CN108535208A - The control method and control system of gas detecting device - Google Patents
The control method and control system of gas detecting device Download PDFInfo
- Publication number
- CN108535208A CN108535208A CN201810261277.8A CN201810261277A CN108535208A CN 108535208 A CN108535208 A CN 108535208A CN 201810261277 A CN201810261277 A CN 201810261277A CN 108535208 A CN108535208 A CN 108535208A
- Authority
- CN
- China
- Prior art keywords
- particle
- formula
- gas
- control method
- neural network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000013528 artificial neural network Methods 0.000 claims abstract description 44
- 238000005457 optimization Methods 0.000 claims abstract description 19
- 239000002245 particle Substances 0.000 claims description 56
- 230000000739 chaotic effect Effects 0.000 claims description 21
- 230000035772 mutation Effects 0.000 claims description 17
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 claims description 15
- 230000000694 effects Effects 0.000 claims description 10
- 238000011156 evaluation Methods 0.000 claims description 8
- 239000000428 dust Substances 0.000 claims description 4
- 230000008569 process Effects 0.000 claims description 3
- 230000008859 change Effects 0.000 claims description 2
- 239000003016 pheromone Substances 0.000 claims description 2
- 238000001514 detection method Methods 0.000 abstract description 17
- 230000032683 aging Effects 0.000 abstract description 4
- 238000012360 testing method Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 239000003245 coal Substances 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000001537 neural effect Effects 0.000 description 2
- 230000002028 premature Effects 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000005315 distribution function Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000012843 least square support vector machine Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 239000000843 powder Substances 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 238000005406 washing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/3504—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light for analysing gases, e.g. multi-gas analysis
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/0265—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric the criterion being a learning criterion
- G05B13/027—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric the criterion being a learning criterion using neural networks only
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Automation & Control Theory (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Biochemistry (AREA)
- Analytical Chemistry (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Chemical & Material Sciences (AREA)
- Feedback Control In General (AREA)
Abstract
A kind of control method of gas detecting device provided by the invention, includes the following steps：S100 obtains gas management signal；S200, the PSO Neural Network algorithm dynamic that Small Population optimization is carried out to the gas management signal compensate, and export gas management concentration.The control method of above-mentioned gas detecting device, the PSO Neural Network dynamic for introducing Small Population optimization compensates, it is existing since sampled analogue circuits drift, reference voltage have the problems such as fluctuation, component aging to reduce accuracy of detection to solve traditional gas indicator, improves the accuracy of detection of gas detecting device.
Description
Technical field
The present invention relates to gas management technical field, the control method more particularly to a kind of gas detecting device and control
System.
Background technology
The safety in production of underground coal mine is the major premise in entire progress of coal mining, and dangerous imflammable gas-gas
The accurate detection of gas is wherein main content.Usually gas management is infrared based on infrared-sensing type gas indicator
Gas indicator is detected gas density using the reflection of light, principle of interference, eliminates electromagnetic interference.A kind of common watt
The gas chamber of this detector, using monochromatic light road air chamber structure, infrared beam passes through gas chamber only once, although monochromatic light road gas chamber is longer
But since its simple and stable structure is reliable so having pervasive meaning.Traditionally, there are sampled analogue circuits for gas indicator
Drift, reference voltage have the problems such as fluctuation, component aging, can influence gas management precision.In addition, various powder under complex working condition
The testing conditions of dirt, gas and different humidity temperature cause the accuracy of detection of detector relatively low detector in the presence of interference.
Invention content
Based on this, it is necessary to drift about there are sampled analogue circuits for traditional gas indicator and dry be led there are various
The problem for causing the accuracy of detection of detector relatively low, provides a kind of control method and control system of gas detecting device.
A kind of control method of gas detecting device provided by the invention, includes the following steps：
S100 obtains gas management signal；
S200, the PSO Neural Network algorithm dynamic that Small Population optimization is carried out to the gas management signal compensates, defeated
Go out gas management concentration.
In one of the embodiments, tent maps are introduced in the PSO Neural Network algorithm of Small Population optimization
Neural network algorithm generates Chaos Variable.
In one of the embodiments, the model of the neural network algorithm include first layer, the second layer, third layer and
4th layer；
The first layer is input layer, and the corresponding input signal of the input layer includes the telecommunications of the gas management signal
Number, external environment temperature-humidity signal, non-methane gas and dust disturbing factor signal and electromagnetic interference signal；
The second layer is used to describe the membership function and Center Parameter of the input layer；
The third layer is used to correspond to the number of nodes of the neural network；
The described 4th layer gas density for corresponding output.
In one of the embodiments, it is described based on PSO Neural Network algorithm to the gas management signal into action
State compensation process, including following sub-step：
S210, particle populations initialization, including population total number of particles N, solution space dimension D, the PSO Neural Network
Initial value be that D ties up Chaos Variable, assign initial position, speed, inertia weight to each particle；
S220, E iteration before being carried out according to the speed of PSO Neural Network algorithm more new formula and location update formula,
The preceding F particle for selecting position optimal according to fitness value is as the center of circle for dividing Small Population and initializes F Small Population
Radius, obtain the Chaos Variable after the E times iteration according to chaotic mutation formula, E, F are positive integer；
Wherein, chaotic mutation formula is：
Z'k=β Zk+(1-β)Ω*；
In formula,Ω*For optimal solution chaos vector in D dimension spaces, value is between 0 to 1, X=
(x1,x2,...,xD) it is chaos vector, it is the vector after chaotic disturbance；β is impact factor, influence
It is degree of variation of the chaotic mutation to particle；
S230 evaluates the part of each Small Population particle according to evaluation function formula using Small Population Policy iteration algorithm
Optimal value and global optimum, the individual optimal value and global optimum of comparison position, if individual optimal value is better than global optimum
Value then replaces global optimum with individual optimal value；Otherwise replace individual optimal value with global optimum；
S240, chaotic mutation global optimum form Chaos Variable, and the value range of Chaos Variable is between 0 to 1；It calculates
The desired value of each particle, judge the desired value whether reach ideal optimizing effect and iterations be more than or equal to it is default most
The half of big iterations；If it is not, then return to step S220, if so, thening follow the steps S250；
S250 after carrying out preset H iteration, judges whether are global optimum's fitness and Small Population draw fitness
Less than preset positive number G；If so, algorithm terminates, gas management concentration is exported；If it is not, then return to step S220 continues
Iteration optimization.
The speed more new formula is in one of the embodiments,：
The location update formula is：
In formula, i represents i-th of particle；D indicates d dimension spaces；V represents the speed of single particle；P is the position of particle；
K is kth time iteration；Introduce Boundary Variables of the variable x as limitation d dimension spaces, xminRepresent minimum boundary, xmaxRepresent maximum
Boundary；W is inertia weight；c1、c2For positive accelerator coefficient；r1、r2Random number between being 0 to 1.
The inertia weight formula is in one of the embodiments,：
In formula, wkFor the inertia weight of kth time iteration；kmaxFor final iterations.
The evaluation function is in one of the embodiments,：
In formula, fiFor the fitness for being originally when introducing Small Population concept, SiThe information between single particle and remaining particle
The summation of element.
The E is the positive integer between 8 to 20 in one of the embodiments,；The F be 80 to 200 between it is just whole
Number, c1And c2Equal value is 2.
In one of the embodiments, in step S250, the iterations of the preset H iteration watt as needed
This monitoring accuracy determines；The preset positive integer G is less than or equal to 0.001.
The present invention also provides a kind of control systems of gas detecting device, and the control system includes controller, described
Controller is for realizing control method as described above.
The control method of above-mentioned gas detecting device, the PSO Neural Network dynamic for introducing Small Population optimization compensate, solution
Traditional gas indicator of having determined is existing since sampled analogue circuits drift, reference voltage have fluctuation, component aging etc. to ask
Topic reduces accuracy of detection, improves the accuracy of detection of gas detecting device.
Description of the drawings
In order to illustrate the technical solutions in the embodiments of the present application or in the prior art more clearly, below will be to institute in embodiment
Attached drawing to be used is needed to be briefly described, it should be apparent that, the accompanying drawings in the following description is only one described in the present invention
A little embodiments, for those of ordinary skill in the art, other drawings may also be obtained based on these drawings.
Fig. 1 is the control method flow chart of the gas detecting device of one embodiment of the invention；
Fig. 2 is the control method step S200 flow charts of the gas detecting device of one embodiment of the invention；
Fig. 3 is the neural network algorithm model of the control method of the gas detecting device of one embodiment of the invention；
Fig. 4 is the white noise acoustic jamming fitting result chart of the control method of one embodiment of the invention gas detecting device；
Fig. 5 is the white noise acoustic jamming fitting result chart of least square method supporting vector machine.
Specific implementation mode
In order to make the purpose , technical scheme and advantage of the present invention be clearer, by the following examples, it and combines attached
The control method and control system of the gas detecting device of the present invention is further elaborated in figure.It should be appreciated that herein
Described specific examples are only used to explain the present invention, is not intended to limit the present invention.
It drifts about there are sampled analogue circuits for traditional gas indicator and there are the various dry inspections for leading to detector
Survey the relatively low problem of precision, although structurally can as possible consider thorough carry out error compensation in hardware design, gas gas chamber,
But this solution is on the one hand of high cost, detector is complicated, and on the other hand, also it is necessary to by establishing corresponding mend
Control method is repaid to further increase its accuracy of detection, detection is made to optimize, reaches optimal detection result.
Refering to Figure 1, the control method of the gas detecting device of an embodiment provided by the invention, including following step
Suddenly：
S100 obtains gas management signal；
S200, the PSO Neural Network algorithm dynamic that Small Population optimization is carried out to gas management signal compensate, output watt
This detectable concentration.
The control method of above-mentioned gas detecting device, the PSO Neural Network dynamic for introducing Small Population optimization compensate, solution
Traditional gas indicator of having determined is existing since sampled analogue circuits drift, reference voltage have fluctuation, component aging etc. to ask
Topic reduces accuracy of detection, improves the accuracy of detection of gas detecting device.
It is asked although neural network algorithm is a kind of relatively conventional non-linear multiple coupled strong external interference factor that is applied to
Dissatisfied algorithm certainly is inscribed, but neural network algorithm needs big-sample data to be trained early period, measuring accuracy is influenced, in of the invention
The example group's neural network algorithm optimized using Small Population, is optimized by Small Population, is changed to PSO Neural Network algorithm
It is kind, optimal velocity and the position of neighborhood particle can be taken into account in certain radius, reduces data volume used in training, accelerate to receive
Speed is held back, detection efficiency is improved.
Further, it as a kind of optional embodiment, is introduced in the PSO Neural Network algorithm of Small Population optimization
Tent maps neural network algorithm generates Chaos Variable.Although being carried out to PSO Neural Network algorithm by Small Population optimization
Improve, optimal velocity and the position of neighborhood particle can be taken into account in certain radius, increases the local search ability of population, still
Later stage still can precocity be absorbed in local extremum, the embodiment by introduce tent maps neural network algorithm, can be used as small
The PSO Neural Network algorithm of swarm optimization increases the tool of ability of searching optimum, reinforces the population nerve of Small Population optimization
Ergodic, the randomness of network algorithm eliminate of inferior quality particle and obtain optimal particle.After introducing tent maps neural network algorithm,
Have the advantages that fast iteration speed, strong antijamming capability, topological structure are simple to the dynamic compensation of gas management signal, Neng Goushi
Now firedamp sensor is more preferably dynamically compensated.
Optionally, please continue to refer to shown in Fig. 3, the model of the neural network algorithm of one embodiment of the invention includes first
Layer, the second layer, third layer and the 4th layer；
First layer is input layer, and the corresponding input signal of input layer includes the electric signal of gas management signal, external environment
Temperature-humidity signal, non-methane gas and dust disturbing factor signal and electromagnetic interference signal；
The second layer is used to describe the membership function and Center Parameter of input layer；
Third layer is used to correspond to the number of nodes of neural network；
The 4th layer of gas density for corresponding output.
In the present embodiment, algorithm optimization be the second layer membership function and Center Parameter and third layer to the 4th
The connection weight of layer.Optionally, mean square error is further taken as evaluation function to adjust the training effect of iterative algorithm.
The non-methane gas interference signal in the various external worlds, temperature-humidity signal, dust interference signal, electromagnetic interference signal also conduct
The input of neural network can be trained neural network under collective effect, to which related interference factors are introduced control system
In system, the accuracy of neural network compensation is improved.
Shown in Fig. 2, as a kind of optional embodiment, gas is examined based on PSO Neural Network algorithm
Signal is surveyed into Mobile state compensation process, including following sub-step：
S210, particle populations initialize, including population total number of particles N, solution space dimension D, at the beginning of PSO Neural Network
Initial value is that D ties up Chaos Variable, assigns initial position, speed, inertia weight to each particle.
S220, E iteration before being carried out according to the speed of PSO Neural Network algorithm more new formula and location update formula,
The preceding F particle for selecting position optimal according to fitness value is as the center of circle for dividing Small Population and initializes F Small Population
Radius, obtain the Chaos Variable after the E times iteration according to chaotic mutation formula, E, F are positive integer；
Wherein, chaotic mutation formula is Z'k=β Zk+(1-β)Ω*；In formula,Ω*For D dimension spaces
Middle optimal solution chaos vector, value is between 0 to 1, X=(x1,x2,...,xD) be chaos vector, be after chaotic disturbance to
Amount；β is impact factor, and influence is degree of variation of the chaotic mutation to particle.
In the present embodiment, when the particle of Small Population is absorbed in local extremum after improvement, pass through chaotic mutation formula
Chaotic disturbance is added to help the example of Small Population after improving to enhance ability of searching optimum again.It can be with by chaotic mutation formula
Find out, particle is being gradually reduced with the influence of the increase chaotic mutation of iterations so that algorithm has stronger in the early stage
Ability of searching optimum, and can be restrained rapidly in the later stage and search out optimal value.It is combined with Small Population split plot design, tent maps god
Through network chaos algorithm, the value of radius r is in optimized selection in global scope, gives full play to the traversal of the chaos algorithm
Property；The chaotic mutation under original state can substantially shorten the iterations of chaotic mutation algorithm with the cooperation of Small Population algorithm simultaneously
And improve its secondary search precision.
Optionally, E is the positive integer between 8 to 20；F is the positive integer between 80 to 200, and c1 and the equal values of c2 are 2.
E chooses smaller numerical value to carry out Small Population optimization as early as possible, and the choosing value range of F is related with input quantity range, is not easy excessive
Or it is too small, crossing the finally obtained accuracy of ambassador reduces, too small, cannot embody the characteristics of Small Population optimizes.
In the present embodiment, the value that the value of E is 10, F is 100.Updated according to the speed of PSO Neural Network algorithm
Formula and location update formula carry out preceding 10 iteration, and preceding 100 particles for selecting position optimal according to fitness value are made
To divide the center of circle of Small Population and initializing the radiuses of 100 Small Populations, according to chaotic mutation formula obtain the 10th iteration it
Chaos Variable afterwards.
Optionally, speed more new formula isLocation updating is public
Formula is：In formula, i represents i-th of particle；D indicates d dimension spaces；V represents the speed of single particle；P is particle
Position；K is kth time iteration；Introduce Boundary Variables of the variable x as limitation d dimension spaces, xminRepresent minimum boundary, xmax
Represent maximum boundary；W is inertia weight；c1、c2For positive accelerator coefficient；r1、r2Random number between being 0 to 1.
It is updated in formula and location update formula from speed it can be found that also there is existing individual variable in particle cluster algorithm
Global variable, each particle is intended to chase the position and speed of optimal particle in population in algorithm evolution iteration.W inertia
Weight determines that population is searched in global scope in still subrange and scans for that optionally, inertia weight formula selects
The linear decrease weight equation is taken to beIn formula, wkFor the inertia weight of kth time iteration；kmaxFor
Final iterations.
It can be seen that the increase weight coefficient with iterations is gradually increasing, that is, particle local search energy
Power is being reinforced until making algorithmic statement.When the particle cluster algorithm of this classics optimizes problem with other biological intelligence
Energy algorithm such as ant group algorithm, ant colony algorithm is the same, is easily trapped into local extremum, algorithm premature convergence problem occurs.
The present invention is optimized using Small Population, and entire population is divided into multiple Small Populations first, chooses each microspecies
The particle with the optimal advantage of Position And Velocity is the center of circle in group, with dik=| | Xi-Xk| | (i, k=1,2 ..., N i ≠ k) be
For the border circular areas of maximum radius as a Small Population, wherein N is population total number of particles.
S230 evaluates the part of each Small Population particle according to evaluation function formula using Small Population Policy iteration algorithm
Optimal value and global optimum, the individual optimal value and global optimum of comparison position, if individual optimal value is better than global optimum
Value then replaces global optimum with individual optimal value；Otherwise replace individual optimal value with global optimum.
Further, in order to which the neighborhood search ability for increasing each individual selects information functionInformation interchange medium as each individual and another individual.In formula, r is kind
Group's radius, α are information function control parameter.It is contemplated that if a particle has n such individuals and its exchange of information,
It is evident that it is most closely to be contacted between particle similar in its position and speed.
Further, as a kind of optional embodiment, the evaluation function of Small Population individual isIn formula, fiFor
Original fitness to introduce when Small Population concept, SiThe pheromones summation between single particle and remaining particle.
From information function and evaluation function it can be found that if particle particle similar with its is more, information
Plain summation SiBigger, its corresponding fitness function is with regard to smaller, then there is this particle of similar characteristic to work as in iterative evolution
In can gradually be eliminated with the reduction of fitness, what is remained is the higher individual of fitness, in this way optimize after
Particle populations are iterated calculating again.This partitioning algorithm had both maintained the diversity of population population, while also washing in a pan competition
The mechanism of eliminating is introduced into algorithm, is that population is carried out evolution screening in fact, has both been avoided the premature convergence problem of algorithm or has been accelerated algorithm
Convergence rate.
S240, chaotic mutation global optimum form Chaos Variable, and the value range of Chaos Variable is between 0 to 1；It calculates
The desired value of each particle judges whether desired value reaches ideal optimizing effect and iterations are more than or equal to default maximum and change
The half of generation number；If it is not, then return to step S220, no to be, S250 is thened follow the steps.
S250 after carrying out preset H iteration, judges whether are global optimum's fitness and Small Population draw fitness
Less than preset positive number G；If so, algorithm terminates, heat promotees gas management concentration；If it is not, then return to step S220 continues
Iteration optimization.
As a kind of optional embodiment, in step S250, the iterations of preset H iteration gas as needed
Monitoring accuracy determines.Still optionally further, in step S250, it is positive number as small as possible to preset positive integer G.Optionally, it presets
Positive integer G be less than or equal to 0.001, to improve compensation accuracy.
In the present invention, 5~60 DEG C of temperature in use, the various concentration methane gas and humidity that gradient is 0.5 DEG C are 20%
~70%, the various concentration methane gas that gradient is 10% is trained and detects to compensation model, and least square is supported
The detection method of vector machine compensation as a comparison case, is detected, testing result is as shown in table 1 under identical condition.
1 neural network of the present invention of table compensation detection compensates testing result with comparative example least square support vector machines
As shown in table 1, the difference between the gas density that control method of the present invention measures and Real Gas concentration obviously compared with
Small, absolute error maximum is no more than 0.021, and maximum relative error is no more than 0.08；And least square method supporting vector machine compensates
The data error of detection is much larger than the present invention, absolutely proves that the neural network algorithm of the present invention can be with effective compensation gas in data
Trueness error when measurement improves accuracy of detection.
It is the fitting effect of white noise function pair algorithm shown in Fig. 4 to Fig. 5, white noise isBeing one, just too distribution function, normal distribution are most common disturbing factors, we can be it
Regard the interference of outer bound pair gas detecting device as.Please comparison diagram 3 and Fig. 4 it is found that neural network algorithm model of the present invention it is quasi-
The fitting effect with obvious effects better than least square method supporting vector machine model is closed, point is almost ideal with the fitting effect of line.
The compensation model of the Small Population optimization population neural algorithm proposed by the present invention for introducing tent maps neural algorithm,
By the simulated training to model, and the test to methane gas concentration under various concentration, fully demonstrate the supplementary model
Compensation effect.And noise jamming fitting result chart illustrates that the anti-jamming effectiveness of this model is good.
Several embodiments of the invention above described embodiment only expresses, the description thereof is more specific and detailed, but simultaneously
Cannot the limitation to the scope of the claims of the present invention therefore be interpreted as.It should be pointed out that for those of ordinary skill in the art
For, without departing from the inventive concept of the premise, various modifications and improvements can be made, these belong to the guarantor of the present invention
Protect range.Therefore, the protection domain of patent of the present invention should be determined by the appended claims.
Claims (10)
1. a kind of control method of gas detecting device, which is characterized in that include the following steps：
S100 obtains gas management signal；
S200, the PSO Neural Network algorithm dynamic that Small Population optimization is carried out to the gas management signal compensate, output watt
This detectable concentration.
2. control method according to claim 1, which is characterized in that in the PSO Neural Network of Small Population optimization
Tent maps neural network algorithm is introduced in algorithm, generates Chaos Variable.
3. control method according to claim 2, which is characterized in that the model of the neural network algorithm includes first
Layer, the second layer, third layer and the 4th layer；
The first layer is input layer, the corresponding input signal of the input layer include the gas management signal electric signal,
External environment temperature-humidity signal, non-methane gas and dust disturbing factor signal and electromagnetic interference signal；
The second layer is used to describe the membership function and Center Parameter of the input layer；
The third layer is used to correspond to the number of nodes of the neural network；
The described 4th layer gas density for corresponding output.
4. control method according to claim 3, which is characterized in that the PSO Neural Network algorithm that is based on is to described
Gas management signal is into Mobile state compensation process, including following sub-step：
S210, particle populations initialize, including population total number of particles N, solution space dimension D, at the beginning of the PSO Neural Network
Initial value is that D ties up Chaos Variable, assigns initial position, speed, inertia weight to each particle；
S220, E iteration before being carried out according to the speed of PSO Neural Network algorithm more new formula and location update formula, according to
The optimal preceding F particle in fitness value selection position is as the center of circle for dividing Small Population and initializes the half of F Small Population
Diameter, the Chaos Variable after the E times iteration is obtained according to chaotic mutation formula, and E, F are positive integer；
Wherein, chaotic mutation formula is：
Z'k=β Zk+(1-β)Ω*；
In formula,Ω*For optimal solution chaos vector in D dimension spaces, value is between 0 to 1, X=(x1,
x2,...,xD) it is chaos vector, it is the vector after chaotic disturbance；β is impact factor, and influence is mixed
Degree of variation of the ignorant variation to particle；
S230 evaluates the local optimum of each Small Population particle according to evaluation function formula using Small Population Policy iteration algorithm
Value and global optimum, the individual optimal value and global optimum of comparison position, if individual optimal value better than if global optimum
Replace global optimum with individual optimal value；Otherwise replace individual optimal value with global optimum；
S240, chaotic mutation global optimum form Chaos Variable, and the value range of Chaos Variable is between 0 to 1；It calculates each
The desired value of particle judges whether the desired value reaches ideal optimizing effect and iterations are more than or equal to default maximum and change
The half of generation number；If it is not, then return to step S220, if so, thening follow the steps S250；
S250 after carrying out preset H iteration, judges whether global optimum's fitness and Small Population draw fitness are less than
Preset positive number G；If so, algorithm terminates, gas management concentration is exported；If it is not, then return to step S220 continues iteration
Optimization.
5. control method according to claim 4, which is characterized in that the speed more new formula is：
The location update formula is：
In formula, i represents i-th of particle；D indicates d dimension spaces；V represents the speed of single particle；P is the position of particle；K is
Kth time iteration；Introduce Boundary Variables of the variable x as limitation d dimension spaces, xminRepresent minimum boundary, xmaxRepresent maximum side
Boundary；W is inertia weight；c1、c2For positive accelerator coefficient；r1、r2Random number between being 0 to 1.
6. control method according to claim 5, which is characterized in that the inertia weight formula is：
In formula, wkFor the inertia weight of kth time iteration；kmaxFor final iterations.
7. control method according to claim 4, which is characterized in that the evaluation function is：
In formula, fiFor the fitness for being originally when introducing Small Population concept, SiThe pheromones between single particle and remaining particle
Summation.
8. control method according to claim 5, which is characterized in that the E is the positive integer between 8 to 20；The F is
Positive integer between 80 to 200, c1And c2Equal value is 2.
9. control method according to claim 4, which is characterized in that in step S250, the preset H iteration changes
Generation number gas monitor precision as needed determines；The preset positive integer G is less than or equal to 0.001.
10. a kind of control system of gas detecting device, which is characterized in that the control system includes controller, the control
Device is for realizing the control method as described in claim 1 to 9 any one.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810261277.8A CN108535208A (en) | 2018-03-28 | 2018-03-28 | The control method and control system of gas detecting device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810261277.8A CN108535208A (en) | 2018-03-28 | 2018-03-28 | The control method and control system of gas detecting device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108535208A true CN108535208A (en) | 2018-09-14 |
Family
ID=63483750
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810261277.8A Pending CN108535208A (en) | 2018-03-28 | 2018-03-28 | The control method and control system of gas detecting device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108535208A (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104834215A (en) * | 2015-03-24 | 2015-08-12 | 浙江师范大学 | Variation particle swarm optimized BP neural network proportion integration differentiation (PID) control algorithm |
| CN105404783A (en) * | 2015-12-01 | 2016-03-16 | 吉林大学 | Blind source separation method |
| CN106874999A (en) * | 2017-02-15 | 2017-06-20 | 湖南省瞬渺通信技术有限公司 | A kind of full flowing water particle swarm optimization algorithm implementation method based on FPGA |
-
2018
- 2018-03-28 CN CN201810261277.8A patent/CN108535208A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104834215A (en) * | 2015-03-24 | 2015-08-12 | 浙江师范大学 | Variation particle swarm optimized BP neural network proportion integration differentiation (PID) control algorithm |
| CN105404783A (en) * | 2015-12-01 | 2016-03-16 | 吉林大学 | Blind source separation method |
| CN106874999A (en) * | 2017-02-15 | 2017-06-20 | 湖南省瞬渺通信技术有限公司 | A kind of full flowing water particle swarm optimization algorithm implementation method based on FPGA |
Non-Patent Citations (1)
| Title |
|---|
| 常静等: "《基于SPCPSO 神经网络红外瓦斯传感器自动补偿研究》", 《煤炭技术》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106093724B (en) | A kind of shelf depreciation Wavelet noise-eliminating method based on hybrid particle swarm | |
| CN114969995B (en) | Rolling bearing early fault intelligent diagnosis method based on improved sparrow search and acoustic emission | |
| Sikora | Induction and pruning of classification rules for prediction of microseismic hazards in coal mines | |
| CN114692505A (en) | Error prediction method and device for capacitor voltage transformer | |
| CN106569272B (en) | A kind of seismic attribute fusion method that dimension is risen based on data character space | |
| CN106991509A (en) | Log Forecasting Methodology based on radial basis function neural network model | |
| CN111709350B (en) | Low-frequency oscillation modal parameter identification method and system based on FCM clustering | |
| Wong et al. | Reservoir characterization using support vector machines | |
| CN104570109A (en) | Method for reservoir petroleum gas prediction | |
| CN106446514A (en) | Fuzzy theory and neural network-based well-log facies recognition method | |
| Zhang et al. | A novel comprehensive model of suitability analysis for matching area in underwater geomagnetic aided inertial navigation | |
| CN108535208A (en) | The control method and control system of gas detecting device | |
| Zhu et al. | Quantitative evaluation of low-permeability gas reservoirs based on an improved fuzzy-gray method | |
| Zhou et al. | Application of Kriging algorithm based on ACFPSO in geomagnetic data interpolation | |
| CN112862169B (en) | Method and device for predicting free oil content of continental phase shale | |
| CN108668254A (en) | The area positioning method of WiFi signal feature based on improved BP | |
| CN112464848B (en) | Information flow abnormal data monitoring method and device based on density space clustering | |
| CN108509735A (en) | A kind of Cylinder Liner-piston Ring break-in trend prediction method | |
| Winkler | A power-law fault model for Bayesian network geosteering in complex geology | |
| CN110320573B (en) | Logging parameter construction method and system reflecting reservoir productivity | |
| Thinh et al. | GIS-based multiple criteria decision analysis and optimization for land suitability evaluation | |
| CN112147679A (en) | Lithology prediction method and device based on elastic parameters under fuzzy logic framework | |
| CN115508511B (en) | Sensor self-adaptive calibration method based on full-parameter feature analysis of gridding equipment | |
| Gedeon et al. | Use of linguistic petrographical descriptions to characterise core porosity: contrasting approaches | |
| CN113176613B (en) | Multi-information fusion low-frequency model building method based on three-level body control |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180914 |