion monitoring and diagnosis method based on feature extraction and fuzzy C-means cluster |
-
2014
- 2014-09-09 CN CN201410455410.5A patent/CN104200109B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102680017A (en) * | 2012-05-31 | 2012-09-19 | 潍柴动力股份有限公司 | Fault diagnostic method and diagnostic device of sensor |
| CN102944418A (en) * | 2012-12-11 | 2013-02-27 | 东南大学 | Wind turbine generator group blade fault diagnosis method |
| CN103886316A (en) * | 2014-02-20 | 2014-06-25 | 东南大学 | Combustion monitoring and diagnosis method based on feature extraction and fuzzy C-means cluster |
Non-Patent Citations (3)
| Title |
|---|
| 可能性聚类算法研究;周巧萍;《中国优秀硕士学位论文全文数据库信息科技辑》;20081115(第11期);正文第13-15页 * |
| 基于小波分析的风机故障诊断;胡汉辉等;《中南大学学报(自然科学版)》;20071231;第38卷(第6期);1169-1173 * |
| 模糊数学在设备故障诊断领域的应用及常用方法;王书纯等;《湖北师范学院学报(自然科学版)》;20050331;第25卷(第3期);76页第2节-77页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104200109A (en) | 2014-12-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104200109B (en) | A kind of city rail vehicle air-conditioning system method for diagnosing faults and device | |
| CN106596123A (en) | Device fault diagnosis method, device fault diagnosis device, and device fault diagnosis system | |
| CN106662072B (en) | Wind-driven generator method for monitoring state and system | |
| CN107599783A (en) | A kind of environment inside car management system and its control method | |
| CN107806690B (en) | A kind of domestic air conditioning method for diagnosing faults based on Bayesian network | |
| CN204693738U (en) | Air conditioner with separate hot pump actual motion energy efficiency detection system | |
| CN108285071A (en) | A kind of elevator Gernral Check-up method based on Bayesian network | |
| CN106338406A (en) | On-line monitoring and fault early-warning system and method for traction electric transmission system of train | |
| CN103901882A (en) | Online monitoring fault diagnosis system and method of train power system | |
| CN105973626A (en) | Evaluation and prediction method and apparatus for operation energy efficiency of host of central air-conditioning system | |
| CN106021771A (en) | Method and device for diagnosing faults | |
| CN104133143B (en) | A kind of Guangdong power system diagnostic system and method calculating platform based on Hadoop cloud | |
| CN106352475A (en) | Training sample collection method and device of air conditioner neutral network and air conditioning system | |
| CN103345637A (en) | On-line detecting system and method of hydrophobicity of composite insulator of electric transmission line | |
| CN110163075A (en) | A kind of multi-information fusion method for diagnosing faults based on Weight Training | |
| CN110503153A (en) | Photovoltaic system method for diagnosing faults based on differential evolution algorithm and support vector machines | |
| CN110850717B (en) | Neural network heat pump defrosting control device and control method by utilizing fan current | |
| CN108444056A (en) | Central air-conditioning monitoring system and its monitoring method | |
| CN110823291A (en) | Method and system for monitoring indoor temperature and humidity environment of building based on K-means clustering algorithm | |
| CN110779264A (en) | Defrosting control system and control method based on fan vibration and information fusion | |
| CN207225013U (en) | A kind of car inside health environment intelligent management platform and automobile | |
| CN110221173A (en) | A kind of power distribution network intelligent diagnosing method based on big data driving | |
| CN106502121A (en) | A kind of intelligent home control system | |
| CN117196349A (en) | Inpatient bed information intelligent management system | |
| CN108805192B (en) | Monitoring data analysis method based on hierarchical network structure |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |