CN105782071B - A kind of water-isolation slurry pump method for diagnosing faults based on probabilistic neural network - Google Patents
A kind of water-isolation slurry pump method for diagnosing faults based on probabilistic neural network Download PDFInfo
- Publication number
- CN105782071B CN105782071B CN201610123651.9A CN201610123651A CN105782071B CN 105782071 B CN105782071 B CN 105782071B CN 201610123651 A CN201610123651 A CN 201610123651A CN 105782071 B CN105782071 B CN 105782071B
- Authority
- CN
- China
- Prior art keywords
- water
- neural network
- slurry pump
- probabilistic neural
- class
- 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
Classifications
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F04—POSITIVE - DISPLACEMENT MACHINES FOR LIQUIDS; PUMPS FOR LIQUIDS OR ELASTIC FLUIDS
- F04D—NON-POSITIVE-DISPLACEMENT PUMPS
- F04D15/00—Control, e.g. regulation, of pumps, pumping installations or systems
- F04D15/0077—Safety measures
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F05—INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
- F05D—INDEXING SCHEME FOR ASPECTS RELATING TO NON-POSITIVE-DISPLACEMENT MACHINES OR ENGINES, GAS-TURBINES OR JET-PROPULSION PLANTS
- F05D2260/00—Function
- F05D2260/80—Diagnostics
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F05—INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
- F05D—INDEXING SCHEME FOR ASPECTS RELATING TO NON-POSITIVE-DISPLACEMENT MACHINES OR ENGINES, GAS-TURBINES OR JET-PROPULSION PLANTS
- F05D2270/00—Control
- F05D2270/70—Type of control algorithm
- F05D2270/709—Type of control algorithm with neural networks
Landscapes
- Engineering & Computer Science (AREA)
- Mechanical Engineering (AREA)
- General Engineering & Computer Science (AREA)
- Structures Of Non-Positive Displacement Pumps (AREA)
Abstract
The invention discloses a kind of water-isolation slurry pump method for diagnosing faults based on probabilistic neural network, includes the following steps：(1) fault category of water-isolation slurry pump is divided into m classes；(2) N of the i-th class is acquirediGroup training sample data；(3) training sample data are normalized, obtain input vector X；(4) X is connected to mode layer, obtains the output of mode layer；(5) the same class variable that mode layer is sent is added up and is summed by summation layer, obtains the probability density value of the i-th class；(6) trained probabilistic neural network model is obtained；(7) current gathered data is inputted into the probabilistic neural network model, obtains the current failure classification output of water-isolation slurry pump.The configuration of the present invention is simple, fast convergence rate, training time are short, it is not easy to converge to part, stability is high, and sample supplemental capabilities are strong.
Description
Technical field
The invention belongs to the fault diagnosis field of water-isolation slurry pump, more particularly to a kind of water based on probabilistic neural network
Ore slurry pump method for diagnosing faults is isolated.
Background technology
Currently, various mine resources are in remote mountain areas, ore, which is transported to smeltery, needs long-distance transportation.Ore is defeated
It is one of important technique in ore dressing process to send, and is generally conveyed using water-isolation slurry pump.
The conveying principle of water-isolation slurry pump is that solid matter (ore) and liquid (clear water) are mutually mixed into ore pulp,
Power is provided using clarified water pump 11, is conveyed in slurry transportation pipeline.This method is transported relative to railway transportation and highway
The defeated advantage in cost, while having the characteristics that energy conservation and environmental protection.
The structure of water-isolation slurry pump is as shown in Figure 1.Water-isolation slurry pump includes clear water reserviors 12, clarified water pump equipped with clear water
11, inlet valve 10, back-water valve (BWV) 9, sealing chamber 5, the first shut-off valve 3, the second shut-off valve 4 and slurry transportation pipeline 2；Wherein sealing chamber
5 are internally provided with floating ball 6 moving up and down, and sealing chamber 5 is separated into superposed first cavity 7 and is located at by the floating ball 6
Second cavity 8 of lower part, the first cavity 7 and the second cavity 8 are not connected；The outlet of clear water reserviors 12 passes sequentially through 11 He of clarified water pump
Inlet valve 10 is communicated with the first cavity 7, and the entrance of clear water reserviors 12 is communicated by back-water valve (BWV) 9 with the first cavity 7；First shut-off valve 3
Outlet and the entrance of the second shut-off valve 4 communicated with the second cavity 8, the entrance of the first shut-off valve 3 and the second shut-off valve 4 go out
Mouth is communicated with slurry transportation pipeline 2；One end of slurry transportation pipeline 2 is communicated with the Slurry Bin 1 equipped with ore pulp, and the other end leads to
The destination of ores lifting.
When conveying ore, power is provided by clarified water pump 11, so that clear water is isolated with ore pulp by sealing chamber 5.Suck ore pulp
When, inlet valve 10 is closed, opens back-water valve (BWV) 9, ore pulp is pressed into sealing chamber 5 under the effect of gravity, from the first shut-off valve 3, on floating ball 6
It rises, the clear water in the first cavity 7 is pressed into clear water reserviors 12 by back-water valve (BWV) 9.When ore pulp is discharged, back-water valve (BWV) 9 is closed, opens water inlet
Clear water in clear water reserviors 12 is squeezed into the first cavity 7 with certain pressure, floating ball 6 is pushed to pass through ore pulp by valve 10, clarified water pump 11
Second shut-off valve 4 is pressed into slurry transportation pipeline 2, is then delivered to destination by slurry transportation pipeline 2.Whole work process
In, the opening and closing of inlet valve 10 and back-water valve (BWV) 9 is controlled by program, realizes the sucking ore pulp of water-isolation slurry pump and ore pulp is discharged
Process.
Water-isolation slurry pump is the important equipment in ore slurry pipeline transmission process, ore pulp can be caused to exist when it breaks down
Precipitation, can cause ore pulp delivery duct to block when serious in ore pulp delivery duct.Remote monitoring is carried out to water-isolation slurry pump, is examined
Its disconnected operating status, remedial measure can be made early when breaking down by being conducive to staff, and it is such to reduce line clogging
The possibility that major accident occurs.
Water-isolation slurry pump is a complicated nonlinear system, it is difficult to accurate mathematical model be set up, because of the invention
A kind of accuracy is high, can the method for diagnosing faults of quick diagnosis water-isolation slurry pump be of great significance.
The method for diagnosing faults of water-isolated slurry pump is the diagnostic method based on BP neural network at present, due to BP nerve nets
Network algorithm the convergence speed is slow, needs the longer training time；It is easy to converge to local minimum point, leads to failure to train；Study and
Memory has unstability, and strong to the dependence of sample, if increasing learning sample, trained network just needs from the beginning to open
Begin to train, pervious weights and threshold value are not remembered.
Probabilistic neural network is the neural network of one mode classification, is a kind of feedforward neural network.With BP neural network
It compares, probabilistic neural network is simple in structure, fast convergence rate, and the training time is short, it is not easy to converge to part, stability is high, sample
This supplemental capabilities is strong.The present invention is exactly from the advantages of probabilistic neural network, obtaining a kind of water based on probabilistic neural network
Ore slurry pump method for diagnosing faults is isolated.
It is illustrated in figure 2 the basic structure of probabilistic neural network.Probabilistic neural network is by input layer I, mode layer II, summation
Layer III and output layer IV form.Input vector is sent after treatment in input layer I to mode layer II.The neural network of mode layer II
Number is equal to the sum of the number of training of each classification, wherein gijFor j-th of the i-th class output of mode layer II, i=1,2 ...,
m；J=1,2 ... Ni, i is fault category number, and m is fault category total number, NiFor the total number of the training sample of the i-th class.It asks
It is equal to fault category number with the neural network number of layer III, summation layer III distinguishes each class probability density from mode layer II
Phase adduction is sent after averaging to output layer IV.Output layer IV receives all kinds of probability density that summation layer III exports and judges to obtain defeated
Do well classification.
Invention content
Since existing water-isolation slurry pump method for diagnosing faults is all based on BP neural network algorithm, and BP neural network
Algorithm the convergence speed is slow, and the training time is long；It is easy to converge to local minimum point, leads to failure to train；Learning and memory has not
Stability is strong to the dependence of sample.It is an object of the present invention in view of the above shortcomings of the prior art, provide one kind and be based on
The water-isolation slurry pump method for diagnosing faults of probabilistic neural network.
In order to solve the above technical problems, the technical solution adopted in the present invention is：
A kind of water-isolation slurry pump method for diagnosing faults based on probabilistic neural network, includes the following steps：
The fault category of water-isolation slurry pump is divided into m classes by step (1), m >=2, and one kind in the m classes represents normal shape
State, remaining class representing fault state；
Step (2) acquires the N of the i-th classiGroup training sample data, wherein i are fault category, i=1,2 ..., m；
Step (3) is to the NiGroup training sample data are normalized, and obtain input vector X；
X is connected to mode layer by step (4), obtains the output of mode layer
Wherein gij(X) it is j-th of output of the i-th class of mode layer, p is the dimension of input vector, and σ is smoothing parameter, Xij
For the weights of j-th of sample vector of the i-th class in a network；
The same class variable that mode layer is sent is added up and is summed by step (5) summation layer, and the probability for obtaining the i-th class is close
Angle value is
Step (6) output layer receives the probability density value of summation layer, and passes through formula max (fi(X)) water segregation mine is obtained
The sample fault category of stock pump exports theoretical value；If sample fault category exports theoretical value and exports actual value with sample fault category
It is inconsistent, then change XijSize, until to export actual value consistent for sample fault category output theoretical value and sample fault category；
Store XijValue, obtain trained probabilistic neural network model；
Current gathered data is inputted the probabilistic neural network model by step (7), obtains the current of water-isolation slurry pump
Fault category exports.
The fault category in the step (1) includes normal condition, outer tube blocked state and water as a preferred method,
Pump 3 class of state of wear.
Probabilistic neural network is a kind of neural network for pattern classification, for be only applicable to when fault diagnosis field can
With the model classified to failure, the fault category of water-isolation slurry pump is divided into normally by the present invention by a large amount of experiment
3 class of state, outer tube blocked state and water pump state of wear, achieves good fault diagnosis effect.
The normalization processing method in the step (3) is Z-score standardized methods as a preferred method,.
The training sample data and current gathered data include clear water pump motor current, slurry as a preferred method,
Plasma discharge amount, the pressure of slurry transportation pipeline of body conveyance conduit.
σ=0.1 as a preferred method,.
Compared with prior art, the configuration of the present invention is simple, fast convergence rate, training time are short, it is not easy to part is converged to,
Stability is high, and sample supplemental capabilities are strong.
Description of the drawings
Fig. 1 is the structural schematic diagram of water-isolation slurry pump.
Fig. 2 is the basic structure of probabilistic neural network.
Fig. 3 is the diagnostic result of training data.
Fig. 4 is the diagnostic result of verify data.
Wherein, 1 is Slurry Bin, and 2 be slurry transportation pipeline, and 3 be the first shut-off valve, and 4 be the second shut-off valve, and 5 be sealing chamber,
6 be floating ball, and 7 be the first cavity, and 8 be the second cavity, and 9 be back-water valve (BWV), and 10 be inlet valve, and 11 clarified water pumps, 12 be clear water reserviors, and I is
Input layer, II is mode layer, and III is summation layer, and IV is output layer.
Specific implementation mode
One embodiment of the present invention includes the following steps：
The fault category of water-isolation slurry pump is divided into m classes by step (1), m >=2, and one kind in the m classes represents normal shape
State, remaining class representing fault state；
Step (2) acquires the N of the i-th classiGroup training sample data, wherein i are fault category, i=1,2 ..., m；
Step (3) is to the NiGroup training sample data carry out Z-score and standardize normalized, obtain input vector
X；
X is connected to mode layer by step (4), obtains the output of mode layer
Wherein gij(X) be mode layer j-th of the i-th class output, p is the dimension of input vector, and σ is smoothing parameter, σ=
0.1, XijFor the weights of j-th of sample vector of the i-th class in a network；
The same class variable that mode layer is sent is added up and is summed by step (5) summation layer, and the probability for obtaining the i-th class is close
Angle value is
Step (6) output layer receives the probability density value of summation layer, and passes through formula max (fi(X)) water segregation mine is obtained
The sample fault category of stock pump exports theoretical value；If sample fault category exports theoretical value and exports actual value with sample fault category
It is inconsistent, then change XijSize, until to export actual value consistent for sample fault category output theoretical value and sample fault category；
Store XijValue, obtain trained probabilistic neural network model；
Current gathered data is inputted the probabilistic neural network model by step (7), obtains the current of water-isolation slurry pump
Fault category exports.
The training sample data and current gathered data include the plasma discharge of clear water pump motor current, slurry transportation pipeline
Amount, the pressure of slurry transportation pipeline.
In order to verify the feasibility of the present invention, in test the operating status of water-isolation slurry pump is carried out using the present invention
Diagnosis.In experiment, the fault category of water-isolation slurry pump is divided into normal condition, outer tube blocked state and water pump state of wear 3
Class, wherein normal condition number are state 1, and outer tube blocked state number is state 2, and water pump state of wear number is state 3, if
Corresponding status number is successively increased when there are other states.
