CN104483714B - Circulating fluidized bed boiler break coal detection device and detection method - Google Patents
Circulating fluidized bed boiler break coal detection device and detection method Download PDFInfo
- Publication number
- CN104483714B CN104483714B CN201410604626.3A CN201410604626A CN104483714B CN 104483714 B CN104483714 B CN 104483714B CN 201410604626 A CN201410604626 A CN 201410604626A CN 104483714 B CN104483714 B CN 104483714B
- Authority
- CN
- China
- Prior art keywords
- controller
- coal
- alarm lamp
- detection
- lever
- 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
- Fluidized-Bed Combustion And Resonant Combustion (AREA)
- Control Of Steam Boilers And Waste-Gas Boilers (AREA)
Abstract
The invention discloses a circulating fluidized bed boiler break coal detection device and a detection method. The detection device comprises a controller electrically connected with each electric appliance in a boiler, m cameras respectively arranged at m coal bunker coal mouths of the boiler, a memory, a first alarm lamp, a second alarm lamp, a third alarm lamp, an alarm horn, a coal receiving board which is arranged in the lower portion of a coal bunker of the boiler and opens downwards in the horn-shaped mode, a lever of which one end penetrates through a strip hole of the side wall of the coal bunker and is hinged with the upper end of the coal receiving board, an elastic sealing sleeve used for sealing a gap between the lever and the strip hole, a vertical rod connected with the other end of the lever and a hammer arranged on the vertical rod. The detection device has advantages of high detection accuracy, effective error operation avoidance performance, good adjustability and good adaptability.
Description
Technical field
The coal fault detection technique field the present invention relates to CFB CFBB breaks, especially relating to one kind can be fast
The disconnected coal detection means of CFBB of speed, the accurately disconnected coal fault of detection CFB CFBB and detection method.
Background technology
Under the dual-pressure of energy and environment, recirculating fluidized bed (CFB) boiler is so that its fuel tolerance is strong, efficiency of combustion
Height, load adjustment ability are good, the low distinguishing feature of pollutant emission, have obtained quick development in China.Ended for the end of the year 2011,
CFBB 3000 multiple stage of the existing different capabilities of China, puts into commercial operation more than 90000MW according to incompletely statistics.
The 300MW level Circulating Fluidized Bed Boilers having put into operation reach 40 multiple stage, have put into operation and in the 300MW grade recycle fluid bed built
Boiler controller system summation has then reached 110 multiple stage.Additionally, the maximum supercritical circulating fluidized bed pot of 600MW of single-machine capacity in the world
Stove unit Sichuan Baima recirculating fluidized bed Demonstration Station project passes through 168 hours trial operations at full capacity in April, 2013
Successfully put into operation.It is therefore contemplated that CFBB will obtain bigger development in China.
Although CFB boiler just employs low flow velocity, blast furnace hearth, middle Matter Transfer multiplying power at the beginning of design, lays to grow and defend combustion
The measures such as band, but have the reason the disconnected coal fault of frequent generation that is in operation, generation disconnected coal fault：The moisture of coal is excessive, granularity not
All blocking, bridgings etc..
When there is disconnected coal fault, if processing the fluctuation that will cause boiler operatiopn state not in time, or even cause to stop
The generation of stove accident.
Chinese patent mandate publication number：CN103439081A, authorizes publication date on December 11st, 2013, discloses a kind of life
Material CFBB flow behavior measurement method is it is characterised in that comprise the steps：Set up for simulating biomass
Euler's two-fluid model of CFBB flow behavior；Structural parameters according to biomass recirculating fluidized bed boiler are set up
The full scale model of described biomass recirculating fluidized bed boiler；The full scale model of described biomass recirculating fluidized bed boiler is entered
Row stress and strain model sets up the grid model of described biomass recirculating fluidized bed boiler；Obtain described biomass recirculating fluidized bed boiler
Corresponding gas parameter, particle parameter, boundary condition parameter, initial wind speed；According to described Euler's two-fluid model, grid mould
Gas in type, gas parameter, particle parameter, boundary condition parameter, biomass recirculating fluidized bed boiler described in initial wind speed simulation
Gu biphase flow process；By the flow process of described simulation, measure the velocity field regularity of distribution of gas-particle two-phase in burner hearth, obtain
Flow behavior to biomass recirculating fluidized bed boiler.The weak point of this invention is, single function it is impossible to for detecting whether
Disconnected coal.
Content of the invention
The goal of the invention of the present invention is to overcome CFBB in prior art disconnected coal accident easily
Not enough, there is provided a kind of disconnected coal of CFBB that can quickly, accurately detect the disconnected coal fault of CFB CFBB
Detection means and detection method.
To achieve these goals, the present invention employs the following technical solutions：
A kind of CFBB breaks coal detection means, including the controller electrically connecting with each electrical part in boiler,
It is respectively arranged on m photographic head at coal mouth under m coal bunker of boiler, memorizer, the first alarm lamp, the second alarm lamp, the 3rd report
What warning lamp, alarming horn, horn-like downwards in the coal bunker bottom of boiler opened connect coal plate, one end passes through located at coal bunker side
The bar hole of wall simultaneously and connects the hinged lever in coal plate upper end, the elastic packing for the gap between closure lever and bar hole
Cover the montant being connected with the lever other end and the weight on montant；Lever is connected with coal bunker side wall by articulated slab, leans on
The coal bunker side wall of nearly bar hole top edge is provided with the postive stop baffle with lever mated, near the coal bunker side wall of bar hole lower limb
It is provided with the limit switch with lever mated；Controller respectively with memorizer, alarming horn, the first alarm lamp, the second alarm lamp,
3rd alarm lamp, limit switch and the electrical connection of each photographic head.
For evading the unreliable and limitation of single disconnected coal detection method, the present invention is by machinery, image and software detection phase
In conjunction with when the Indexes Abnormality of 2 kinds of methods any in 3 kinds of detection methods, controller then controls alarming horn to send alarm signal；Respectively
Individual alarm lamp flashes for providing disconnected coal to note pointing out to operator, reminds operator to note observing；Detection is more accurate,
It is prevented effectively from maloperation.
Therefore, the present invention has detection accuracy height, is prevented effectively from maloperation；Improve the safety of boiler operatiopn and steady
Qualitatively feature.
Preferably, described lever includes one end stretches into the fix bar of coal bunker, expansion link and on the fix bar other end
Adapter sleeve；Several length adjustment through holes are equipped with adapter sleeve and expansion link, adapter sleeve passes through bolt with expansion link even
Connect..
The structure setting of lever, makes the adjustable length of lever, strong applicability.
Preferably, described montant bottom is provided with screw rod, described screw rod is from top to bottom sequentially provided with upper cap nut, upper gear
Plate, lower baffle plate and at least one lower nut；Described weight is sheathed on the screw rod between overhead gage, lower baffle plate, if weight includes
Dry block counterweight disc；Screw diameter is more than the diameter of montant.
Using front, repeatedly adjust counterweight and expansion link collapsing length it is ensured that when no coal falls, lever is due to weight
Reason depresses the contact of limit switch.
Preferably, described limit switch is provided with the L shaped plate that right-hand member is folded upward at, L shaped plate left part is opened with located at spacing
The gripper shoe shut is hinged, and L shaped plate left end is connected with limit switch by spring, and L shaped plate right-hand member is matched with lever；Spacing
The contact of switch is matched with L shaped plate lower surface right part.
Preferably, the coal bunker side wall below lever is provided with triangle bracket, described limit switch is located on bracket.
A kind of detection method of the disconnected coal detection means of CFBB, comprises the steps：
(1-1) machinery disconnected coal detection：
(1-1-1) it is previously provided with safety time threshold value W in controller；Controller obtains the signal of telecommunication of limit switch detection,
And calculate the time T that the contact of limit switch persistently depressed by lever；As T ＞ W, controller is made and is currently at first and prejudges coal
The judgement of state；As T≤W, the current judgement being to have coaly state made by controller；
(1-1-2) carry out first under state of activation and prejudge coal control：
Make boiler when controller and prejudge, by first, the judgement that coaly state switchs to have coaly state, then detection means enters activation
State；
In active state, as T ＞ W, controller controls the first alarm lamp flicker；
In active state, as T≤W, controller controls the first alarm lamp to stop flicker；
(1-2) image disconnected coal detection：
(1-2-1) m photographic head shoots the image that at lower coal mouth, coal flows respectively, and each image is stored and deposits by controller
It is handled as follows in reservoir and to image：
Controller utilizes impulse response function yij=med (xI+r, j+s(r, s) ∈ A) picture signal f (x, y) is filtered
Ripple；Wherein, A is filtering window, its a size of N=(2k+1) × (2k+1)；{xijIt is image manifold, med is median function,
The gray value of each pixel is set to the intermediate value of all pixels point gray value in this point neighborhood window by median function；Will
In digital picture or Serial No., the value of any is replaced with the Mesophyticum of each point value in a neighborhood of this point, allows the pixel value of surrounding
Close actual value, thus eliminate isolated noise spot.(r, s) is coordinate, xI+r, j+sCorresponding for pixel in filter window A
Gray value；
(1-2-2) motion detection region parameter is set：
A is set to 1-99 hough transform region by controller, and the interframe setting motion detection is divided into B；
(1-2-3) motor function detection：
In each macro block that controller sets in each detection zone, each picture element is (x, y), and difference threshold is Ta；Institute
The difference stating the picture element T moment with T-n moment brightness Y is MX, y(T)=| YX, y(T)-YX, y(T-n)|；
When | | YX, y(T)-YX, y(T-n) | | >=Ta, then set L=0；Otherwise, set L=1；Controller utilizes formula IMsum
=∑ L calculates and obtains macro block difference coefficient IMsum；Controller calculates each macro block difference coefficient IM in each detection zonesum
Sum, obtains the difference coefficient value of each detection zone, and is calculated the difference coefficient A of filtering window A furthersum；
(1-2-4)
Controller is detected to present image using step (1-1-3) at interval of B frame；
A when the image of any one photographic headsum≤ threshold value, then controller control the second alarm lamp flicker；When all shootings
The A of the image of headsum＞ threshold value, then controller control the second alarm lamp stopping flicker；
(1-3) software disconnected coal detection：
(1-3-1) be provided with memorizer there are 11 input nodes X=[X1 ..., X11], single hidden layer is 9 node Y
=[Y1 ..., Y9], the BP neural network model of 1 output node Z, wherein, be stored with memorizer q bar learning sample, net
Network performance objective error SSE≤0.00001, train epochs are at least d step；
(1-3-2) initialize BP neural network model：
(1-3-2-1) weight and threshold values initialize：Controller is produced using Gauss random function and meets normal distribution, all
Be worth for 0, variance is 1 and random number initialization weight W in interval [0,1] for the spanij, threshold values θ and θj, i=1 ...,
11；J=1 .., 9；Setting network performance error is ε；
(1-3-2-2) variable normalization：It is provided with corresponding respectively with 11 input nodes X1 ..., X11 in controller
11 groups of XMaxAnd XMin, controller is using formula X '=(X-XMin)/(XMax-XMin) respectively calculate q bar learning sample X1 ...,
The normalized value X ' of X111, X '2..., X '11；
It is provided with the Z corresponding with output contact Z in controllermaxAnd Zmin, controller is using formula Z '=(Z-ZMin)/
(ZMax-ZMin) calculate q bar learning sample Z normalized value Z '；Obtain through normalized s group learning sample；
(1-3-3) train BP neural network model：
(1-3-3-1) controller inputs s group learning sample in BP neural network model, and s initial value is 1, sets s
The target output value of group sample is Z '；
Using formulaCalculate hidden layer neuron output；
Using formulaCalculate output layer neuron reality output Zr；Wherein, function
Using formulaCalculate single sample deviation Es；
(1-3-3-2) start successively reversely to adjust weight and threshold values from output layer：
Controller makes weight W of output layerjIncrease by 0.6 × δ × Yj+0.45×ΔWj, Δ WjThe power increasing for previous adjustment
Weight, Δ Wo=0；Wherein, δ=(Z '-Zr)×Zr×(1-Zr),
Controller makes the weight of hidden layer increase by 0.6 × δj×Xi+0.45×ΔWij；
Wherein, δj=Yj×(1-Yj)×(δj×Wj)；
(1-3-3-3) as s ＜ q, s value increase by 1, return to step (1-3-3-1) are made；Otherwise proceed to step (1-3-3-4)；
(1-3-3-4) utilize formulaCalculate total error Et, wherein p is sample sequence number；
Work as Et≤ ε or study step number are less than d, and training terminates, and obtains the BP neural network model training；Otherwise proceed to step
Suddenly (1-3-3-1)；
(1-3-4) controller Real-time Collection main steam flow, feedwater flow, coal-supplying amount, bed temperature, combustion chamber draft, primary air pressure,
Primary air fan electric current, secondary wind pressure, overfire air fan electric current, air-introduced machine electric current and 11 technological parameters of exhaust gas temperature, using step
(1-2-2-2) 11 technological parameters are normalized, and each is sent into instruction through the technological parameter of normalized
In 11 input nodes of the BP neural network model perfected, obtain model output valve Zr, recycle formula V=Zr×(ZMax-
ZMin)+ZMinRenormalization obtains real-time oxygen content V, and controller calculatesWherein V1For setting in memorizer and currently main vapour
The related targeted oxygen content of flow；
(1-3-5) whenPersistent period is more than T1Second, then controller controls the 3rd alarm lamp flicker；
WhenPersistent period is more than T2Second, then controller controls the 3rd alarm lamp to stop flicker；
(1-4) when all flicker or any two flickers of first, second, third alarm lamp, controller is made disconnected coal and is judged,
And control alarming horn to report to the police.
Preferably, when main steam flow is 29% to 31.5%, V1For 8.2% to 8.9%；When main steam flow is
When 37.5% to 41%, V1For 7.5% to 8.1%；When main steam flow is 47.5% to 53%, V1For 6.75% to 7.1%；
When main steam flow is 57.5% to 61%, V1For 5.75% to 6.1%；When main steam flow is 67.5% to 73%, V1For
4.65% to 5.2%；When main steam flow is 78.5% to 83%, V1For 4.0% to 4.4%；When main steam flow be 87% to
When 92%, V1For 3.5% to 3.9%；When main steam flow is 92% to 100%, V1For 3.3% to 3.5%.
Preferably, q is 500 to 560.
Preferably, c is 1.18 to 1.32；D is 10000 to 11200.
Preferably, T1For 12 to 17；T2For 2 to 3.
Therefore, the present invention has the advantages that：
(1) detection accuracy is high, is prevented effectively from maloperation；
(2) improve the safety and stability of boiler operatiopn；
(3) adjustability is good, and the suitability is good.
