CN102080569B - Distributed optical fiber temperature measurement-based fire early warning method for belt conveyor - Google Patents

Abstract

The invention relates to the field of coal mine safety, in particular to a distributed optical fiber temperature measurement-based fire early warning method for a belt conveyor. The method comprises the following steps of: dividing a temperature measuring optical fiber into a temperature measuring point, a channel and an area, and processing temperature data measured by the temperature measuring optical fiber by an absolute temperature early warning method, an area relative temperature difference early warning method, an area normal distribution early warning method, a measuring point temperature rise slope early warning method and a measuring point temperature rise variation trend early warning method to acquire a channel threshold early warning characteristic value, an area threshold early warning characteristic value, an area relative temperature early warning characteristic value, an area temperature normal distribution statistical characteristic value, a measuring point threshold early warning characteristic value, a measuring point temperature rise slope early warning characteristic value and a measuring point temperature rise accumulated trend early warning characteristic value; inputting the characteristic values and the measuring point temperature rise accumulated trend early warning characteristic value into a back propagation (BP) neural network model; and outputting a warning coefficient, an early warning coefficient and a safety coefficient by using the BP neural network model.

Description

Belt Conveyor Fire method for early warning based on distributed optical fiber temperature measurement

Technical field

The present invention relates to safety of coal mines field, particularly a kind of fire alarm method for mine belt conveyor.

Background technology

Mine belt conveyor is due to long use, easily bring out fire, its fire inducement mainly contains following several respects: 1. cylinder skids: under normal operation, the slip of the relative cylinder surface of adhesive tape is generally no more than 2%, if there is adhesive tape slipping, between adhesive tape and cylinder, exist larger relative sliding, and cause friction Temperature Rise, as do not found in time, once rising to adhesive tape ignition temperature, temperature will cause fire. the reason that common drive roll skids has: head coal piling after buying securities with all one's capital, skid-proof rubber serious wear on cylinder, full load starting after stopping in emergency etc., the main cause that driven cylinder skids is that roller bearings is badly damaged or stuck.2. carrying roller fault: in rubber conveyer firing accident, cause one of main cause that the too high adhesive tape that ignites of carrying roller temperature rise is Mine-used Belt Conveyor fire in recent years because carrying roller is stuck.Because the ambient conditions of down-hole is poor, in Idler bearing, be easy to enter dust, the life-span is shorter, and after bearing damage is stuck, adhesive tape will continue friction on carrying roller surface, make carrying roller accumulate a large amount of heats.Carrying roller temperature rises to very high, thus roasting combustion adhesive tape, and the beans-and bullets shooter burning coal powder of going forward side by side, expands rapidly the intensity of a fire.In investigation, also find. be mainly that snub pulley does not turn and easily causes that adhesive tape catches fire, because snub pulley, particularly near the snub pulley at tail place, bracket is overhead very near, very easily by clear and bright with on the coal dust that scatters surround, there is coal piling, cause the radiating condition variation of carrying roller, make stuck carrying roller temperature rise too high.3. spontaneous combustion of pulverized coal: exist a large amount of dumps on rubber conveyer ground along the line, owing to contacting and oxidative heat generation with air, in the situation that radiating condition is not smooth, the heat that oxidation generates is greater than the heat distributing to surrounding, cause the temperature of coal to raise gradually and cause " glowing ", once the burning-point that reaches coal, will spontaneous combustion form the intensity of a fire.4. outside burning things which may cause a fire disaster: except the above-mentioned fire being caused by sealing-tape machine itself, the Another reason that causes sealing-tape machine fire is extraneous unexpected burning things which may cause a fire disaster, as artificial naked light, wire short-circuiting, Underground Substation catch fire, unreasonable Site Welding, large spoil fall and produce fire by friction below adhesive tape etc.