Acquire the data of 5 groups of normal operating conditions, the data of 5 groups of outer tube blocked states, the number of 5 groups of water pump state of wear
According to data value is as shown in table 1.
The training of 1 water-isolated slurry pump of table and verify data
Every group of state takes the training of the probabilistic neural network diagnostic model of 4 groups of carry out water-isolation slurry pumps, uses each
The remaining one group of data of operating status carry out model verification.The diagnostic result of training data such as Fig. 3, the diagnostic result of verify data
Such as Fig. 4, as can be seen from Figure, using the present invention, the accuracy of water-isolation slurry pump fault diagnosis is up to 100%.
Claims (5)
1. a kind of water-isolation slurry pump method for diagnosing faults based on probabilistic neural network, which is characterized in that include the following steps：
The fault category of water-isolation slurry pump is divided into m classes by step (1), m >=2, and one kind in the m classes represents normal condition,
Remaining class representing fault state；
Step (2) acquires the N of the i-th classiGroup training sample data, wherein i are fault category, i=1,2 ..., m；
Step (3) is to the NiGroup training sample data are normalized, and obtain input vector X；
X is connected to mode layer by step (4), obtains the output of mode layer
Wherein gij(X) it is j-th of output of the i-th class of mode layer, p is the dimension of input vector, and σ is smoothing parameter, XijIt is i-th
The weights of j-th of sample vector of class in a network；
The same class variable that mode layer is sent is added up and is summed by step (5) summation layer, obtains the probability density value of the i-th class
For
Step (6) output layer receives the probability density value of summation layer, and passes through formula max (fi(X)) water-isolation slurry pump is obtained
Sample fault category exports theoretical value；Differ with sample fault category output actual value if sample fault category exports theoretical value
It causes, then changes XijSize, until to export actual value consistent for sample fault category output theoretical value and sample fault category；Storage
XijValue, obtain trained probabilistic neural network model；
Current gathered data is inputted the probabilistic neural network model by step (7), obtains the current failure of water-isolation slurry pump
Classification exports.
2. the water-isolation slurry pump method for diagnosing faults based on probabilistic neural network as described in claim 1, which is characterized in that
Fault category in the step (1) includes 3 class of normal condition, outer tube blocked state and water pump state of wear.
3. the water-isolation slurry pump method for diagnosing faults based on probabilistic neural network as described in claim 1, which is characterized in that
Normalization processing method in the step (3) is Z-score standardized methods.
4. the water-isolation slurry pump method for diagnosing faults based on probabilistic neural network as described in claim 1, which is characterized in that
The training sample data and current gathered data include that clear water pump motor current, the plasma discharge amount of slurry transportation pipeline, slurry are defeated
Send the pressure of pipeline.
5. the water-isolation slurry pump method for diagnosing faults based on probabilistic neural network as described in claim 1, which is characterized in that
σ=0.1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610123651.9A CN105782071B (en) | 2016-03-04 | 2016-03-04 | A kind of water-isolation slurry pump method for diagnosing faults based on probabilistic neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610123651.9A CN105782071B (en) | 2016-03-04 | 2016-03-04 | A kind of water-isolation slurry pump method for diagnosing faults based on probabilistic neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105782071A CN105782071A (en) | 2016-07-20 |
| CN105782071B true CN105782071B (en) | 2018-07-13 |
Family
ID=56387521
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610123651.9A Active CN105782071B (en) | 2016-03-04 | 2016-03-04 | A kind of water-isolation slurry pump method for diagnosing faults based on probabilistic neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105782071B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106650932B (en) * | 2016-12-23 | 2019-05-28 | 郑州云海信息技术有限公司 | A kind of the intelligent trouble classification method and device of data center's monitoring system |
| CN107563069A (en) * | 2017-09-06 | 2018-01-09 | 国电联合动力技术有限公司 | A kind of wind power generating set intelligent fault diagnosis method |
| CN108803555B (en) * | 2018-03-20 | 2020-08-21 | 北京航空航天大学 | Sub-health online identification and diagnosis method based on performance monitoring data |
| CN108953172A (en) * | 2018-08-10 | 2018-12-07 | 湖南柿竹园有色金属有限责任公司 | A kind of separate pump automatic control system |
| CN109063785B (en) * | 2018-08-23 | 2021-03-16 | 国网河北省电力有限公司沧州供电分公司 | Charging pile fault detection method and terminal equipment |
| CN109751173A (en) * | 2019-01-16 | 2019-05-14 | 哈尔滨理工大学 | Hydraulic turbine operation method for diagnosing faults based on probabilistic neural network |
| CN109978048B (en) * | 2019-03-22 | 2021-05-07 | 大唐环境产业集团股份有限公司 | Fault analysis and diagnosis method for slurry circulating pump of desulfurizing tower |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201496284U (en) * | 2009-08-20 | 2010-06-02 | 楼生高 | Submerged pump controller |
| CN101666315B (en) * | 2009-09-30 | 2011-09-21 | 杨爱民 | Tubular shaft type centrifugal pump |
| CN102900675A (en) * | 2012-11-09 | 2013-01-30 | 昆山北极光电子科技有限公司 | System for detecting working condition and diagnosing fault of centrifugal pump |
| CN103886405B (en) * | 2014-02-20 | 2017-02-22 | 东南大学 | Boiler combustion condition identification method based on information entropy characteristics and probability nerve network |
| CN105134619B (en) * | 2015-09-28 | 2017-03-01 | 北京航空航天大学 | A kind of fault diagnosis based on wavelet energy, manifold dimension-reducing and dynamic time warping and health evaluating method |
-
2016
- 2016-03-04 CN CN201610123651.9A patent/CN105782071B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN105782071A (en) | 2016-07-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105782071B (en) | A kind of water-isolation slurry pump method for diagnosing faults based on probabilistic neural network | |
| CN102169077B (en) | Hybrid intelligent soft measuring method of overflow granularity index in wet grinding process | |
| CN106523260B (en) | A kind of sequence of Hydropower Plant efficiency and load distribution method based on guide vane opening | |
| CN103544389A (en) | Fault tree and fuzzy neural network based automobile crane fault diagnosis method | |
| CN102633124B (en) | Saltation ash conveying control method | |
| CN201842507U (en) | Concentrated phase pneumatic conveying self-diagnosis self-removal dredging system | |
| CN208254546U (en) | A kind of geothermal tail water recharge test macro | |
| CN206146768U (en) | Three phase flow erosion corrosion test device | |
| CN209622502U (en) | A kind of pipe leakage data acquisition device and detection positioning device | |
| CN209344267U (en) | Sour cycle battery formation device | |
| CN106382467A (en) | Long-distance liquid oxygen supply system and precooling method thereof | |
| CN106285503B (en) | Acid gas well flow string inner wall medium corrosion-resistant automatic painting apparatus and method | |
| CN115526196A (en) | Method for dynamically diagnosing working condition of electric submersible pump on line based on real-time data and model | |
| CN104295892B (en) | CO2Fluid reservoir supercharge control system | |
| CN207962092U (en) | A kind of food-grade carbon-dioxide charging device with security control function | |
| CN206309289U (en) | Acid gas well flow string inwall medium corrosion-resistant automatic painting apparatus | |
| CN207406904U (en) | Slurry balance shield muddy water circulation line valve intelligent automatic-control device | |
| CN208182733U (en) | A kind of appearance water process unit | |
| CN205842219U (en) | A kind of device for slurry transportation | |
| CN205146170U (en) | Recovery tower cauldron prevent blockking up counterflush system | |
| CN211649838U (en) | Skid-mounted device and crude oil gathering and transportation equipment with same | |
| CN109657909A (en) | A kind of desulfurization based on big data grinds system ratio of water to material method of adjustment and system | |
| CN217425998U (en) | Control system for temperature of powder conveying gas | |
| CN104775799B (en) | Profile control method and system for thickening well thin oil | |
| CN219291634U (en) | Cyclone shunt system with anti-blocking function |
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 | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20190624 Address after: 410019 No. 299, Mu Lin East Road, Yuhua District, Changsha, Hunan Patentee after: Changsha Huahengyuan Information Technology Co., Ltd. Address before: No. 199 Jiefang Road, Changsha, Hunan Province Patentee before: Changsha Engineering & Research Institute of Nonferrous Metallurgy Co., Ltd. |
|
| TR01 | Transfer of patent right | ||
| CP03 | Change of name, title or address |
Address after: 410205 room a788, 7 / F, building 9, Changsha Zhongdian Software Park, No.39, Jianshan Road, Changsha hi tech Development Zone, Changsha City, Hunan Province Patentee after: Huachu Intelligent Technology (Hunan) Co., Ltd Address before: 410019 Hunan province Changsha Yuhua District, Magnolia Road No. 299 Patentee before: Changsha Huahengyuan Information Technology Co., Ltd. |
|
| CP03 | Change of name, title or address |