Brief description
Fig. 1 is a kind of structural representation of the present invention；
Fig. 2 is a kind of theory diagram of the present invention；
Fig. 3 is a kind of flow chart of embodiments of the invention；
Fig. 4 is a kind of structural representation of the amplification at A in Fig. 1.
In figure：Controller 1, alarming horn 2, connect coal plate 3, bar hole 4, lever 5, elastic seal cartridge 6, located at perpendicular
Weight 7 on bar, articulated slab 8, postive stop baffle 9, limit switch 10, fix bar 11, expansion link 12, adapter sleeve 13, length
Degree adjusts through hole 14, upper cap nut 15, overhead gage 16, lower baffle plate 18, lower nut 19, counterweight disc 17, L shaped plate 20, props up
Fagging 21, spring 22, contact 23, bracket 24, coal bunker 25, montant 26, the first alarm lamp 27, the second alarm lamp 28,
Memorizer 29, photographic head 30, the 3rd alarm lamp 31.
Specific embodiment
The present invention will be further described with reference to the accompanying drawings and detailed description.
Embodiment as shown in Figure 1 and Figure 2 is a kind of disconnected coal detection means of CFBB, including each with boiler
The controller 1 of individual electrical part electrical connection, be respectively arranged on 8 photographic head 30 at coal mouth under 8 coal bunkers of boiler, memorizer 29,
First alarm lamp 27, the second alarm lamp 28, the 3rd alarm lamp 31, alarming horn 2, downward in coal bunker 25 bottom of boiler
Horn-like open connect coal plate 3, one end passes through located at the bar hole 4 of coal bunker side wall and the lever 5 hinged with connecing coal plate upper end, uses
Montant 26 that the elastic seal cartridge 6 in the gap between closure lever and bar hole is connected with the lever other end and located at montant
On weight 7；Lever is connected with coal bunker side wall by articulated slab 8, near bar hole top edge coal bunker side wall be provided with
The postive stop baffle 9 of lever mated, the coal bunker side wall near bar hole lower limb is provided with the limit switch 10 with lever mated；As
Shown in Fig. 2, controller respectively with memorizer, alarming horn, the first alarm lamp, the second alarm lamp, the 3rd alarm lamp, limit switch
With the electrical connection of each photographic head.
As shown in figure 1, lever includes one end stretches into the fix bar 11 of coal bunker, expansion link 12 and on the fix bar other end
Adapter sleeve 13；Length adjustment through hole 14 is equipped with adapter sleeve and expansion link, adapter sleeve is connected with expansion link by bolt.
Montant bottom is provided with screw rod, and described screw rod is from top to bottom sequentially provided with upper cap nut (15), overhead gage 16, lower baffle plate
18 descend nut 19 with least one；Weight is sheathed on the screw rod between overhead gage, lower baffle plate, and weight includes 5 pieces of counterweight discs
17；Screw diameter is more than the diameter of montant.
As shown in figure 4, limit switch is provided with the L shaped plate 20 that right-hand member is folded upward at, L shaped plate left part with located at limit switch
On gripper shoe 21 hinged, L shaped plate left end is connected with limit switch by spring 22, and L shaped plate right-hand member is matched with lever；Limit
The contact 23 of bit switch is matched with L shaped plate lower surface right part.
Coal bunker side wall below lever is provided with triangle bracket 24, and described limit switch is located on bracket.Postive stop baffle
The bending tabular extending horizontally and bending downwards, the circular in cross-section of the lower limb of postive stop baffle.First alarm lamp and
Two alarm lamps are all using stroboscopic lamp.
As shown in figure 3, detection method comprises the steps：
Step 100, machinery disconnected coal detection：
Step 110, is previously provided with safety time threshold value W in controller；Controller obtains the signal of telecommunication of limit switch detection,
And calculate the time T that the contact of limit switch persistently depressed by lever；As T ＞ W, controller is made and is currently at first and prejudges coal
The judgement of state；As T≤W, the current judgement being to have coaly state made by controller；W is 12 seconds；
Step 120, carries out first and prejudges coal control under state of activation：
Make boiler when controller and prejudge, by first, the judgement that coaly state switchs to have coaly state, then detection means enters activation
State；
In active state, as T ＞ W, controller controls the first alarm lamp flicker；
In active state, as T≤W, controller controls the first alarm lamp to stop flicker；
Step 200, image disconnected coal detection：
Step 210,8 photographic head shoot the image of coal flowing at lower coal mouth respectively, and each image is stored by controller
It is handled as follows in memorizer and to image：
Controller utilizes impulse response function yij=med (xI+r, j+s(r, s) ∈ A) picture signal f (x, y) is filtered
Ripple；Wherein, A is filtering window, its a size of N=(2k+1) × (2k+1)；{xijIt is image manifold, med is median function,
(r, s) is coordinate, xI+r, j+sFor the corresponding gray value of pixel in filter window A；
Step 220, arranges motion detection region parameter：
A is set to 1-99 hough transform region by controller, and the interframe setting motion detection is divided into B；
Step 230, motor function detects：
In each macro block that controller sets in each detection zone, each picture element is (x, y), and difference threshold is Ta；Institute
The difference stating the picture element T moment with T-n moment brightness Y is
MX, y(T)=| YX, y(T)-YX, y(T-n)|；
When | | YX, y(T)-YX, y(T-n) | | >=Ta, then set L=0；Otherwise, set L=1；Controller utilizes formula IMsum
=∑ L calculates and obtains macro block difference coefficient IMsum；Controller calculates each macro block difference coefficient IM in each detection zonesum
Sum, obtains the difference coefficient value of each detection zone, and is calculated the difference coefficient A of filtering window A furthersum；
Step 240,
Controller is detected to present image using step 230 at interval of B frame；Image when any one photographic head
Asum≤ threshold value, then controller control the second alarm lamp flicker；
A when the image of all photographic headsum＞ threshold value, then controller control the second alarm lamp stopping flicker；
Step 300, software disconnected coal detection：
Step 310, is provided with memorizer and has 11 input node X=areas 1 ..., X1], single hidden layer be 9 node Y
=[Y1 ..., Y9], the BP neural network model of 1 output node Z, wherein, the q=500 bar that is stored with memorizer learns sample
This, network performance objective error SSE≤0.00001, train epochs are at least d=10000 step；
Step 320, initializes BP neural network model：
Step 321, weight is initialized with threshold values：Controller is produced using Gauss random function and meets normal distribution, average
It is 1 and random number initialization weight W in interval [0,1] for the span for 0, varianceij, threshold values θ and θj, i=1 ...,
11；J=1 .., 9；Setting network performance error is ε；
Step 322, variable normalization：It is provided with controller and difference corresponding 11 in 11 input nodes X1 ..., X11
Group XMaxAnd XMin, controller is using formula X '=(X-XMin)/(XMax-XMin) respectively calculate q bar learning sample X1 ..., X11
Normalized value X '1, X '2..., X '11；
It is provided with the Z corresponding with output contact Z in controllermaxAnd Zmin, controller is using formula Z '=(Z-ZMin)/
(ZMax-ZMin) calculate q bar learning sample Z normalized value Z '；Obtain through normalized s group learning sample；
Step 330, trains BP neural network model：
Step 331, controller inputs s group learning sample in BP neural network model, and s initial value is 1, sets s
The target output value of group sample is Z '；
Using formulaCalculate hidden layer neuron output；
Using formulaCalculate output layer neuron reality output Zr；Wherein, function
Using formulaCalculate single sample deviation Es；
Step 332, starts successively reversely to adjust weight and threshold values from output layer：
Controller makes weight W of output layerjIncrease by 0.6 × δ × Yj+0.45×ΔWj, Δ WjThe power increasing for previous adjustment
Weight, Δ Wo=0；Wherein, δ=(Z '-Zr)×Zr×(1-Zr),
Controller makes the weight of hidden layer increase by 0.6 × δj×Xi+0.45×ΔWij；
Wherein, δj=Yj×(1-Yj)×(δj×Wj)；
Step 333, as s ＜ q, makes s value increase by 1, return to step 331；Otherwise proceed to step 334；
Step 334, using formulaCalculate total error Et, wherein p is sample sequence number；
Work as Et≤ ε or study step number are less than d, and training terminates, and obtains the BP neural network model training；Otherwise proceed to step
Rapid 331；
Step 340, controller Real-time Collection main steam flow, feedwater flow, coal-supplying amount, bed temperature, combustion chamber draft, First air
Pressure, primary air fan electric current, secondary wind pressure, overfire air fan electric current, air-introduced machine electric current and 11 technological parameters of exhaust gas temperature, using step
Suddenly (1-2-2-2) is normalized to 11 technological parameters, and each is sent into through the technological parameter of normalized
In 11 input nodes of the BP neural network model training, obtain model output valve Zr, recycle formula V=Zr×(ZMax-
ZMiin)+ZMinRenormalization obtains real-time oxygen content V, and controller calculatesWherein V1For setting in memorizer and currently leading
The related targeted oxygen content of steam flow amount；
Step 350, whenPersistent period is more than T1Second, then controller is made and is currently at second and prejudges coaly state
Judge, and control the second alarm lamp flicker；
WhenPersistent period is more than T2Second, then controller controls the second alarm lamp to stop flicker；
Step 400, disconnected coal judges and reports to the police：
When first, second alarm lamp all flashes, controller is made disconnected coal and is judged, and controls alarming horn to report to the police.
Wherein, when main steam flow is 29% to 31.5%, V1For 8.2% to 8.9%；When main steam flow be 37.5% to
When 41%, V1For 7.5% to 8.1%；When main steam flow is 47.5% to 53%, V1For 6.75% to 7.1%；When main steam flow
When measuring as 57.5% to 61%, V1For 5.75% to 6.1%；When main steam flow is 67.5% to 73%, V1For 4.65% to
5.2%；When main steam flow is 78.5% to 83%, V1For 4.0% to 4.4%；When main steam flow is 87% to 92%, V1
For 3.5% to 3.9%；When main steam flow is 92% to 100%, V1For 3.3% to 3.5%；C is 1.2, T1For 11, T2For 3.
It should be understood that the present embodiment is only illustrative of the invention and is not intended to limit the scope of the invention.In addition, it is to be understood that
After having read the content of present invention instruction, those skilled in the art can make various changes or modifications to the present invention, these etc.
Valency form equally falls within the application appended claims limited range.
Claims (9)
1. the detection method of the disconnected coal detection means of a kind of CFBB, the disconnected coal detection means of CFBB includes
M photographic head at coal mouth under the controller (1) electrically connecting with each electrical part in boiler, m coal bunker being respectively arranged on boiler
(30), memorizer (29), the first alarm lamp (27), the second alarm lamp (28), the 3rd alarm lamp (31), alarming horn (2), located at
What horn-like downwards in coal bunker (25) bottom of boiler opened connect coal plate (3), one end passes through the bar hole located at coal bunker side wall
(4) and and connect the hinged lever (5) in coal plate upper end, between closure lever and bar hole the elastic seal cartridge (6) in gap,
The montant (26) being connected with the lever other end and the weight (7) on montant；Lever passes through articulated slab (8) and coal bunker side wall phase
Connect, be provided with the postive stop baffle (9) with lever mated near the coal bunker side wall of bar hole top edge, near bar hole lower limb
Coal bunker side wall be provided with the limit switch (10) with lever mated；Controller respectively with memorizer, alarming horn, first warning
Lamp, the second alarm lamp, the 3rd alarm lamp, limit switch and the electrical connection of each photographic head；It is characterized in that, comprise the steps：
(1-1) machinery disconnected coal detection：
(1-1-1) it is previously provided with safety time threshold value W in controller；Controller obtains the signal of telecommunication of limit switch detection, and counts
Calculate the time T that the contact of limit switch persistently depressed by lever；As T ＞ W, controller is made and is currently at first and prejudges coaly state
Judgement；As T≤W, the current judgement being to have coaly state made by controller；
(1-1-2) carry out first under state of activation and prejudge coal control：
Make boiler when controller and prejudge, by first, the judgement that coaly state switchs to have coaly state, then detection means enters activation shape
State；
In active state, as T ＞ W, controller controls the first alarm lamp flicker；
In active state, as T≤W, controller controls the first alarm lamp to stop flicker；
(1-2) image disconnected coal detection：
(1-2-1) m photographic head shoots the image that at lower coal mouth, coal flows respectively, and each image is stored memorizer by controller
In and image is handled as follows：
Controller utilizes impulse response function yij=med (xI+r, j+s(r, s) ∈ A) picture signal f (x, y) is filtered；Its
In, A is filtering window, its a size of N=(2k+1) × (2k+1)；{xijIt is image manifold, med is median function, (r, s)
For coordinate, xI+r, j+sFor the corresponding gray value of pixel in filter window A；
(1-2-2) motion detection region parameter is set：
A is set to 1-99 hough transform region by controller, and the interframe setting motion detection is divided into B；
(1-2-3) motor function detection：
In each macro block that controller sets in each detection zone, each picture element is (x, y), and difference threshold is Ta；Described as
The vegetarian refreshments T moment is M with the difference of T-n moment brightness YX, y(T)=| YX, y(T)-YX, y(T-n)|；
When | | YX, y(T)-YX, y(T-n) | | >=Ta, then set L=0；Otherwise, set L=1；Controller utilizes formula IMsum=∑
L calculates and obtains macro block difference coefficient IMsum；Controller calculates each macro block difference coefficient IM in each detection zonesumSum,
Obtain the difference coefficient value of each detection zone, and be calculated the difference coefficient A of filtering window A furthersum；
(1-2-4)
Controller is detected to present image using step (1-1-3) at interval of B frame；Image when any one photographic head
Asum≤ threshold value, then controller control the second alarm lamp flicker；A when the image of all photographic headsum＞ threshold value, then controller control
Make the second alarm lamp and stop flicker；
(1-3) software disconnected coal detection：
(1-3-1) be provided with memorizer there are 11 input nodes X=[X1 ..., X11], single hidden layer is 9 node Y=
The BP neural network model of [Y1 ..., Y9], 1 output node Z, wherein, be stored with memorizer q bar learning sample, network
Performance objective error SSE≤0.00001, train epochs are at least d step；
(1-3-2) initialize BP neural network model：
(1-3-2-1) weight and threshold values initialize：Controller meets normal distribution using the generation of Gauss random function, average is
0th, variance is 1 and random number initialization weight W in interval [0,1] for the spanij, threshold values θ and θj, i=1 ..., 11；j
=1 .., 9；Setting network performance error is ε；
(1-3-2-2) variable normalization：It is provided with controller and corresponding 11 groups respectively in 11 input nodes X1 ..., X11
XMaxAnd XMin, controller is using formula X '=(X-XMin)/(XMax-XMin) calculate X1 ..., X11 of q bar learning sample respectively
Normalized value X '1, X '2..., X '11；
It is provided with the Z corresponding with output contact Z in controllermaxAnd Zmin, controller is using formula Z '=(Z-ZMin)/(ZMax-
ZMin) calculate q bar learning sample Z normalized value Z '；Obtain through normalized s group learning sample；
(1-3-3) train BP neural network model：
(1-3-3-1) controller inputs s group learning sample in BP neural network model, and s initial value is 1, sets s group sample
This target output value is Z '；
Using formulaCalculate hidden layer neuron output；
Using formulaCalculate output layer neuron reality output Zr；Wherein, function
Using formulaCalculate single sample deviation Es；
(1-3-3-2) start successively reversely to adjust weight and threshold values from output layer：
Controller makes weight W of output layerjIncrease by 0.6 × δ × Yj+0.45×ΔWj, Δ WjThe weight increasing for previous adjustment, Δ
Wo=0；Wherein, output layer error delta=(Z '-Zr)×Zr×(1-Zr),
Controller makes the weight of hidden layer increase by 0.6 × δj×Xi+0.45×ΔWij；
Wherein, δj=Yj×(1-Yj)×(δj×Wj)；
(1-3-3-3) as s ＜ q, s value increase by 1, return to step (1-3-3-1) are made；Otherwise proceed to step (1-3-3-4)；
(1-3-3-4) utilize formulaCalculate total error Et, wherein p is sample sequence number；
Work as Et≤ ε or study step number are less than d, and training terminates, and obtains the BP neural network model training；Otherwise proceed to step (1-
3-3-1)；
(1-3-4) controller Real-time Collection main steam flow, feedwater flow, coal-supplying amount, bed temperature, combustion chamber draft, primary air pressure, once
Blower fan electric current, secondary wind pressure, overfire air fan electric current, air-introduced machine electric current and 11 technological parameters of exhaust gas temperature, using step (1-2-
2-2) 11 technological parameters are normalized, and each is sent into and train through the technological parameter of normalized
In 11 input nodes of BP neural network model, obtain model output valve Zr, recycle formula V=Zr×(ZMax-ZMin)+ZMin
Renormalization obtains real-time oxygen content V, and controller calculatesWherein V1Related to current main steam flow for set in memorizer
Targeted oxygen content；
(1-3-5) whenPersistent period is more than T1Second, then controller controls the 3rd alarm lamp flicker；
WhenPersistent period is more than T2Second, then controller controls the 3rd alarm lamp to stop flicker；
(1-4) when all flicker or any two flickers of first, second, third alarm lamp, controller is made disconnected coal and is judged, and controls
Alarming horn processed is reported to the police.
2. the detection method of the disconnected coal detection means of CFBB according to claim 1, is characterized in that, described thick stick
Bar includes fix bar (11), expansion link (12) and the adapter sleeve (13) on the fix bar other end that coal bunker is stretched in one end；Even
Several length adjustment through holes (14) are equipped with female connector and expansion link, adapter sleeve is connected with expansion link by bolt.
3. the detection method of the disconnected coal detection means of CFBB according to claim 1, is characterized in that, described perpendicular
Bar bottom is provided with screw rod, and described screw rod is from top to bottom sequentially provided with upper cap nut (15), overhead gage (16), lower baffle plate (18) and extremely
A few lower nut (19)；Described weight is sheathed on the screw rod between overhead gage, lower baffle plate, and weight includes some pieces of balancing disks
Piece (17)；Screw diameter is more than the diameter of montant.
4. the detection method of the disconnected coal detection means of CFBB according to claim 1, is characterized in that, described limit
Bit switch is provided with the L shaped plate (20) that right-hand member is folded upward at, L shaped plate left part and gripper shoe (21) hinge on limit switch
Connect, L shaped plate left end is connected with limit switch by spring (22), and L shaped plate right-hand member is matched with lever；The contact of limit switch
(23) match with L shaped plate lower surface right part.
5. the detection method of the disconnected coal detection means of CFBB according to claim 1, is characterized in that, under lever
The coal bunker side wall of side is provided with triangle bracket (24), and described limit switch is located on bracket.
6. the detection method of the disconnected coal detection means of CFBB according to claim 1, is characterized in that, when main vapour
When flow is 29% to 31.5%, V1For 8.2% to 8.9%；When main steam flow is 37.5% to 41%, V1For 7.5% to
8.1%；When main steam flow is 47.5% to 53%, V1For 6.75% to 7.1%；When main steam flow is 57.5% to 61%
When, V1For 5.75% to 6.1%；When main steam flow is 67.5% to 73%, V1For 4.65% to 5.2%；When main steam flow is
When 78.5% to 83%, V1For 4.0% to 4.4%；When main steam flow is 87% to 92%, V1For 3.5% to 3.9%；Work as master
When steam flow amount is 92% to 100%, V1For 3.3% to 3.5%.
7. the detection method of the disconnected coal detection means of CFBB according to claim 1, is characterized in that, q is 500
To 560.
8. the detection method of the disconnected coal detection means of CFBB according to claim 1, is characterized in that, c is
1.18 to 1.32；D is 10000 to 11200.