Domestic current rubber conveyer Integrated Protection System kind is many, and the manufacturer that scale is larger mainly contains Tianjin Huaning electronics, Bake, Tianjin, Chongqing Mei Ke institute, Changzhou connection power automation, science and technology Changzhou, world automation research institute etc.Domestic application has in the commonplace technology of rubber conveyer fire monitoring: temperature (containing memorial alloy temperature pick up), smog, carbonomonoxide concentration, flame detector and heat-sensitive cable technology.Wherein mainly still take smog, Carbon Monoxide Detection and temperature as main, realize the key positions such as rubber conveyer head, tail are carried out to condition of a fire detection.But current smog and Carbon Monoxide Detection technology exist very large limitation, when reality is used at the scene, false alarm rate is very high, and result of use is bad; Temperature detection technology has multiple, from traditional Pt resistance to technology such as memorial alloy and non-contacting infrared thermometrics, all can not realize distributed, the monitoring continuously in whole transportation lane.

Summary of the invention

In view of this, in order to address the above problem, the invention discloses a kind of Belt Conveyor Fire method for early warning based on distributed optical fiber temperature measurement, can to whole transportation lane carry out distributed, continuously monitoring.

The object of the present invention is achieved like this: the Belt Conveyor Fire method for early warning based on distributed optical fiber temperature measurement, be arranged on ribbon conveyer temperature-measuring optical fiber is distributed, wherein, at the crucial thermometric of ribbon conveyer position, temperature-measuring optical fiber is installed by temperature-measuring optical fiber heat conducting clamp, and the described Belt Conveyor Fire method for early warning based on distributed optical fiber temperature measurement comprises the steps:

1) temperature-measuring optical fiber is divided into point for measuring temperature, passage, region, utilize absolute temperature early warning method, the poor early warning method of region relative temperature, region normal distribution early warning method, the temperature data that survey mark temperature rise slope early warning method and survey mark temperature rise variation tendency early warning method record temperature-measuring optical fiber is processed, obtain passage thresholding early warning characteristic value, region thresholding early warning characteristic value, region relative temperature early warning characteristic value, regional temperature normal distribution statistical characteristics, survey mark thresholding early warning characteristic value, survey mark temperature rise slope early warning characteristic value and survey mark temperature rise accumulative total trend early warning characteristic value,

2) by passage thresholding early warning characteristic value, region thresholding early warning characteristic value, region relative temperature early warning characteristic value, regional temperature normal distribution statistical characteristics, survey mark thresholding early warning characteristic value, survey mark temperature rise slope early warning characteristic value and survey mark temperature rise accumulative total trend early warning characteristic value input BP neural network model;

3) BP neural network model output alarm coefficient, early warning coefficient and safety factor.

Further, described BP neutral net;

Further, described BP neutral net hidden layer number is 5, is limited to 10000 on train epochs, weights coefficient is got the random number between [1,1], learning rate η=0.5, inertia correction coefficient alpha=0.2, weight convergence factor ξ=0.001 and error convergence factor-beta=0.05;

Further, described absolute temperature early warning method specifically comprises the steps: measured temperature and the predefined threshold signal of temperature-measuring system of distributed fibers to compare, once signal amplitude, surpass thresholding, output early warning characteristic value is 1, otherwise output early warning characteristic value is the ratio of measured temperature and thresholding;

Further, relative temperature poor early warning method in described region specifically comprises the steps: the region to dividing, and the temperature distribution history in this region is carried out to local maximum point retrieval; The adjacent extreme point being retrieved is compared, if adjacent pole temperature difference surpasses thresholding, exporting early warning characteristic value is 1, otherwise output early warning characteristic value is the ratio of pole temperature difference value and thresholding;

Further, it is N that described region normal distribution early warning method comprises the steps: to establish sensor fibre effective length in region, this region memory is at N temperature spot, temperature sequence x (n) in this region is carried out to descending sort, and the heavy w=5% of weighting, the temperature spikes/low-points of this sequence is respectively leached to N.w, sample average and the sample variance of utilizing filtered temperature sequence to carry out normal distribution are calculated, obtain the temperature Density Function of Normal Distribution in this region, utilize Density Function of Normal Distribution, determine that each temperature spot is at the normal distribution probability in this region, according to predetermined probabilities border limit, normal distribution probability is judged, if the normal distribution probability of certain temperature spot in this region is greater than predetermined probabilities border limit, exporting early warning characteristic value is 1, otherwise output early warning characteristic value is the ratio on distribution probability and probability border,

Further, in described survey mark temperature rise slope early warning method, the computational methods of slope adopt differential method:

$k\left(t\right)=\frac{\mathrm{dx}\left(t\right)}{\mathrm{dt}};$

The slope threshold s of survey mark is divided into two-stage s1, s2, and the value of two-stage thresholding is respectively:

s ₁＝0.11℃/min

s ₂＝0.1755℃/min

The survey mark temperature rise slope early warning characteristic value K[y (t) exporting] as follows:

$K\left[y\left(t\right)\right]=\left\{\begin{array}{ccc}1.0& k\left(t\right)\&GreaterEqual;{\mathrm{<figure-callout\; id="2"\; label="s"\; filenames="HDA0000037530890000014.png"\; state="\{\{state\}\}">s</figure-callout>}}_2& \\ 0.5& (1+\frac{k\left(t\right)-s_1}{s_2-s_1})& {\mathrm{<figure-callout\; id="2"\; label="s"\; filenames="HDA0000037530890000014.png"\; state="\{\{state\}\}">s</figure-callout>}}_2>k\left(t\right)>s_1\\ 0.0& k\left(t\right)<s_1& \end{array}\right.;$

Further, described survey mark temperature rise variation tendency early warning method specifically comprises the steps: to define the relative difference function d (n) between temperature and its stationary value:

d(n)＝x(n)-RW；

The temperature that x (n) is survey mark, Rw is the stationary value before survey mark;

Definition sum function a (n):

α(n)＝[α(n-1)+1]u(k(n-1)-s _k)；

Wherein u is unit-step function, and Sk is a predefined thresholding, and its assurance only, when survey mark temperature data changes greatly, is just carried out the computing of temperature rise accumulative total;

Definition temperature rise accumulative total function:

$g\left(n\right)=d\left(n\right)u(\mathrm{\&alpha;}\left(n\right)-N)u(\frac{23.5}{\mathrm{\&Delta;T}}-\mathrm{\&alpha;}\left(n\right));$

The value of N is:

$N=0.638\&times;\frac{23.5}{\mathrm{\&Delta;T}};$

Temperature rise accumulative total early warning feature is output as:

$A\left[g\left(n\right)\right]=\left\{\begin{array}{cc}1& g\left(n\right)>s\\ \frac{g\left(n\right)}{s}& g\left(n\right)<s\end{array}\right.;$

In setting formula, s is threshold value.

Beneficial effect of the present invention is as follows:

1, can to whole transportation lane carry out temperature distributed, continuously monitoring and early warning.

2, seven kinds of temperature pre-warnings are carried out to combination, can greatly improve the accuracy of early warning.

Accompanying drawing explanation

In order to make the object, technical solutions and advantages of the present invention clearer, below in conjunction with accompanying drawing, the present invention is described in further detail:

Fig. 1 is thermometric main frame electrical connection schematic diagram;

The temperature survey principle schematic of Fig. 2 optical fiber temperature-measurement main frame;

Fig. 3 is a kind of internal construction schematic diagram of temperature-measuring optical fiber heat conducting clamp;

Fig. 4 is the external structure schematic diagram of fixture shown in Fig. 3;

Fig. 5 is the structural representation of another kind of temperature-measuring optical fiber heat conducting clamp;

Fig. 6 is the installment state schematic diagram of heat conducting clamp;

Fig. 7 is the enlarged diagram that heat conducting clamp is arranged on V-arrangement snub pulley place;

Fig. 8 shows the Belt Conveyor Fire method for early warning schematic flow sheet based on distributed optical fiber temperature measurement.

The specific embodiment

Fig. 1 is in the Belt Conveyor Fire method for early warning based on distributed optical fiber temperature measurement, the electrical connection schematic diagram of distribution type fiber-optic, and as shown in Figure 1, temperature-measuring optical fiber main frame has mouthful output of an amp-light mouth of basis/electricity, a power input interface.Light mouth is linked into main frame through ethernet optical fiber transceiver; Electricity mouth, after safety barrier isolation, is exported intrinsic safety signal.Main frame has two-way optical fiber input measurement passage, is respectively 1,2 tunnels.The grade of the accessible 50Hz alternating voltage of power input interface is: 127V/220V/380V/660V AC; Power supply divides two-way output, and a road output 24VDC is for inner DTS main frame, and a road output 24VDC is for ethernet optical fiber transceiver.

Optical fiber temperature-measurement main frame in the present embodiment adopts the Raman time-domain reflection technology of light, its temperature survey principle as shown in Figure 2, when laser pulse is during along Optical Fiber Transmission, its rear orientation light will return transmitting terminal along optical fiber transmission, by gathering, resolve rear orientation light, extract Raman-stoke occulting light and Raman-anti-Stockton light intensity, thereby obtained optical fiber along road temperature; Measure the time that arrives transmitting terminal to scattered light simultaneously, just can calculate survey mark at light the position along road.

Fire early-warning system of the present invention is to utilize heat conducting clamp conduction key point temperature detection and the direct surveyed area environment temperature of optical fiber, both detections combine and realize rubber conveyer non-blind area temperature monitoring, realize belt conveyor fire earlier detection, early warning under coal mine.Fig. 3 is a kind of internal construction schematic diagram of temperature-measuring optical fiber heat conducting clamp, Fig. 4 is the external structure schematic diagram of this temperature-measuring optical fiber heat conducting clamp, as shown in the figure, the main body of this temperature-measuring optical fiber heat conducting clamp is that optical fiber is taken in dish 1, this optical fiber is taken in dish 1 and is adopted the metal material that thermal conductivity is good to make, as red copper etc., the medium position of taking in dish 1 at optical fiber is provided with for coiling the chamber 1a that gives out light fine, through evidence, only have and sensor fibre is coiled in to optical fiber takes in that dish is upper could obtain sufficient heat, therefore in the situation that adopting coil modes of emplacement, also must guarantee that the placement length of optical fiber in optical fiber is taken in dish will reach 1m, therefore dish is given out light, fine chamber design must meet the requirement of optical fiber minimum disc loop diameter D >=6.5cm, to meet optical fiber consume,

As further improvement, this temperature-measuring optical fiber heat conducting clamp also comprises a protection box body, this protection box body comprises upper cartridge body 2 and lower box body 3, one end of upper cartridge body 2 and lower box body 3 is hinged, the other end is closed clamping relatively, and optical fiber is taken in dish and 1 is arranged on lower box body 2 inside, in the horizontal direction of lower box body 2, be provided with optical fiber entrance with optical fiber outlet and in the bottom of lower box body 2, be provided with optical fiber and take in the heat-conducting piece 4 that dish is connected, this heat-conducting piece 4 can adopt takes in optical fiber that dish 1 identical material is made and both can be made into one, the effect of this heat-conducting piece is the temperature conduction at thermometric position to optical fiber to take in dish 1, in the base plane of lower box body 2, be provided with fastening devices for convenience detach 5, from practicality and operability, fastening devices 5 in the present embodiment has adopted magnetic suction disc, thus can be very like a cork by heat conducting clamp ontological adsorption theory at the thermometric position that needs key monitoring.

Fig. 5 is the structural representation of another kind of temperature-measuring optical fiber heat conducting clamp, as shown in the figure, this temperature-measuring optical fiber heat conducting clamp has adopted heat pipe-type structure, utilize heat pipe the heat of heat-conducting piece to be transmitted to the sensor fibre being attached on heat pipe, it has mainly comprised conducting strip 6, heat pipe 7 and fixed support device 8, in the present embodiment, temperature-measuring optical fiber is located in inside heat pipe, conducting strip 6 selects the direct laminating of copper sheet and thermometric position to carry out heat conduction, and heat pipe is connected with conducting strip 6 and heat pipe 7 is fixed on the top of fixed support device 8 by heat pipe pressing plate 7a, fixed support device 8 is fixed near thermometric position by retainer nut 9.

By accurate positioning optical waveguides, the effective optical fiber of 0.5m is arranged on heat pipe, thereby guarantees that this 0.5m optical fiber, in the scope of spatial resolution 1m, can realize the temperature rise of converting 1m optical fiber by 0.5m optical fiber.Because the heat dissipation characteristics of heat pipe self is become reconciled, therefore heat pipe and temperature-measuring optical fiber must be arranged with to windproof sheath on close at least part of naked section of heat pipe, just can guarantee that sensor fibre absorbs maximum heats.

Fig. 6 is heat conducting clamp installment state schematic diagram, Fig. 7 is the enlarged diagram that heat conducting clamp is arranged on V-arrangement snub pulley place, as shown in the figure, temperature-measuring optical fiber heat conducting clamp 10 can be arranged on the cylinders such as rubber belt conveyor head, tail and ribbon conveyer each carrying roller (only having marked the local circumstance that is arranged on upper supporting roller place in Fig. 6) along the line, and particular location comprises: axle head inside the represented V-type snub pulley going out in head drum bearing (ball) cover, aircraft tail cylinder end cap, tension drum end cap and Fig. 7.And the temperature sensing optical fiber that is arranged on thermometric heat conducting clamp outside is monitored rubber conveyer space environment temperature along the line, emphasis monitoring return belt below float coal temperature, tail coal piling regional temperature.

The Belt Conveyor Fire method for early warning based on distributed optical fiber temperature measurement of the present embodiment, at each thermometric position, all by temperature-measuring optical fiber heat conducting clamp, temperature-measuring optical fiber is installed, the thermal conductive surface of temperature-measuring optical fiber heat conducting clamp directly contacts with thermometric position and the temperature at this position is directly conducted to temperature-measuring optical fiber, wherein, at the crucial thermometric of ribbon conveyer position, temperature-measuring optical fiber is installed by temperature-measuring optical fiber heat conducting clamp, in practice, concrete crucial thermometric position comprises head drum bearing (ball) cover, aircraft tail cylinder end cap, tension drum end cap; Axle head inside V-type snub pulley; The described Belt Conveyor Fire method for early warning based on distributed optical fiber temperature measurement comprises the steps:

1) temperature-measuring optical fiber is divided into point for measuring temperature, passage, region, utilize absolute temperature early warning method, the poor early warning method of region relative temperature, region normal distribution early warning method, the temperature data that survey mark temperature rise slope early warning method and survey mark temperature rise variation tendency early warning method record temperature-measuring optical fiber is processed, obtain passage thresholding early warning characteristic value, region thresholding early warning characteristic value, region relative temperature early warning characteristic value, regional temperature normal distribution statistical characteristics, survey mark thresholding early warning characteristic value, survey mark temperature rise slope early warning characteristic value and survey mark temperature rise accumulative total trend early warning characteristic value,

2) by passage thresholding early warning characteristic value, region thresholding early warning characteristic value, region relative temperature early warning characteristic value, regional temperature normal distribution statistical characteristics, survey mark thresholding early warning characteristic value, survey mark temperature rise slope early warning characteristic value and survey mark temperature rise accumulative total trend early warning characteristic value input BP neural network model;

3) BP neural network model output alarm coefficient, early warning coefficient and safety factor.Described BP neutral net hidden layer number is 5, is limited to 10000 on train epochs, and weights coefficient gets [1,1] random number between, learning rate η=0.5, inertia correction coefficient alpha=0.2, weight convergence factor ξ=0.001 and error convergence factor-beta=0.05.

Step 2) in, being described as follows of each warning algorithm:

Absolute temperature early warning method

This algorithm compares measured temperature and the predefined threshold signal of temperature-measuring system of distributed fibers, once signal amplitude, surpasses thresholding, just exports early warning characteristic signal.

The Threshold detection method algorithm of fibre optic temperature sensor: the temperature of establishing fibre optic temperature sensor is x (t), near environment temperature this temperature pick up is h (t), comparison signal is y (t), and T is transforming function transformation function (heat conduction model), and it can be expressed as:

y(t)＝T[x(t)-h(t)]+x(t) (1-1)

$D\left[y\left(t\right)\right]=\left\{\begin{array}{cc}1& y\left(t\right)>s\\ y\left(t\right)/s& y\left(t\right)\&le;s\end{array}\right.---(1-2)$

D[y (t) wherein]=1 represent that temperature rise transfinites.S in formula is divided into Three Estate, and it is first order early warning line that s1=50 is set respectively; S2=70 is second level early warning line, and s3=90 is third level early warning line.And for the non-fibre optic temperature sensor survey mark in temperature-measuring system of distributed fibers, environment temperature point h (t) algorithm is as follows:

$D\left[h\left(t\right)\right]=\left\{\begin{array}{cc}1& h\left(t\right)>e\\ h\left(t\right)/e& h\left(t\right)\&le;e\end{array}\right.---(1-3)$

Parameter e in formula has represented the threshold value of environment temperature, and the setting of parameter e is divided into passage and region two-stage.Passage thresholding is set to e _t=80, D[h (t)]=1 represent channel environment temperature over-range, D[h (t)]=h (t)/er represents the ratio between channel environment temperature and thresholding; Regional temperature thresholding e _raccording to the otherness of environment, its parameter setting is generally less than e _t; Specifically arrange and need determine according to regional temperature characteristic.

Region relative temperature algorithm

Region relative temperature algorithm is applicable to a certain specific region that rubber conveyer has been divided, and the temperature distribution history in this region is carried out to local maximum point retrieval; The adjacent extreme point being retrieved is compared.If certain on temperature distribution history is some x (n), its condition that meets extreme value is as follows:

$\left\{\begin{array}{c}x\left(n\right)>x(n-1)\\ x\left(n\right)>x(n+1)\end{array}\right.---(1-4)$

The extreme point x (n) that meets this condition is designated as to f (n).Due to the continuity of Temperature Distribution, if adjacent pole temperature difference surpasses thresholding s, meet formula (1-4), export early warning characteristic value.

$D\left[f\left(n\right)\right]=\left\{\begin{array}{cc}1& [f\left(n\right)-f(n-1)>s]\left|\right[f\left(n\right)-f(n-1)>s]\\ \max /s& \max (f\left(n\right)-f(n-1),f\left(n\right)-f(n-1))\&le;s\end{array}\right.---(1-5)$

Region relative threshold s is traditionally arranged to be 4, D[f (n)]=1, represent that region relative temperature transfinites; D[f (n)]=max/s, the ratio of expression region relative temperature and thresholding.When if region relative temperature transfinites, need to calculate this temperature x (n) at the shared weighted value A of this regional temperature thresholding et _w[x (n)], that is:

$A_w\left[x\left(n\right)\right]=\left\{\begin{array}{cc}\frac{x\left(n\right)}{e_r}& 0<x\left(n\right)<e_r\\ 1& x\left(n\right)\&GreaterEqual;e_r>0\end{array}\right.---(1-6)$

A in formula _wthe span of [x (n)] be (0,1].Simultaneously by D[f (n)] and A _w[x (n)] is as the feature early warning output of region relative temperature algorithm.

Regional temperature normal distribution warning algorithm

Suppose that in this region, sensor fibre effective length is N, this region memory is at N temperature spot, and wherein N is at least greater than 100.In order to filter the impact of noise, the temperature sequence x (n) in this region is carried out to descending sort, and the heavy w=5% of weighting, the temperature spikes/low-points of this sequence is respectively leached to N.w.Sample average and the sample variance of utilizing filtered temperature sequence to carry out normal distribution are calculated.

$\stackrel{\&OverBar;}{E}=\underset{n=i}{\overset{N(1-2w)}{\mathrm{\&Sigma;}}}x\left(n\right)---(1-7)$

$S_{\mathrm{\&sigma;}2}=\frac{1}{N(1-2w)-1}\underset{n=i}{\overset{N(1-2w)}{\mathrm{\&Sigma;}}}{[x\left(n\right)-\stackrel{\&OverBar;}{E}]}^2---(1-8)$

Thereby the temperature Density Function of Normal Distribution that obtains this region is as follows:

$f\left[x\left(n\right)\right]=\frac{1}{\sqrt{2\mathrm{\&pi;}{S}_{\mathrm{\&sigma;}2}}}\exp [-\frac{{[x\left(n\right)-\stackrel{\&OverBar;}{E}]}^2}{2S_{\mathrm{\&sigma;}2}}]---(1-9)$

Utilize Density Function of Normal Distribution, determine that each temperature spot is at the normal distribution probability f[x in this region (n)], the empirical probability border limit p=0.01 obtaining according to experiment, determines whether warning, that is:

As f[x (n)] while being greater than p, S[f[x (n)]]=1 output of the feature early warning as regional temperature normal distribution warning algorithm.

Survey mark temperature rise slope warning algorithm

The computational methods of slope adopt differential method:

$k\left(t\right)=\frac{\mathrm{dx}\left(t\right)}{\mathrm{dt}}---(1-11)$

In practice, slope is generally used Δ X (t)=X (t ₂)-X (t ₁), Δ T=t ₂-t ₁time in temperature-measuring system of distributed fibers sampling interval, t ₂for current time, t ₁for previous moment.The reliability and the antijamming capability that in order to improve temperature slope, calculate, carried out average and delay process by survey mark temperature data.

Slope threshold s for survey mark is divided into two-stage s1, s2, and the value of two-stage thresholding is respectively:

s ₁＝0.11(℃/min)

s ₂＝0.1755(℃/min)

Survey mark temperature rise slope early warning feature is exported K[y (t)] as follows:

$K\left[y\left(t\right)\right]=\left\{\begin{array}{ccc}1.0& k\left(t\right)\&GreaterEqual;s_2& \\ 0.5& (1+\frac{k\left(t\right)-s_1}{s_2-s_1})& {\mathrm{<figure-callout\; id="2"\; label="s"\; filenames="HDA0000037530890000014.png"\; state="\{\{state\}\}">s</figure-callout>}}_2>k\left(t\right)>s_1\\ 0.0& k\left(t\right)<s_1& \end{array}\right.---(1-12)$

Survey mark temperature rise variation tendency algorithm

The temperature of survey mark is x (n), and the stationary value before survey mark is Rw, the relative difference function d (n) between definition temperature and its stationary value.

d(n)＝x(n)-RW (1-13)

In order to overcome the interference such as noise, Cai consider that survey mark temperature data only has while significantly changing, calculate meaningfully, introduced a sum function a (n):

α(n)＝[α(n-1)+1]u(k(n-1)-s _k) (1-14)

Wherein u is unit-step function, Sk=0.11 (℃/min) be a predefined thresholding, its assurance only, when survey mark temperature data changes greatly, is just carried out the computing of temperature rise accumulative total.Can define temperature rise accumulative total function:

$g\left(n\right)=d\left(n\right)u(\mathrm{\&alpha;}\left(n\right)-N)u(\frac{23.5}{\mathrm{\&Delta;T}}-\mathrm{\&alpha;}\left(n\right))---(1-15)$

N is a value relevant with calculating temperature rise accumulative total siding-to-siding block length.Only when α (n) >=N, calculate temperature rise accumulative total, as α (n) >=23.5/ Δ T, g (n)=0, reenters the next accumulative total cycle.Choosing the temperature rise cumulative time is 23.5min, and Δ T is the time in temperature data sampling interval, adds certain redundancy, and the value of N is at present:

$N=0.638\&times;\frac{23.5}{\mathrm{\&Delta;T}}---(1-16)$

Temperature rise accumulative total early warning feature is output as:

$A\left[g\left(n\right)\right]=\left\{\begin{array}{cc}1& g\left(n\right)>s\\ \frac{g\left(n\right)}{s}& g\left(n\right)<s\end{array}\right.---(1-17)$

S=25 in setting formula, A[g (n)] for the output of temperature rise accumulative total early warning feature, for characterizing the weighted value of temperature rise accumulative total degree.

The foregoing is only the present invention that is preferably not limited to of the present invention, obviously, those skilled in the art can carry out various changes and modification and not depart from the spirit and scope of the present invention the present invention.Like this, if within of the present invention these are revised and modification belongs to the scope of the claims in the present invention and equivalent technologies thereof, the present invention is also intended to comprise these changes and modification interior.

Claims (7)

1. the Belt Conveyor Fire method for early warning based on distributed optical fiber temperature measurement, it is characterized in that: be arranged on ribbon conveyer temperature-measuring optical fiber is distributed, wherein, at the crucial thermometric of ribbon conveyer position, temperature-measuring optical fiber is installed by temperature-measuring optical fiber heat conducting clamp, and the described Belt Conveyor Fire method for early warning based on distributed optical fiber temperature measurement comprises the steps:

1) temperature-measuring optical fiber is divided into point for measuring temperature, passage, region, utilize absolute temperature early warning method, the poor early warning method of region relative temperature, region normal distribution early warning method, the temperature data that survey mark temperature rise slope early warning method and survey mark temperature rise variation tendency early warning method record temperature-measuring optical fiber is processed, obtain passage thresholding early warning characteristic value, region thresholding early warning characteristic value, region relative temperature early warning characteristic value, regional temperature normal distribution statistical characteristics, survey mark thresholding early warning characteristic value, survey mark temperature rise slope early warning characteristic value and survey mark temperature rise accumulative total trend early warning characteristic value,

2) by passage thresholding early warning characteristic value, region thresholding early warning characteristic value, region relative temperature early warning characteristic value, regional temperature normal distribution statistical characteristics, survey mark thresholding early warning characteristic value, survey mark temperature rise slope early warning characteristic value and survey mark temperature rise accumulative total trend early warning characteristic value input BP neural network model;

3) BP neural network model output alarm coefficient, early warning coefficient and safety factor.

2. a kind of Belt Conveyor Fire method for early warning based on distributed optical fiber temperature measurement as claimed in claim 1, it is characterized in that: described BP neutral net hidden layer number is 5, is limited to 10000 on train epochs, and weights coefficient gets [1,1] random number between, learning rate

=0.5, inertia correction coefficient

=0.2, the weight convergence factor

=0.001 and the error convergence factor

=0.05.

3. a kind of Belt Conveyor Fire method for early warning based on distributed optical fiber temperature measurement as claimed in claim 1 or 2, it is characterized in that: described absolute temperature early warning method specifically comprises the steps: measured temperature and the predefined threshold signal of temperature-measuring system of distributed fibers to compare, once signal amplitude, surpass thresholding, output early warning characteristic value is 1, otherwise output early warning characteristic value is the ratio of measured temperature and thresholding.

4. a kind of Belt Conveyor Fire method for early warning based on distributed optical fiber temperature measurement as claimed in claim 1 or 2, it is characterized in that: relative temperature poor early warning method in described region specifically comprises the steps: the region to dividing, and the temperature distribution history in this region is carried out to local maximum point retrieval; The adjacent extreme point being retrieved is compared, if adjacent pole temperature difference surpasses thresholding, exporting early warning characteristic value is 1, otherwise output early warning characteristic value is the ratio of pole temperature difference value and thresholding.

5. a kind of Belt Conveyor Fire method for early warning based on distributed optical fiber temperature measurement as claimed in claim 1 or 2, it is characterized in that: it is N that described region normal distribution early warning method comprises the steps: to establish sensor fibre effective length in region, this region memory is at N temperature spot, by the temperature sequence x(n in this region) carry out descending sort, and the heavy w=5% of weighting, the temperature spikes/low-points of this sequence is respectively leached

sample average and the sample variance of utilizing filtered temperature sequence to carry out normal distribution are calculated, obtain the temperature Density Function of Normal Distribution in this region, utilize Density Function of Normal Distribution, determine that each temperature spot is at the normal distribution probability in this region, according to predetermined probabilities border limit, normal distribution probability is judged, if the normal distribution probability of certain temperature spot in this region is greater than predetermined probabilities border limit, exporting early warning characteristic value is 1, otherwise output early warning characteristic value is the ratio on distribution probability and probability border.

6. a kind of Belt Conveyor Fire method for early warning based on distributed optical fiber temperature measurement as claimed in claim 1 or 2, is characterized in that: in described survey mark temperature rise slope early warning method, the computational methods of slope adopt differential method:

；

The slope threshold s of survey mark is divided into two-stage s1, s2, and the value of two-stage thresholding is respectively:

=0.11 ℃/min

=0.1755 ℃/min

The survey mark temperature rise slope early warning characteristic value of exporting

as follows:

。

7. a kind of Belt Conveyor Fire method for early warning based on distributed optical fiber temperature measurement as claimed in claim 2, is characterized in that: described survey mark temperature rise variation tendency early warning method specifically comprises the steps: to define the relative difference function d (n) between temperature and its stationary value:

；

The temperature that x (n) is survey mark, Rw is the stationary value before survey mark;

Definition sum function a (n):

；

Wherein u is unit-step function, and Sk is a predefined thresholding, and its assurance only, when survey mark temperature data changes greatly, is just carried out the computing of temperature rise accumulative total;

Definition temperature rise accumulative total function:

；

The value of N is:

；

Temperature rise accumulative total early warning feature is output as:

；

In setting formula, s is threshold value.

Images