CN101825523A - Wet method ball mill belly expansion fault prediction device and method - Google Patents

Abstract

The invention relates to a wet method ball mill belly expansion fault prediction device and a method. The invention is characterized in that the device comprises a vibration signal detecting device and a monitoring computer, wherein the vibration signal detecting device is arranged on the surface of a tube wall of a ball mill and has the wireless communication function, the monitoring computer is arranged near the ball mill and has the wireless communication function, an acceleration sensor, an analog-to-digital conversion unit, a central processing unit, a data storage unit and a wireless data sending unit are prearranged in the vibration signal detecting device, a wireless data receiving unit, a ball mill belly expansion fault model training unit and a ball mill belly expansion fault prediction unit are prearranged in the monitoring computer, the central processing unit controls the analog-to-digital conversion unit to carry out analog-to-digital conversion on vibration acceleration signals collected by the acceleration sensor and to store the vibration acceleration signals into the data storage unit, the central processing unit calculates the data stored in the data storage unit, and sends the data to the wireless data receiving unit through the wireless data sending unit, and the wireless data receiving unit transmits the received data to the belly expansion fault model training unit and the fault prediction unit.

Description

A kind of ball mill by wet process belly expansion fault prediction device and method

Technical field

The present invention relates to a kind of fault prediction device and method, particularly about a kind of ball mill by wet process belly expansion fault prediction device and method.

Background technology

Bowl mill is an important milling equipment during mineral resources are produced, and its major function is that ore is ground into the needed granularity of subsequent production technology.Bowl mill is divided into two kinds of dry method ball mill and ball mill by wet process, and the key distinction is whether need to add water during ore grinding as middle break-in medium, and the discharge mode of ore in the bowl mill.Need not add water during dry method ball mill work, the ore that grinds in time be discharged bowl mill by air blast; Need add water during ball mill by wet process work, utilize the flowability and the interior angle of inclination of bowl mill that grind ore and the composite ore pulp of water that the ore that grinds is discharged bowl mill by sieve aperture.

Under the normal operating conditions, the material total amount in the bowl mill should remain on comparatively within the reasonable range, and the inventory that enters bowl mill should equate with the inventory that bowl mill is discharged.If the material total amount is too many in the bowl mill, surpass the maximum tonnage capacity that bowl mill itself allows, and made material to discharge, when being full of bowl mill, can make bowl mill lose the ore grinding ability, " tripe rises " fault takes place.The tripe fault that rises not only influences the ordinary production of ore grinding and subsequent job, and can damage bowl mill, even interrupts whole mineral products production process, therefore, prevents that the bowl mill tripe fault that rises is most important.

For dry method ball mill, the increase of material will directly influence the leeway of bowl mill entrance and exit in the bowl mill, therefore can in time find whether stop up in the bowl mill by detecting the entrance and exit leeway, the tripe thereby whether the prediction bowl mill rises.Material in the ball mill by wet process is the potpourri of ore, steel ball and water, and wherein water mixes with the ore that is ground easily and discharges bowl mill as middle break-in medium, will not be accumulated in the bowl mill and grind ore.Therefore, than dry method ball mill, ball mill by wet process is difficult to predict the tripe fault that whether rises by detecting comparatively intuitively physical quantity.

Cause the rise factor of tripe fault of ball mill by wet process more, comprise the sand return amount of steel ball amount in the bowl mill, raw ore mine-supplying quantity, bowl mill inlet confluent, stage equipment and ore hardness etc.In actual mechanical process, the ore grinding workman needs often to observe the ball mill by wet process feed port and whether tells material, whether the ore grinding noise changes, phenomenons such as whether the bowl mill drive current reduces suddenly, the sign of tripe fault comprehensively judges whether to rise, this has greatly increased the ore grinding working strength of workers, and can be because of ore grinding workman's working experience and sense of responsibility different, the generation of the tripe fault that is difficult to avoid fully to rise.

According to ball mill by wet process operator's experience, also there are some devices to be used to the detection and the forecast of ball mill belly expansion fault at present, for example ore grinding noise measuring and bowl mill drive current detect.Because the operating noise of the on-the-spot miscellaneous equipment of ore grinding is very big, noise detection device is subject to disturb, so effect is relatively poor.And ball mill by wet process is only when rising by tripe, in the bowl mill material just along with barrel is done centrifugal motion, thereby cause the drive motor acting to die-off, so drive current detects the prediction effect that can not play the tripe fault that rises.Therefore, for ball mill by wet process, also lack the effective belly expansion fault prediction method and apparatus of practical application at present.

Summary of the invention

At the problems referred to above, the purpose of this invention is to provide a kind of ball mill by wet process belly expansion fault prediction device and method, this method and device can be in bowl mill operations, predict the bowl mill tripe fault that whether will rise in real time, rapidly and accurately.

For achieving the above object, the present invention takes following technical scheme: a kind of ball mill by wet process belly expansion fault prediction device, it is characterized in that: it comprises that one is arranged on the vibration signal pick-up unit with wireless communication function on bowl mill barrel surface, and one is arranged near the supervisory control comuter with wireless communication function the bowl mill; Preset acceleration transducer, analog digital converting unit, central processing unit, data storage cell and wireless data transmission unit in the described vibration signal pick-up unit; Preset wireless data receiving element, ball mill belly expansion fault model training unit and ball mill belly expansion fault prediction unit in the described supervisory control comuter; Described central processing unit is controlled the vibration acceleration signal that described analog digital converting unit collects described acceleration transducer and is carried out the digital-to-analog conversion sampling, and be stored to described data storage cell, described central processing unit is to calculating through the vibration acceleration sampled value after the digital-to-analog conversion simultaneously, obtain the average intensity value of vibration signal, average intensity value is sent to the described wireless data receiving element that presets in the described supervisory control comuter by described wireless data transmission unit then; Described wireless data receiving element sends the average intensity value of receiving to ball mill belly expansion fault model training unit, obtains training pattern

And the model that obtains is preset in the described ball mill belly expansion fault prediction unit, described wireless data receiving element directly sends the follow-up average intensity value of receiving to described ball mill belly expansion fault prediction unit, carries out the evaluation prediction of ball mill belly expansion fault.

The low-limit frequency response of described acceleration transducer is less than the rotating speed of bowl mill, and the highest frequency response of described acceleration transducer is higher than the highest frequency of steel-ball collision barrel generation vibration acceleration signal in the bowl mill.

The sensitive axes of described acceleration transducer is vertical with the bowl mill barrel at its place, installation site.

A kind of ball mill by wet process belly expansion fault prediction method of said apparatus, it may further comprise the steps: 1) a vibration signal pick-up unit with wireless communication function is set on bowl mill barrel surface, a supervisory control comuter with wireless communication function is set near bowl mill; 2) analog digital converting unit that presets in the vibration signal pick-up unit is set, the sample frequency that the vibration acceleration signal that acceleration transducer is collected carries out the analog digital conversion, and the sensitive axes direction position identical with gravity direction of the acceleration transducer that presets in the vibration signal pick-up unit be designated as the A point, the position opposite with gravity direction is designated as the C point; 3) the interior ore total amount w of bowl mill is set _oThe steel ball amount and the water yield, bowl mill work, steel ball constantly collides barrel and produces vibration acceleration signal x in the cylindrical shell, the central processing unit that presets in vibration signal pick-up unit control analog digital converting unit degree of will speed up sensor acquisition to vibration acceleration signal carry out the digital-to-analog conversion sampling, and be stored to data storage cell, central processing unit is according to the vibration acceleration signal sampled value through digital-to-analog conversion of storing in the data storage cell, calculate the average intensity value β of the vibration acceleration signal that the bowl mill barrel ordered to A by the C point, and average intensity value β is sent to supervisory control comuter by wireless data transmission unit; 4) the wireless data receiving element in the supervisory control comuter receives the average intensity value β of the bowl mill barrel vibration acceleration signal that is sent by wireless data transmission unit in the vibration signal pick-up unit, and send ball mill belly expansion fault model training unit and ball mill belly expansion fault prediction unit to, be respectively applied for the foundation and the failure prediction of the tripe fault model that rises; 5) constantly increase the interior ore total amount of bowl mill, repeating step 3) and step 4), and in ball mill belly expansion fault model training unit under the different ore total amounts of record, the vibration signal average intensity value that corresponding bowl mill barrel is ordered to A by the C point, constitute sample data group { β (i), w _o(i) }, i=1 wherein, 2 ..., l is for increasing the experiment group of bowl mill raw ore mine-supplying quantity; K group experiment (during 1＜k＜l), w _oThe upper limit of ore total amount when (k) being the bowl mill operate as normal, when the experiment of l group, the ore total amount in the bowl mill is w _o(l), bowl mill is in the tripe malfunction that rises; When in the ball mill belly expansion fault prediction unit, noting experiment of k group and the experiment of l group respectively, the interior corresponding ore total amount w of bowl mill _o(k), w _o(l); 6) in ball mill belly expansion fault model training unit, preset ore total amount estimated value in the bowl mill

The quadratic polynomial model:

${\hat{w}}_o=p\left(\mathrm{\β}\right)=a_0+a_1\mathrm{\β}+a_2{\mathrm{\β}}^2---\left(1\right)$

Wherein,

Be ore total amount w in the bowl mill _oEstimated value, calculate by β; a ₀, a ₁, a ₂Be undetermined parameter, according to sample data group { β (i), w _o(i) }, try to achieve undetermined parameter a ₀, a ₁, a ₂, determine formula (1), ball mill belly expansion fault model training unit sends the formula of determining (1) to the ball mill belly expansion fault prediction unit, and stores, to be used for the prediction of ball mill belly expansion fault; 7) bowl mill man-hour, the vibration signal pick-up unit detects and calculates the average intensity value β of vibration acceleration signal in real time, and average intensity value β sent to supervisory control comuter by wireless data transmission unit, ball mill belly expansion fault prediction unit in the supervisory control comuter is brought average intensity value β in the formula of determining in the step 6) (1) into, calculates the estimated value of ore total amount in the bowl mill With this estimated value The upper limit w of ore total amount during with the bowl mill operate as normal _o(k) compare, thereby whether the ore total amount of in time judging in the bowl mill is in normal range, if And continue to increase near w _o(l), then forecast ore total amount overrun in the bowl mill, will the tripe fault that rises occur.

The concrete computing method of described step 3) are: 1. in central processing unit, put storage mark position SP=0, and slip mean filter parameter N is set; 2. detect the vibration acceleration signal of bowl mill barrel by acceleration transducer, and be vibration acceleration signal sampled value x (k), k=1 by the analog digital converting unit digital-to-analog conversion, 2,, k is counting of vibration acceleration signal, and is stored in the data storage cell; 3. central processing unit is according to the vibration acceleration signal sampled value x (k) that stores in the data storage cell, k=1, and 2 ..., calculate the acceleration a that acceleration transducer produces because of gravity by the slip mean filter _g(k), when k＜N,

$a_g\left(k\right)=\frac{1}{k}\underset{i=1}{\overset{k}{\mathrm{\Σ}}}x\left(i\right)$

When the N of k,

$a_g\left(k\right)=\frac{1}{N}\underset{i=k-N+1}{\overset{k}{\mathrm{\Σ}}}x\left(i\right)$

And obtain a respectively _g(k) maximal value a _{G max}With minimum value a _{G min}4. judge by central processing unit whether the storage mark position is SP=0, if SP=0 then judges a _g(k) whether be minimum value, if a _g(k) non-minimum value a _{G min}, then directly return step 2.; Work as a _g(k) be minimum value a _{G min}The time, illustrate that acceleration transducer is in the C point, then is provided with storage mark position SP=1, and returns step 2.; If 5. storage mark position SP=1 then judges a _g(k) whether be maximal value a _{G max}If: a _g(k) non-maximal value a _{G max}, illustrate acceleration transducer at the C point between the A point, then calculate

Be vibration acceleration signal x (k), k=1,2 ... deduct the acceleration a that acceleration transducer produces because of gravity _g(k) value after, promptly

And will Store in the data storage cell, return step 2.; If 6. storage mark position SP=1, and a _g(k) be maximal value, illustrate that acceleration transducer is positioned at the A point, SP=1 is made as SP=0 with the storage mark position, calculate the bowl mill barrel by the C point to the vibration signal average intensity value β between the A point,

$\mathrm{\β}=\frac{1}{M}\underset{j=1}{\overset{M}{\mathrm{\Σ}}}\left|\stackrel{\‾}{x}\left(j\right)\right|$

Wherein, M is that the bowl mill barrel is C point counting to the vibration signal sampled value between the A point; 7. by wireless data transmission unit vibration signal average intensity value β is sent to supervisory control comuter, and the data in the storage unit that clears data, return step 2., infinite loop quits work until bowl mill.

Described step 1. in, the span of slip mean filter parameter N is 500≤N≤1000.

Described step 2) in, the sample frequency f that the vibration acceleration signal that the analog digital converting unit that presets in the vibration signal pick-up unit collects acceleration transducer carries out digital-to-analog conversion is set to f＞2f _Max, f _MaxProduce the highest frequency of vibration acceleration signal for steel-ball collision barrel in the bowl mill.

Undetermined parameter a in the described formula (1) ₀, a ₁, a ₂Try to achieve, should make ore total amount w in the bowl mill _oWith ore total amount estimated value in the bowl mill

Difference Q minimum, that is:

$Q_{\min }=\underset{i=1}{\overset{l}{\mathrm{\Σ}}}{(w_o\left(i\right)-{\hat{w}}_o\left(i\right))}^2=\underset{i=1}{\overset{l}{\mathrm{\Σ}}}{(w_o\left(i\right)-p\left(\mathrm{\β}\left(i\right)\right))}^2---\left(2\right)$

The present invention is owing to take above technical scheme, it has the following advantages: 1, the present invention is owing to the barrel outside surface at bowl mill is equipped with the vibration signal pick-up unit, therefore the barrel vibration acceleration signal in the time of can directly detecting the bowl mill rotating operation, detected signal can more direct reaction sphere grinding machine inside running status, detection sensitivity is higher, and have stronger antijamming capability, work is not subject to the environmental impact of production scene, has stronger practicality.2, the present invention has adopted the acceleration transducer of frequency range broad, when detecting bowl mill barrel vibration acceleration signal, can determine the zone of steel-ball collision barrel in the bowl mill easily, thereby reduced the data volume of vibration signals collecting and processing, reduced hardware cost accordingly, realized easily.3, the present invention's vibration signal pick-up unit of being installed in bowl mill barrel outside surface has wireless communication function, can carry out wireless data transmission with near the supervisory control comuter the bowl mill, can realize real-time monitoring to the bowl mill running status, avoided necessity for realizing that the transmission of barrel vibration data is transformed bowl mill, system realizes that cost is low, is easy to promote.4, the present invention handles the vibration acceleration signal that the vibration signal pick-up unit collects, and calculate the average intensity value of vibration acceleration signal, average intensity value sends supervisory control comuter to, calculate the tripe fault model that rises by ball mill belly expansion fault model training unit that presets in the supervisory control comuter and ball mill belly expansion fault prediction unit, therefore, can be according to the tripe fault model that rises, with the vibration acceleration signal that collects in real time be used to the to rise forecast analysis of tripe fault.5, the present invention only needs to be provided with the vibration signal pick-up unit on ball mill by wet process, at the bowl mill annex supervisory control comuter is installed, the tripe fault that whether can rise in the time of can working to bowl mill is predicted, and this type of bowl mill is except that using in the non-ferrous metal ore field, also be widely used in fields such as chemical industry, metallurgy, therefore, but detection method provided by the present invention and device have certain generalization.The present invention is skillfully constructed, and is accurate and practical, and the operational efficiency that can greatly improve institute's watch-dog is also effectively avoided the accident of equipment operation, therefore, can be widely used in the monitoring system in fields such as non-ferrous metal ore field, chemical industry and metallurgy.

Description of drawings

Steel-ball collision barrel position and acceleration transducer sensitive axes direction synoptic diagram when Fig. 1 is bowl mill of the present invention work

Fig. 2 is acceleration transducer of the present invention produces acceleration because of gravity a cycle synoptic diagram

Fig. 3 is a prediction unit synoptic diagram of the present invention

Fig. 4 is a preset module synoptic diagram in the vibration signal pick-up unit of the present invention

Fig. 5 is a preset module synoptic diagram in the supervisory control comuter of the present invention

Fig. 6 is that the present invention detects and calculate the bowl mill barrel by the schematic flow sheet of C point to A point vibration acceleration signal mean intensity β

Fig. 7 is the synoptic diagram that concerns that barrel vibration acceleration signal x sampled value and vibration signal were counted when the ore total amount was 22kg in the bowl mill in the embodiment of the invention

Fig. 8 is that the interior ore total amount of bowl mill is the acceleration a that produces because of gravity in the 22kg brief acceleration sensor in the embodiment of the invention _gConcern synoptic diagram with vibration signal is counted

Fig. 9 is that vibration acceleration signal x rejected in the acceleration transducer because of gravity generation acceleration a when the ore total amount was 22kg in the bowl mill in the embodiment of the invention _gAfter value

Concern synoptic diagram with vibration is counted

Figure 10 be in the embodiment of the invention in the bowl mill ore total amount when 22kg is increased to 50kg, ore total amount w in barrel vibration signal mean intensity β and the bowl mill _oCorresponding relation figure

Embodiment

Below in conjunction with drawings and Examples the present invention is described in detail.

The present invention is by being arranged on the vibration acceleration signal that acceleration transducer on ball mill by wet process (hereinafter to be referred as bowl mill) the barrel outside surface detects barrel, predicting that according to the mean intensity of vibration acceleration signal whether bowl mill is with the tripe fault that rises.

The acceleration transducer that the present invention adopts satisfies following the requirement: the low-limit frequency response of acceleration transducer

(Hz of unit) is lower than the rotational speed omega of bowl mill _r(rotating speed unit is converted to " revolutions per second ", so that compare with cps Hz), the highest frequency response

(Hz of unit) is higher than the highest frequency f of steel-ball collision barrel generation vibration acceleration signal in the bowl mill _Max(Hz of unit).

The present invention is based on following principle: as shown in Figure 1 and Figure 2, owing to be installed in the low-limit frequency response of the acceleration transducer of bowl mill barrel outside surface

Rotational speed omega less than bowl mill _r, and acceleration transducer is along with bowl mill barrel rotation, the angle of its sensitive axes direction and gravity direction is cyclical variation, when the sensitive axes direction is opposite with gravity direction, shown in the C point, in the acceleration transducer because of the acceleration a of gravity generation _gReach minimum; When the sensitive axes direction is identical with gravity direction, shown in the A point, the acceleration a that produces because of gravity in the acceleration transducer _gReach maximum; And the position of bowl mill steel-ball collision barrel in rotary course, as the C point to shown in the A point, when material in the bowl mill continues to increase the trend that the tripe fault that rises is arranged, C point to the mean intensity of A point vibration acceleration signal will continue to reduce, therefore, when bowl mill is worked, detect and calculate the average intensity value β of C point to A point vibration acceleration signal, whether the material total amount that can directly reflect bowl mill inside is in normal condition, thereby whether the prediction bowl mill is with the tripe fault that rises.

Concrete prediction steps of the present invention is as follows:

1) as shown in Figure 3, barrel outside surface at bowl mill 1 is installed a vibration signal pick-up unit 2 with wireless communication function, one supervisory control comuter 3 with wireless communication function is installed near bowl mill 1, and supervisory control comuter 3 is used to receive the data that vibration signal pick-up unit 2 transmits.As shown in Figure 4, preset the acceleration transducer 21, analog digital converting unit 22, central processing unit 23, data storage cell 24 and the wireless data transmission unit 25 that meet above-mentioned requirements in the vibration signal pick-up unit 2, wherein, the sensitive axes of acceleration transducer 21 is vertical with bowl mill 1 barrel at its place, installation site.As shown in Figure 5, preset wireless data receiving element 31, ball mill belly expansion fault model training unit 32 and ball mill belly expansion fault prediction unit 33 in the supervisory control comuter 3.

2) in vibration signal pick-up unit 2, the sample frequency f that vibration acceleration signal that 22 pairs of acceleration transducers 21 of analog digital converting unit collect carries out digital-to-analog conversion is set, be f＞2f _Max

3) ore total amount w in the bowl mill 1 is set _o, steel ball amount and the water yield.Bowl mill 1 operate as normal, steel ball constantly collides barrel and produces vibration acceleration signal x in the cylindrical shell, acceleration transducer 21 in the vibration signal pick-up unit 2 is gathered vibration acceleration signal, central processing unit 23 control analog digital converting unit 22 are carried out the digital-to-analog conversion sampling with the vibration acceleration signal that collects, and be stored in the data storage cell 24, central processing unit 23 is according to the vibration acceleration signal sampled value after digital-to-analog conversion of storage in the data storage cell 24, calculate the average intensity value β of the vibration acceleration signal that bowl mill 1 barrel ordered to A by the C point, and average intensity value β is sent to supervisory control comuter 3 by wireless data transmission unit 25.

4) the wireless data receiving element in the supervisory control comuter 3 31 receives the average intensity value β of the bowl mill barrel vibration acceleration signal that wireless data transmission unit 25 sends in the vibration signal pick-up unit 2, and send ball mill belly expansion fault model training unit 32 and ball mill belly expansion fault prediction unit 33 to, to be respectively applied for the foundation and the failure prediction of the tripe fault model that rises.

5) constantly increase ore total amount in the bowl mill 1, repeating step 3) and step 4), and in ball mill belly expansion fault model training unit 32 under the different ore total amounts of record, the vibration signal average intensity value that corresponding barrel is ordered to A by the C point, constitute sample data group { β (i), w _o(i) }, i=1 wherein, 2 ..., l is for increasing the experiment group of bowl mill raw ore mine-supplying quantity.Wherein, (during 1＜k＜l), the ore total amount is w in the bowl mill in k group experiment _o(k), w _oThe upper limit of ore total amount when (k) being the bowl mill operate as normal; When the experiment of l group, the ore total amount in the bowl mill is w _o(l), this moment, bowl mill was in the tripe malfunction that rises.Corresponding ore total amount w in the bowl mill when in ball mill belly expansion fault prediction unit 33, noting experiment of k group and the experiment of l group respectively _o(k), w _o(l).

6) in ball mill belly expansion fault model training unit 32, preset ore total amount w in the bowl mill _oEstimated value The quadratic polynomial model:

${\hat{w}}_o=p\left(\mathrm{\β}\right)=a_0+a_1\mathrm{\β}+a_2{\mathrm{\β}}^2---\left(1\right)$

Wherein,

Be ore total amount w in the bowl mill _oEstimated value, calculate by β; a ₀, a ₁, a ₂Be undetermined parameter, a ₀, a ₁, a ₂Try to achieve and should make ore total amount w in the bowl mill _oWith ore total amount estimated value in the bowl mill

Difference Q minimum, that is:

$Q_{\min }=\underset{i=1}{\overset{l}{\mathrm{\Σ}}}{(w_o\left(i\right)-{\hat{w}}_o\left(i\right))}^2=\underset{i=1}{\overset{l}{\mathrm{\Σ}}}{(w_o\left(i\right)-p\left(\mathrm{\β}\left(i\right)\right))}^2---\left(2\right)$

According to sample data group { β (i), w _o(i) } and formula (2) try to achieve undetermined parameter a ₀, a ₁, a ₂, determine formula (1), ball mill belly expansion fault model training unit 32 sends the formula of determining (1) to ball mill belly expansion fault prediction unit 33, and stores, to be used for the prediction of ball mill belly expansion fault.

7) bowl mill 1 work, vibration signal pick-up unit 2 detects and calculates the average intensity value β of vibration acceleration signal in real time, and average intensity value β sent to supervisory control comuter 3 by wireless data transmission unit 25, ball mill belly expansion fault prediction unit 33 in the supervisory control comuter 3 is brought mean intensity β value in the formula of determining in the step 6) (1) into, calculates the estimated value of ore total amount in the bowl mill

With this estimated value

The upper limit w of ore total amount during with the bowl mill operate as normal _o(k) compare, thereby whether the ore total amount of in time judging in the bowl mill 1 is in normal range.If

And continue to increase near w _o(l), then forecast ore total amount overrun in the bowl mill 1, will the tripe fault that rises occur.

As shown in Figure 6, above-mentioned steps 3) concrete computing method be:

1. in central processing unit 23, put storage mark position SP=0, and slip mean filter parameter N is set,, can select 500≤N≤1000 usually in order to obtain filter effect preferably.

2. detect the vibration acceleration signal of bowl mill barrels by acceleration transducer 21, and be vibration acceleration signal sampled value x (k), k=1 by analog digital converting unit 22 digital-to-analog conversions, 2,, k is counting of vibration acceleration signal, and is stored in the data storage cell 24.

3. central processing unit 23 according in the data storage cell 24 storage vibration acceleration signal sampled value x (k), k=1,2 ..., calculate the acceleration a that acceleration transducer 21 produces because of gravity by the slip mean filter _g(k), when k＜N,

$a_g\left(k\right)=\frac{1}{k}\underset{i=1}{\overset{k}{\mathrm{\Σ}}}x\left(i\right)$

When k 〉=N,

$a_g\left(k\right)=\frac{1}{N}\underset{i=k-N+1}{\overset{k}{\mathrm{\Σ}}}x\left(i\right)$

And obtain a respectively _g(k) maximal value a _{G max}With minimum value a _{G min}

4. judge by central processing unit 23 whether the storage mark position is SP=0, if SP=0 then judges a _g(k) whether be minimum value, if a _g(k) non-minimum value a _{G min}, then directly return step 2.; Work as a _g(k) be minimum value a _{G min}The time, illustrate that acceleration transducer 21 is in the C point shown in Fig. 1, then is provided with storage mark position SP=1, and returns step 2..

If 5. storage mark position SP=1 then judges a _g(k) whether be maximal value a _{G max}If: a _g(k) non-maximal value a _{G max}, illustrate acceleration transducer 21 at the point of the C shown in Fig. 1 between the A point, then calculate

Be vibration acceleration signal x (k), k=1,2 ... deduct the acceleration a that acceleration transducer 21 produces because of gravity _g(k) value after, promptly

And will

Store data storage cell 24 into, return step 2..

If 6. storage mark position SP=1, and a _g(k) be maximal value, illustrate that acceleration transducer 21 is arranged in the A point of Fig. 1, SP=1 is made as SP=0 with the storage mark position, calculate the bowl mill barrel by the C point to the vibration signal average intensity value β between the A point,

$\mathrm{\β}=\frac{1}{M}\underset{j=1}{\overset{M}{\mathrm{\Σ}}}\left|\stackrel{\‾}{x}\left(j\right)\right|$

Wherein, M is that the bowl mill barrel is C point counting to the vibration signal sampled value between the A point.

7. by wireless data transmission unit 25 vibration signal average intensity value β is sent to supervisory control comuter 3, and the data in the storage unit 24 that clears data, return step 2., infinite loop quits work until bowl mill.

Enumerate a specific embodiment below:

Select XMQL-Φ 420 * 450 ball mill by wet process among this embodiment, during this bowl mill operate as normal, the ore total amount is 20kg～30kg in the grinding machine, and steel ball is 40kg, and water is 10kg, and drum's speed of rotation is ω _r=57rpm is ω _r=0.95 revolutions per second, be converted to cps and correspond to 0.95Hz, steel-ball collision barrel vibration signal highest frequency f _Max≈ 10000Hz.

1) the vibration signal pick-up unit 2 that will have a wireless communication function is installed in the barrel outside surface of bowl mill 1, and the supervisory control comuter 3 with wireless communication function is installed near bowl mill.

2) selecting frequency response range in the barrel vibration signal pick-up unit 2 is 0.2Hz～12000Hz, and range is ± 2500m/s ²Acceleration transducer so that make

And analog digital converting unit is set carries out the sample frequency f=51200Hz of digital-to-analog conversion＞2f _Max

3) the ore total amount that is provided with in the bowl mill 1 is 22kg, and steel ball is 40kg, and water is 10kg, and slip mean filter parameter N=1000 are by the vibration signal average intensity value β of 2 detections of vibration signal pick-up unit and calculating bowl mill barrel, β=5.98m/s ²At this moment, the acceleration a that produces because of gravity in the sampled value x of barrel vibration acceleration signal, the acceleration transducer _g, vibration acceleration signal sampled value x rejects in the acceleration transducer and produces acceleration a because of gravity _gAfter value

Respectively with the count graph of a relation of k of vibration acceleration signal, as Fig. 7, Fig. 8 and shown in Figure 9.

4) keep the steel ball and the water yield in the bowl mill 1 constant, increase the raw ore mine-supplying quantity, make the interior ore total amount of bowl mill increase to 50kg gradually by 22kg, be full of ore in the bowl mill this moment, is in the tripe malfunction that rises, and writes down ore total amount w in the bowl mill respectively _oWith the barrel vibration signal mean intensity β of correspondence, constitute sample data group { β (i), w _o(i) }, i=1 wherein, 2 ..., 9, amount to 9 groups of experiments, and, w _o(5)=and 30kg, the upper limit of ore total amount during for the bowl mill operate as normal, w _o(9)=and 50kg, for bowl mill is in the ore total amount in the bowl mill when rising by the tripe malfunction.

5) according to sample data group { β (i), w _o(i) }, determine that the ball mill belly expansion fault prediction model is

Barrel vibration signal average intensity value β and the interior ore total amount w of corresponding bowl mill _oRelation, as shown in figure 10.

When 6) bowl mill 1 is worked, barrel vibration signal pick-up unit 2 detection computations vibration signal average intensity value β, and send to supervisory control comuter 3, the tripe fault model that rises that supervisory control comuter 3 is determined according to step 5)

Promptly can calculate the estimated value of ore total amount in the grinding machine

If And continue to increase, near w _o(9)=and 50kg, forecast that then bowl mill 1 will the tripe fault that rises occur.

The various embodiments described above only are used to illustrate the present invention, and wherein the structure of each parts, connected mode etc. all can change to some extent, and every equivalents of carrying out on the basis of technical solution of the present invention and improvement all should not got rid of outside protection scope of the present invention.

Claims (9)

1. ball mill by wet process belly expansion fault prediction device, it is characterized in that: it comprises that one is arranged on the vibration signal pick-up unit with wireless communication function on bowl mill barrel surface, and one is arranged near the supervisory control comuter with wireless communication function the bowl mill;

Preset acceleration transducer, analog digital converting unit, central processing unit, data storage cell and wireless data transmission unit in the described vibration signal pick-up unit;

Preset wireless data receiving element, ball mill belly expansion fault model training unit and ball mill belly expansion fault prediction unit in the described supervisory control comuter;

Described central processing unit is controlled the vibration acceleration signal that described analog digital converting unit collects described acceleration transducer and is carried out the digital-to-analog conversion sampling, and be stored to described data storage cell, described central processing unit is to calculating through the vibration acceleration sampled value after the digital-to-analog conversion simultaneously, obtain the average intensity value of vibration signal, average intensity value is sent to the described wireless data receiving element that presets in the described supervisory control comuter by described wireless data transmission unit then;

Described wireless data receiving element sends the average intensity value of receiving to ball mill belly expansion fault model training unit, obtains training pattern

And the model that obtains is preset in the described ball mill belly expansion fault prediction unit, described wireless data receiving element directly sends the follow-up average intensity value of receiving to described ball mill belly expansion fault prediction unit, carries out the evaluation prediction of ball mill belly expansion fault.

2. a kind of ball mill by wet process belly expansion fault prediction device as claimed in claim 1, it is characterized in that: the low-limit frequency response of described acceleration transducer is less than the rotating speed of bowl mill, and the highest frequency response of described acceleration transducer is higher than the highest frequency of steel-ball collision barrel generation vibration acceleration signal in the bowl mill.

3. a kind of ball mill by wet process belly expansion fault prediction device as claimed in claim 1 or 2 is characterized in that: the sensitive axes of described acceleration transducer is vertical with the bowl mill barrel at its place, installation site.

4. ball mill by wet process belly expansion fault prediction method as device as described in the claim 1～3, it may further comprise the steps:

1) a vibration signal pick-up unit with wireless communication function is set on bowl mill barrel surface, a supervisory control comuter with wireless communication function is set near bowl mill;

2) analog digital converting unit that presets in the vibration signal pick-up unit is set, the sample frequency that the vibration acceleration signal that acceleration transducer is collected carries out the analog digital conversion, and the sensitive axes direction position identical with gravity direction of the acceleration transducer that presets in the vibration signal pick-up unit be designated as the A point, the position opposite with gravity direction is designated as the C point;

3) the interior ore total amount w of bowl mill is set _oThe steel ball amount and the water yield, bowl mill work, steel ball constantly collides barrel and produces vibration acceleration signal x in the cylindrical shell, the central processing unit that presets in vibration signal pick-up unit control analog digital converting unit degree of will speed up sensor acquisition to vibration acceleration signal carry out the digital-to-analog conversion sampling, and be stored to data storage cell, central processing unit is according to the vibration acceleration signal sampled value through digital-to-analog conversion of storing in the data storage cell, calculate the average intensity value β of the vibration acceleration signal that the bowl mill barrel ordered to A by the C point, and average intensity value β is sent to supervisory control comuter by wireless data transmission unit;

4) the wireless data receiving element in the supervisory control comuter receives the average intensity value β of the bowl mill barrel vibration acceleration signal that is sent by wireless data transmission unit in the vibration signal pick-up unit, and send ball mill belly expansion fault model training unit and ball mill belly expansion fault prediction unit to, be respectively applied for the foundation and the failure prediction of the tripe fault model that rises;

5) constantly increase the interior ore total amount of bowl mill, repeating step 3) and step 4), and in ball mill belly expansion fault model training unit under the different ore total amounts of record, the vibration signal average intensity value that corresponding bowl mill barrel is ordered to A by the C point, constitute sample data group { β (i), w _o(i) }, i=1 wherein, 2 ..., l is for increasing the experiment group of bowl mill raw ore mine-supplying quantity; K group experiment (during 1＜k＜l), w _oThe upper limit of ore total amount when (k) being the bowl mill operate as normal, when the experiment of l group, the ore total amount in the bowl mill is w _o(l), bowl mill is in the tripe malfunction that rises; When in the ball mill belly expansion fault prediction unit, noting experiment of k group and the experiment of l group respectively, the interior corresponding ore total amount w of bowl mill _o(k), w _o(l);

6) in ball mill belly expansion fault model training unit, preset ore total amount estimated value in the bowl mill

The quadratic polynomial model:

${\hat{w}}_o=p\left(\mathrm{\β}\right)=a_0+a_1\mathrm{\β}+a_2{\mathrm{\β}}^2---\left(1\right)$

Wherein,

Be ore total amount w in the bowl mill _oEstimated value, calculate by β; a ₀, a ₁, a ₂Be undetermined parameter, according to sample data group { β (i), w _o(i) }, try to achieve undetermined parameter a ₀, a ₁, a ₂, determine formula (1), ball mill belly expansion fault model training unit sends the formula of determining (1) to the ball mill belly expansion fault prediction unit, and stores, to be used for the prediction of ball mill belly expansion fault;

7) bowl mill man-hour, the vibration signal pick-up unit detects and calculates the average intensity value β of vibration acceleration signal in real time, and average intensity value β sent to supervisory control comuter by wireless data transmission unit, ball mill belly expansion fault prediction unit in the supervisory control comuter is brought average intensity value β in the formula of determining in the step 6) (1) into, calculates the estimated value of ore total amount in the bowl mill With this estimated value

The upper limit w of ore total amount during with the bowl mill operate as normal _o(k) compare, thereby whether the ore total amount of in time judging in the bowl mill is in normal range, if And continue to increase near w _o(l), then forecast ore total amount overrun in the bowl mill, will the tripe fault that rises occur.

5. a kind of ball mill by wet process belly expansion fault prediction method as claimed in claim 4, it is characterized in that: the concrete computing method of described step 3) are:

1. in central processing unit, put storage mark position SP=0, and slip mean filter parameter N is set;

2. detect the vibration acceleration signal of bowl mill barrel by acceleration transducer, and be vibration acceleration signal sampled value x (k), k=1 by the analog digital converting unit digital-to-analog conversion, 2,, k is counting of vibration acceleration signal, and is stored in the data storage cell;

3. central processing unit is according to the vibration acceleration signal sampled value x (k) that stores in the data storage cell, k=1, and 2 ..., calculate the acceleration a that acceleration transducer produces because of gravity by the slip mean filter _g(k), when k＜N,

$a_g\left(k\right)=\frac{1}{k}\underset{i=1}{\overset{k}{\mathrm{\Σ}}}x\left(i\right)$

When k 〉=N,

$a_g\left(k\right)=\frac{1}{N}\underset{i=k-N+1}{\overset{k}{\mathrm{\Σ}}}x\left(i\right)$

And obtain a respectively _g(k) maximal value a _GmaxWith minimum value a _Gmin

4. judge by central processing unit whether the storage mark position is SP=0, if SP=0 then judges a _g(k) whether be minimum value, if a _g(k) non-minimum value a _Gmin, then directly return step 2.; Work as a _g(k) be minimum value a _GminThe time, illustrate that acceleration transducer is in the C point, then is provided with storage mark position SP=1, and returns step 2.;

If 5. storage mark position SP=1 then judges a _g(k) whether be maximal value a _GmaxIf: a _g(k) non-maximal value a _Gmax, illustrate acceleration transducer at the C point between the A point, then calculate

Be vibration acceleration signal x (k), k=1,2 ... deduct the acceleration a that acceleration transducer produces because of gravity _g(k) value after, promptly

And will

Store in the data storage cell, return step 2.;

If 6. storage mark position SP=1, and a _g(k) be maximal value, illustrate that acceleration transducer is positioned at the A point, SP=1 is made as SP=0 with the storage mark position, calculate the bowl mill barrel by the C point to the vibration signal average intensity value β between the A point,

$\mathrm{\β}=\frac{1}{M}\underset{j=1}{\overset{M}{\mathrm{\Σ}}}\left|\stackrel{\‾}{x}\left(j\right)\right|$

Wherein, M is that the bowl mill barrel is C point counting to the vibration signal sampled value between the A point;

7. by wireless data transmission unit vibration signal average intensity value β is sent to supervisory control comuter, and the data in the storage unit that clears data, return step 2., infinite loop quits work until bowl mill.

6. a kind of ball mill by wet process belly expansion fault prediction method as claimed in claim 5 is characterized in that: described step 1. in, the span of slip mean filter parameter N is 500≤N≤1000.

7. as claim 4 or 5 or 6 described a kind of ball mill by wet process belly expansion fault prediction methods, it is characterized in that: described step 2), the sample frequency f that the vibration acceleration signal that the analog digital converting unit that presets in the vibration signal pick-up unit collects acceleration transducer carries out digital-to-analog conversion is set to f＞2f _Max, f _MaxProduce the highest frequency of vibration acceleration signal for steel-ball collision barrel in the bowl mill.

8. as claim 4 or 5 or 6 described a kind of ball mill by wet process belly expansion fault prediction methods, it is characterized in that: undetermined parameter a in the described formula (1) ₀, a ₁, a ₂Try to achieve, should make ore total amount w in the bowl mill _oWith ore total amount estimated value in the bowl mill

Difference Q minimum, that is:

$Q_{\min }=\underset{i=1}{\overset{l}{\mathrm{\Σ}}}{(w_o\left(i\right)-{\hat{w}}_o\left(i\right))}^2=\underset{i=1}{\overset{l}{\mathrm{\Σ}}}{(w_o\left(i\right)-p\left(\mathrm{\β}\left(i\right)\right))}^2---\left(2\right)$

9. a kind of ball mill by wet process belly expansion fault prediction method as claimed in claim 7 is characterized in that: undetermined parameter a in the described formula (1) ₀, a ₁, a ₂Try to achieve, should make ore total amount w in the bowl mill _oWith ore total amount estimated value in the bowl mill

Difference Q minimum, that is:

$Q_{\min }=\underset{i=1}{\overset{l}{\mathrm{\Σ}}}{(w_o\left(i\right)-{\hat{w}}_o\left(i\right))}^2=\underset{i=1}{\overset{l}{\mathrm{\Σ}}}{(w_o\left(i\right)-p\left(\mathrm{\β}\left(i\right)\right))}^2---\left(2\right)$

Images