9. the detection method of the disconnected coal detection means of CFBB according to claim 1, is characterized in that, T1For 12
To 17；T2For 2 to 3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410604626.3A CN104483714B (en) | 2014-10-31 | 2014-10-31 | Circulating fluidized bed boiler break coal detection device and detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410604626.3A CN104483714B (en) | 2014-10-31 | 2014-10-31 | Circulating fluidized bed boiler break coal detection device and detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104483714A CN104483714A (en) | 2015-04-01 |
| CN104483714B true CN104483714B (en) | 2017-02-22 |
Family
ID=52758279
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410604626.3A Active CN104483714B (en) | 2014-10-31 | 2014-10-31 | Circulating fluidized bed boiler break coal detection device and detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104483714B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105953221B (en) * | 2016-06-02 | 2018-03-27 | 陕西华电瑶池发电有限公司 | Coal for circulation fluid bed boiler storehouse coal-blocking-proof device of coal and method |
| CN106202959B (en) * | 2016-07-20 | 2019-03-22 | 广东电网有限责任公司电力科学研究院 | The method for numerical simulation and system of circulating fluidized bed boiler pollutant generating process |
| CN106861897A (en) * | 2017-02-24 | 2017-06-20 | 中国恩菲工程技术有限公司 | The control device of separator |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN2172861Y (en) * | 1993-08-02 | 1994-07-27 | 王国春 | Apparatus for weighing fine coal during flow |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH10176805A (en) * | 1996-12-17 | 1998-06-30 | Hitachi Ltd | Pressurized fluidized bed combustion device for coal |
| CN2292982Y (en) * | 1997-03-06 | 1998-09-30 | 周俊浩 | Foreward grate boiler coal supply device |
| CN2903712Y (en) * | 2006-04-21 | 2007-05-23 | 陈玉国 | Anti-flue reversing blocking-proof type coal supplier of circulation fluid-bed boiler |
| CN201052475Y (en) * | 2007-06-05 | 2008-04-30 | 永城煤电集团有限责任公司 | Coal bunker level automatic control peristalsis breaking preventing device |
| CN102226518B (en) * | 2011-02-01 | 2013-03-06 | 中国科学院工程热物理研究所 | Large-scale circulating fluidized bed boiler |
| CN203319211U (en) * | 2013-05-04 | 2013-12-04 | 大唐淮南洛河发电厂 | Coal caking detection and protection device for redirection rollers of coal conveying belt |
| CN203428477U (en) * | 2013-08-14 | 2014-02-12 | 中国神华能源股份有限公司 | Material-blocking protection device for material conveying system |
-
2014
- 2014-10-31 CN CN201410604626.3A patent/CN104483714B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN2172861Y (en) * | 1993-08-02 | 1994-07-27 | 王国春 | Apparatus for weighing fine coal during flow |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104483714A (en) | 2015-04-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104483714B (en) | Circulating fluidized bed boiler break coal detection device and detection method | |
| US20230383943A1 (en) | Method and Smart System for Fault Detection and Prevention in Industrial Boilers | |
| CN207264430U (en) | Positive-pressure type medium-speed pulverizer fuel pulverizing plant fineness of pulverized coal real-time monitoring system | |
| CN104345118B (en) | Solid propellant many targets line Dynamic Burning Performance Test System and method | |
| CN104459819B (en) | Boiler coal interruption detecting device and method | |
| CN104469284B (en) | The disconnected coal detection method of CFBB | |
| CN107326137B (en) | Blast funnace hot blast stove burns stove process operating parameters multistage matching optimization method at times | |
| CN103877801B (en) | Bag leakage detecting and positioning device and method for pulse bag dust collector | |
| CN101196413A (en) | Device and method for biological safety detection, control and alarm | |
| CN103925614B (en) | A kind of fired power generating unit fast cut-back fuel control method | |
| CN103348119A (en) | Control device for internal combustion engine | |
| CN104864382A (en) | Dry burning-resistant protecting device and method for gas/oil boiler | |
| CN105698158A (en) | Method for detecting four-tube leak of circulating fluidized bed boiler | |
| CN110345463A (en) | A kind of boiler incipient fault recognition methods and device | |
| CN206808100U (en) | A kind of cotton picker spontaneous combustion intelligent early-warning prevention and control system | |
| CN108875165B (en) | Boiler characteristic calibration method based on operation data | |
| CN208642964U (en) | A kind of electric control system of wet electrical dust precipitator | |
| CN106493558A (en) | Electronic throttle idler gear assembly system | |
| CN207231747U (en) | A kind of wind pressure simulator | |
| CN113653607B (en) | Intelligent power station fan stall early warning diagnosis method based on system efficiency model | |
| CN111964043B (en) | Novel chicken manure boiler return bed and monitoring method thereof | |
| CN208076671U (en) | A kind of blast switch detector | |
| CN107505927A (en) | CFB Boiler cigarette equipment fault monitoring method and device based on component | |
| CN105785065A (en) | Power station boiler wind speed on-line monitoring system and speed measuring method | |
| CN106438038A (en) | Controller for internal combustion engine |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |