CN110597053B

CN110597053B - Large mill process control system

Info

Publication number: CN110597053B
Application number: CN201910956040.6A
Authority: CN
Inventors: 樊毅; 管孝强; 王剑
Original assignee: Yunnan Diqing Nonferrous Metals Co ltd
Current assignee: Yunnan Diqing Nonferrous Metals Co ltd
Priority date: 2019-10-09
Filing date: 2019-10-09
Publication date: 2022-09-30
Anticipated expiration: 2039-10-09
Also published as: CN110597053A

Abstract

A large-scale mill process control system relates to mining machinery, in particular to a large-scale mill process control system which avoids distorted signals through discrimination processing and effectively avoids frequent shutdown caused by false alarm and faults. The process control system of the large-scale mill is arranged on the mill and is characterized by comprising an internal framework, an external auxiliary unit and a mill monitoring system, wherein the internal framework, the external auxiliary unit and the mill monitoring system are connected through leads. The invention discriminates and processes the detection signals of all external sensors, avoids the distorted signals from entering a logic linkage control unit and avoids the frequent shutdown of the mill due to false alarm and failure; the mechanical performance predictive control function is added, and the mechanical damage degree of a bearing, a bearing bush, a gear and the like is effectively controlled; the online operation fault processing function is added, so that the faults of gear lubricating oil lubrication, temperature detection and the like can be processed without stopping the machine, and the unnecessary shutdown frequency of the mill is reduced.

Description

Large mill process control system

Technical Field

The invention relates to mining machinery, in particular to a large-scale mill process control system which avoids distorted signals and effectively avoids frequent shutdown caused by false alarm and faults through discrimination processing.

Background

At present, the design concept of a lubricating control system of a copper mine grinding machine for processing more than ten thousand tons per day in China is old and traditional, and most of the lubricating control system is designed by applying signal detection and control under an ideal environment state. The problem that the severe environments such as electromagnetic interference, dust, oil stain, vibration and the like influence the frequent failure and parking is not solved; the mechanical ill-condition prediction of the mill can not be carried out, and once a fault occurs, the mechanical damage is enlarged and repaired, so that the problem of difficulty is solved; the method for processing faults in online operation can not be used for reducing the unplanned shutdown frequency of the mill; meanwhile, after the grinder stops, the problem of difficulty in starting the grinder is solved. The control system for the ultra-large key equipment has the following defects: 1. the functions of discriminating, judging and classifying whether signals of peripheral sensors are distorted or not and pushing control are realized; 2. no predictive control function of mechanical properties; 3. no on-line operation fault handling function, etc. According to incomplete statistics, the influence caused by the fault shutdown of the mill is as follows: 1. each time of fault shutdown, a company carries out parking for about 3 days and good post treatment, more than 10 times of false fault parking occur each year, no unplanned parking is included, the total of 30 days of single-series production is delayed in the whole year, the direct economic loss of the single-series parking in each day is about 200 ten thousand yuan, and the direct economic loss of the single-series parking in the whole year is about 30 x 200 to 6000 ten thousand yuan; 2. the annual running efficiency of the single-series equipment is reduced to about 30/365/2-4.1%; 3. a large amount of parking is good at aftertreatment, additionally increases the high-intensity physical labor of staff under the plateau environment, reduces the physical and mental health indexes of the staff, and increases the safety risk.

The structure of the traditional mill comprises a driving motor, a clutch, a pinion, a bull gear and a mill cylinder body, wherein the driving motor is connected with the pinion through the clutch, the bull gear is installed on the mill cylinder body, the pinion is meshed with the bull gear, bearings are arranged at a feeding end and a discharging end of the mill cylinder body, the driving motor drives the pinion to rotate through the clutch, the pinion drives the bull gear to rotate, the bull gear drives the mill cylinder body to rotate around the bearings, two sets of driving motors, the clutch and the pinion are usually adopted, after connection, two sides of the bull gear are respectively meshed with the two pinions, the mill cylinder body is driven to rotate through the two sets of driving motors, and the rotating stability is effectively improved; and the two ends of the pinion and the two bearings of the mill cylinder are connected with an oil supply pipeline and an oil nozzle to replenish lubricating oil in real time.

Disclosure of Invention

The invention aims to solve the problem that the existing copper mine grinder lubrication control system is unscientific in design, is used in a severe environment and is easily interfered by a distortion signal to cause frequent stoppage, and provides a large grinder process control system which avoids the distortion signal through discrimination processing and effectively avoids frequent stoppage caused by false alarm and faults.

The invention discloses a large-scale mill process control system which is arranged on a mill and is characterized in that the control system comprises an internal framework, an external auxiliary unit and a mill monitoring system, wherein the internal framework, the external auxiliary unit and the mill monitoring system are connected through leads, and the process control system comprises:

the internal framework comprises a logic linkage control unit, an operation state monitoring unit, a discrimination control processing unit, a prediction control processing unit, a data history storage unit and an online fault processing unit, wherein the discrimination control processing unit is respectively connected with the logic linkage control unit and the data history storage unit;

the external auxiliary unit comprises a remote transmission monitoring unit and an audible and visual alarm unit;

the mill monitoring system comprises a temperature sensor, an oil pressure sensor, a flow sensor and a vibration sensor, a plurality of temperature measuring points, oil pressure monitoring points and oil quantity monitoring points are arranged on the mill, three temperature measuring points are respectively arranged on bearings at a feeding end and a discharging end, three temperature measuring points are arranged on the surface of a pinion, two temperature measuring points are respectively arranged at two ends of the pinion, a temperature sensor is arranged on each temperature measuring point, the temperature of the temperature measuring points is collected through the temperature sensor, and the temperature sensor of the temperature measuring points on the surface of the pinion adopts an infrared temperature sensor; flow sensors are arranged on oil nozzles of bearings at two ends of the pinion and two ends of the two cylinders, and oil pressure sensors are arranged on oil supply pipelines; x, Y, Z vibration detection in three directions is established at the driving end, the non-driving end and the vertical direction of the pinion, each pinion is provided with 6 vibration sensors, and two pinions of one mill are provided with 12 vibration sensors; the mill monitoring system is connected with the screening control processing unit, and the obtained monitoring data are transmitted to the screening control processing unit;

the specific control process is as follows:

1) the discrimination control processing unit acquires pinion infrared temperature difference, pinion infrared temperature, oil flow, oil pressure, bearing temperature and vibration range, judges distortion conditions through fuzzy control, judges distortion signals, does not participate in logic linkage control, and enters the historical data recording unit for storage and only for query; if the signal is judged to be a true signal, the data is transmitted to a logic linkage control unit to participate in operation linkage control;

2) after the logic linkage control unit receives the data signal of the discrimination control processing unit, the signal is analyzed, the fault problem is judged, and the optimal use value is searched by utilizing fuzzy control; firstly, setting a wire-breaking temperature of-200 ℃ for each temperature sensor, directly transmitting information to an online fault processing unit when detecting the temperature, and sending an alarm signal; setting the normal use temperature value as 60 ℃, obtaining the ultrahigh temperature as 90 ℃ by utilizing fuzzy control calculation, and sending a fault shutdown instruction of the detection point and sending alarm information when the ultrahigh temperature value is exceeded; setting the normal range of the temperature difference value at the two ends to be 15 and the ultra-large temperature difference value to be 30 according to the information of the infrared temperature sensor in the temperature difference range of the small end of the pinion, sending an alarm signal when the temperature difference value exceeds 15, and sending a fault shutdown instruction when the temperature difference value exceeds 30; setting preset values of low pressure of 40kg/cm3 and ultra-low pressure of 15kg/cm3 in a normal use range for the oil pressure of an oil supply pipeline, and outputting alarm information or a fault shutdown instruction of a corresponding detection point when a signal reaches the preset values;

3) the online fault processing unit mainly aims at two types of fault processing, namely lubricating oil fault processing and temperature detection fault processing, and the specific processing mode is as follows:

(1) the method comprises the following steps of carrying out non-stop on-line processing on lubricating oil faults, wherein the non-stop on-line processing comprises two functions of lubricating oil fault processing waiting intermittence and timing zero setting, presetting waiting intermittence time after a lubricating oil injection period is finished, adding the timing zero setting function into the waiting intermittence time function, carrying out fault processing on a lubricating oil system in the waiting intermittence time, starting the timing zero setting function in the waiting intermittence time, restarting the timing of the waiting intermittence time, ensuring that the fault processing has sufficient time, and supplementing lubricating oil in a manual mode in time in the processing process by manually observing the condition of the lubricating oil until the faults are eliminated;

(2) the method comprises the following steps of (1) temperature detection fault processing, namely monitoring a circuit and a probe of each temperature sensor, displaying broken line information when any one temperature sensor sends out a broken line fault, and prompting maintenance personnel to perform targeted inspection and processing to avoid parking caused by faults generated due to high temperature;

4) the prediction control processing unit is used for setting an exceeding area threshold value for normal operation according to the real-time detection data and the on-line trend record of the historical data and the normal use range index value range of the equipment performance, and when the detection data exceeds the area threshold value, early warning is carried out in advance to inform maintenance personnel to carry out performance analysis and maintenance on mechanical structures such as a bearing, a bearing bush, a gear and the like of the equipment in advance;

5) the acousto-optic alarm unit is matched with an internal framework to carry out early warning on operation and maintenance personnel in time, so that the operation and maintenance personnel can take measures conveniently;

6) the remote transmission monitoring unit transmits the relevant information of the operation of the mill to the distributed control system, and the server of the distributed control system performs the functions of real-time and historical storage, trend query, state information display and control, so that a central control operator can conveniently and timely master the information of one hand of the operation of the mill and accurately issue corresponding instructions.

In the control step 1), the fuzzy control distortion condition judging step is as follows:

1) determining the structure of the fuzzy controller and the input and output linguistic variables, and adopting a two-dimensional fuzzy controller with double input and single output:

inputting language variables: the deviation linguistic variable is E, and the deviation change rate EC;

the deviation e is r-y, r is a distortion set value, and y is a distortion test value;

a deviation change rate ec (k) e (k) -e (k-1), (k) 0,1,2, … …);

outputting language variables: controlling the variable quantity U of the filtering time length in the distortion degree testing process;

2) determining linguistic values of input linguistic variable deviation E, deviation change EC and output linguistic variable control quantity change U:

basic domain of deviation e: [ -20, +20 ];

discrete domain of bias linguistic value E: x { -6., 0 …, +6 };

the quantization factor for the deviation e is: k _e ＝2*6/[20-(-20)]＝0.3；

The linguistic value of the linguistic variable deviation E is selected as: PB, PM, PS, PO, NO, NS, NM, NB 8 linguistic values:

determining membership functions mu (X) describing fuzzy subsets PB, …, NB in the domain of discourse X, and establishing an assignment table of linguistic variables E based thereon

Assignment table for linguistic variable E

Basic domains of variation rate of deviation EC: [ -15, +15 ];

discrete domain of deviation rate of change linguistic value EC: y { -6, -5, ·,0 …, +5, +6 };

the quantization factor for the rate of change of deviation ec is then: k _ec ＝2*6/[15-(-15)]＝0.4；

The linguistic values of the linguistic variable deviation EC are selected as: PB, PM, PS, ZO, NS, NM, NB7 language values;

determining membership functions mu (y) describing the fuzzy subsets PB, …, NB at the domain X, and establishing an assignment table of linguistic variables EC based thereon

Assignment table of language variable EC

Basic domain of control quantity change u: [ -10s, +10s ];

discrete domain of control quantity change linguistic value U: z { -7, -6, ·,0, …, +6, +7 };

the scaling factor of the control quantity change u is then: k _u ＝[10-(-10)]/2*7＝10/7；

The linguistic value of the linguistic variable control variable change U is selected as follows: PB, PM, PS, ZO, NS, NM, NB7 language values;

by means of a long-term summary, the membership functions μ (Z) describing the fuzzy subsets PB, …, NB in the domain of discourse Z are determined, and the assignment table of the linguistic variables U is established based thereon

Assignment table of linguistic variable U

3) Determination of fuzzy control rules:

when the error, namely the test value subtracted from the desired value, is negative and large, the test value is higher than the desired value, when the error change rate is also negative, the error has an increasing trend, and in order to eliminate the existing negative and large error and inhibit the increase of the existing negative and large error as soon as possible, the change of the control quantity is positive and large, namely the control quantity is increased, which means that the filtering time control setting is increased, so that the distortion degree is reduced; when the error is negative and the error change rate is positive, the system has the tendency of reducing the error, and a smaller control quantity is taken to eliminate the error as soon as possible without overshoot; when the error is negative or medium, the change of the control quantity can eliminate the error as soon as possible; based on the principle, the change of the control quantity is the same when the same error is negative; when the error is negative, the system approaches a steady state; when the error change is tiny, the control quantity is selected to be changed to be positive so as to restrain the error from changing to the negative direction; when the error changes to be positive, the system has the tendency of eliminating negative small errors, and the selection control quantity changes to be positive small; obtain a fuzzy control table

Fuzzy control rule table

Obtaining a fuzzy relation R between the distortion degree deviation, the deviation change rate and the filtering time length:

R＝(E×U)○(EC×U)＝E×EC×U＝(NB _E ×NB _EC ×PB _U )∪(NM _E ×NB _EC ×PB _U )∪(NS _E ×NB _EC ×PM _U )∪……

wherein the fuzzy sets in the rules take intersection operation, and the fuzzy sets in the rules take union operation;

4) fuzzification and fuzzy reasoning:

let measured deviations be e (k), ec (k) e (k) -e (k-1), (k) 0,1,2, … …)

Quantized to obtain e ^* ，ec ^*

And e ^* ∈X＝{-6,-5,...,-0,+0…,+5+6},ec ^* E.y { -6, -5.., 0, …, +5+6} will e ^* 、ec ^* Fuzzification to obtain fuzzy subset E ^* 、EC ^*

Then

Namely, the output of the fuzzy controller is the synthesis of an error vector, an error change rate vector and a fuzzy relation;

5) defuzzification of the control quantity:

and (3) generating a deblurring and fuzzy control table, wherein the maximum membership method is adopted to deblur fuzzy control increments to obtain control increments u which take values in a universe Z { -7 ^* The above-mentioned fuzzification, fuzzy inference and fuzzy decision operation are implemented on all the combinations of all the elements of the domain X (X { -6.,..,. 0, +0, …, + 6) and domain Y (Y { -6.,. 0, …, +6}), so as to obtain the one-to-one correspondence between the elements of the domain Z (Z { -7.,. 0, …, +7}), and these correspondences are made into a table, so that the fuzzy control table can be obtained

Fuzzy control look-up table

The fuzzy control table is calculated off-line according to

At E ^* ＝(PB,PM，PS,PO，NO,NS,NM，NB),EC ^* When the output fuzzy vector U of the controller is (PB, PM, PS, ZO, NS, NM, NB), the fuzzy vector U of the output of the controller is calculated ^* Then performing defuzzification according to a maximum membership method, and selecting the control quantity as U ^* The maximum value calculated on (PB, PM, PS, ZO, NS, NM, NB) is U ^* _max Filling the values into the fuzzy control lookup table;

in the case of on-line control, the deviation will be measured in each control period

e(k)、ec(k),(k＝0,1,2,……)

Respectively quantizing to obtain e required by looking up fuzzy control table ^* 、ec ^* Then, the corresponding control amount change u is looked up from the table ^* With a scale factor k _u The change value u of the actual control quantity for the controlled object is obtained by multiplying,

i.e. u-u ^* *k _u ；

The purpose of the fuzzy control is to remove the distortion degree, and all u with e less than or equal to 0 are included ^* In addition to k _u 10/7; the setting of the filtering time is not less than 5s when being combined with the trend oscillogram of the actual operation historical data, and u is equal to u ^* *k _u Calculating for more than or equal to 5s to obtain u ^* Not less than 3.5, so, under the fuzzy controller, all u ^* The control quantity more than 3 can meet the requirement of removing the distortion degree; but u ^* If the selection is too large, the variation value u of the actual control quantity is also too large, namely the filtering time is increased, so that the control system has serious lag of normal signals entering the logic control unit and other units, the system cannot process external normal signals timely, serious overshoot is caused, the superiority of the system is lost, and the control quantity u with the optimal distortion degree ^* It should satisfy: u is more than 3 ^* 4, so the optimal filtering time length for controlling the system signal processing unit to remove the signal distortion is set as 5 s.

In the control step 2), the specific steps of searching the optimal use value set value by using fuzzy control are as follows:

the deviation e is r-y, r is the target value of each set value, and y is the test value of multiple times;

a variation rate ec (k) e (k) -e (k-1), (k) 0,1,2, … …);

outputting language variables: setting the variable quantity U of the length of a control set value in the target stable operation process;

2) determining linguistic values of input linguistic variable deviation E, deviation change EC and output linguistic variable control quantity change U; forming a two-dimensional coordinate system, and performing fuzzy calculation by adopting a triangular or semi-trapezoidal membership function relationship;

3) determination of fuzzy control rules:

when the error is large, the control quantity is selected to eliminate the error as soon as possible; when the error is small, the control amount is selected to prevent overshoot, and according to the above rule, the following fuzzy control table is obtained

Fuzzy control rule table

And according to the control rule, obtaining a fuzzy relation R between the distortion degree deviation, the deviation change rate and the filtering time length:

4) fuzzification and fuzzy reasoning:

let measured deviations be e (k), ec (k) e (k) -e (k-1), (k) 0,1,2, … …)

Quantized to obtain e ^* 、ec ^*

And e ^* ∈X＝{-6,-5,...,-0,+0…,+5+6}

ec ^* ∈Y＝{-6,-5,...,0,…,+5+6}

E is to be ^* And ec ^* Fuzzification to obtain fuzzy subset E ^* And EC ^*

Then

5) defuzzification of the control quantity:

the fuzzy control increment is deblurred by adopting a maximum membership method to obtain a control increment u which takes a value on a discourse domain Z { -7 ^* (ii) a Performing the above-mentioned fuzzification, fuzzy inference and fuzzy decision operation on all combinations of all elements of a domain X (X { -6.,. 0, +0, …, + 6) and a domain Y (Y { -6.,. 0, …, +6}) to obtain one-to-one correspondence between elements in a domain Z (Z { -7.,. 0, …, +7}), and making the correspondence into a fuzzy control lookup table:

fuzzy control look-up table

The fuzzy control lookup table is obtained by off-line calculation:

according to

At E ^* ＝(PB,PM，PS,PO，NO,NS,NM，NB),EC ^* When the output fuzzy vector U of the controller is (PB, PM, PS, ZO, NS, NM, NB), the output fuzzy vector U of the controller is calculated ^* Then performing defuzzification according to a maximum membership method, and selecting the control quantity as U ^* The maximum value calculated on (PB, PM, PS, ZO, NS, NM, NB) is U ^* _max Filling the values into the fuzzy control lookup table;

in the case of on-line control, the measured deviations e (k), ec (k), and (k ═ 0,1,2, and … …) are quantized in each control cycle, and e required for searching the fuzzy control table is obtained ^* 、ec ^* Then, the corresponding control variable change u is looked up from the table ^* Is in proportion toSub k _u Multiplying the control values to obtain a change value u of the actual control quantity for the controlled object, i.e. u-u ^* *k _u ；

6) And obtaining the following set value control setting conclusion according to the fuzzy control design:

(1) searching the optimal set value of infrared temperature difference alarm:

selecting: basic domain of the control variable change u of the alarm value: [ -15, +15 ];

the scale factor of the control quantity change u is: k _u ＝[10-(-10)]/2*7＝15/7；

Deriving u-u from above ^* *k _u Calculated to obtain u ^* 7. The maximum membership value principle is satisfied, so the optimal setting of the alarm value is 15 ℃;

(2) searching the optimal set value of the infrared temperature difference fault shutdown:

selecting: basic domain of the control variable change u of the alarm value: [ -30, +30 ];

the proportional factor of the control amount change u is: k _u ＝[10-(-10)]/2*7＝30/7；

Deriving u-u from above ^* *k _u Is calculated to obtain u ^* And 7, the maximum membership value principle is satisfied, so the optimal setting of the fault shutdown value is 30 ℃.

In the specific control process, a lubricating oil constant temperature control process is also arranged, stable and efficient operation of the mill is required to be kept according to field use conditions, a lubricating control system is required to stably operate at constant temperature, and the temperature of a mill lubricating system is stabilized at 28 ℃ and is in an optimal state through long-time experience summary; the control process utilizes fuzzy calculation to control, and the specific control steps are as follows:

1) determining the structure of a fuzzy controller and input and output linguistic variables, and adopting a two-dimensional fuzzy controller with double input and single output;

the temperature deviation e is r-y, r is a set value of the temperature of the lubricating system, and y is a detected value of the temperature of the lubricating system;

a temperature deviation change rate ec (k) e (k) -e (k-1), (k) 0,1,2, … …);

outputting language variables: controlling the variable quantity U of the action duration of the heater in the constant temperature maintaining process;

basic domain of deviation e: [ -10 ℃, +10 ℃ ];

discrete domain of bias linguistic value E: x { -6., 0 …, +6 };

the quantization factor for the deviation e is: k _e ＝2*6/[10-(-10)]＝0.6；

The linguistic value of the linguistic variable deviation E is selected as: PB, PM, PS, PO, NO, NS, NM and NB 8 language values;

determining the membership function mu (X) of the fuzzy subset PB, …, NB on the domain X as a triangular membership function, and establishing an assignment table of the linguistic variable E based thereon

Assignment table for linguistic variable E

Basic domain of deviation rate of change EC: [ -15, +15 ];

the quantization factor for the rate of change of deviation ec is then: k _ec ＝2*6/[15-(-15)]＝6/15；

Linguistic values for the linguistic variable deviation EC are selected as: PB, PM, PS, ZO, NS, NM and NB7 language values;

determining the membership function mu (y) of the fuzzy subset PB, …, NB on the domain X as a triangular membership function, and establishing an assignment table of the linguistic variables EC according to the function

Assignment table of language variable EC

Basic domain of control quantity change u: [ -60min, +60min ];

discrete domain of control quantity change linguistic value U: z { -7, -6, · 0, …, +6, +7 };

the scaling factor of the control quantity change u is then: k is _u ＝[60-(-60)]/2*7＝60/7；

The linguistic value of the linguistic variable control variable change U is selected as follows: PB, PM, PS, ZO, NS, NM and NB7 language values;

determining the membership function mu (Z) used for describing the fuzzy subsets PB, …, NB on the domain Z as a triangular membership function, and establishing an assignment table of the linguistic variable U according to the function

Assignment table of linguistic variable U

3) Determination of fuzzy control rules:

when the error, namely the desired value minus the detection value, is negative and large, the detection value is higher than the desired value, and when the error change rate is also negative, the error has an increasing trend, so that the existing negative and large error is eliminated and the increase of the error is restrained as soon as possible, and therefore, the change of the control quantity is positive and large; when the error is negative and the error change rate is positive, the system has the tendency of reducing the error, so that a smaller control quantity is taken to eliminate the error as soon as possible without overshoot; when the error is negative, the change of the control quantity can eliminate the error as soon as possible, and based on the principle, the change of the control quantity is the same when the error is negative; when the error is negative, the system is close to a steady state, and if the error changes slightly, the control quantity is selected to be changed to be positive so as to inhibit the error from changing to the negative direction; when the error changes to be positive, the system has the tendency of eliminating negative small errors, and the selection control quantity changes to be positive small;

obtaining a fuzzy control rule table according to the rules

Fuzzy control rule table

And according to the control rule, obtaining a fuzzy relation R between the temperature deviation, the deviation change rate and the working time length of the heater:

4) fuzzification and fuzzy reasoning:

let the measured deviations be e (k), ec (k) ═ e (k) — e (k-1), (k) ═ 0,1,2, … …)

Quantized to obtain e ^* And ec ^* And is and

e ^* ∈X＝{-6,-5,...,-0,+0…,+5+6},

ec ^* ∈Y＝{-6,-5,...,0,…,+5+6}

e is to be ^* And ec ^* Fuzzification to obtain fuzzy subset E ^* And EC ^*

Then

5) defuzzification of the control quantity:

and (3) deblurring the fuzzy control increment by adopting a maximum membership method to obtain a control increment u which takes a value on a universe Z { -7 ^* ；

All combinations of all elements of the discourse field X (X { -6.,. 0, +0, …, + 6) and the discourse field Y (Y { -6.,. 0, …, + 6) are subjected to the fuzzification, fuzzy inference and fuzzy decision operation, so that elements on the discourse field Z (Z { -7.,. 0, …, + 7) can be obtained to be in one-to-one correspondence with the elements, and the correspondence relations are made into a table, so that the fuzzy control lookup table is obtained

Fuzzy control look-up table

And (3) fuzzy control lookup table offline calculation:

according to

in the case of on-line control, the measured deviations e (k), ec (k), and (k) 0,1,2, and … … are quantized individually for each control cycle, and e required for looking up the fuzzy control table is obtained ^* 、ec ^* Then, the corresponding control amount change u is looked up from the table ^* With a scale factor k _u Multiplying the control values to obtain a change value u of the actual control quantity for the controlled object, i.e. u-u ^* *k _u ；

And (3) final control process: the fuzzy control variable quantity is obtained by combining the temperature deviation detected in real time and the deviation change rate, and the required actual control quantity is output through a series of conversion of anti-fuzzy control.

The large-scale mill process control system has the advantages of simple structure, scientific design and convenient use, discriminates and processes detection signals of all external sensors, prevents distorted signals from entering a logic linkage control unit, and prevents mills from being frequently stopped due to false alarm and faults; the mechanical performance predictive control function is added, and the mechanical damage degrees of bearings, bearing bushes, gears and the like are effectively controlled; the online operation fault processing function is added, so that the faults of lubrication of gear lubricating oil, temperature detection and the like can be processed without stopping the machine, and the unnecessary shutdown frequency of the mill is reduced; after the technology is developed and applied, the following effects are obtained: 1. the direct economic benefit of annual production of the company is increased by about 6000 ten thousand yuan; 2. the annual running efficiency of single-series equipment is improved by about 4.1 percent; 3. in a plateau environment, the high-intensity extra physical labor of the staff is reduced, the physical and mental health index of the staff is improved, and the safety risk is reduced.

Drawings

FIG. 1 is a schematic diagram of a control system according to the present invention.

FIG. 2 is a schematic view of the structure of the monitoring system of the mill of the present invention.

The system comprises a logic interlocking control unit 1, an operating state monitoring unit 2, a discrimination control processing unit 3, a prediction control processing unit 4, a data history storage unit 5, an online fault processing unit 6, a remote transmission monitoring unit 7, an audible and visual alarm unit 8, a temperature sensor 9, an oil pressure sensor 10, a flow sensor 11, a bearing 12, a pinion 13 and an oil supply pipeline 14.

Detailed Description

Example 1: a large mill process control system installed on a mill, the control system comprising an internal architecture, an external auxiliary unit and a mill monitoring system, the internal architecture, the external auxiliary unit and the mill monitoring system being connected by wires, wherein:

the internal architecture comprises a logic interlocking control unit 1, an operating state monitoring unit 2, a screening control processing unit 3, a prediction control processing unit 4, a data history storage unit 5 and an online fault processing unit 6, wherein the screening control processing unit 3 is respectively connected with the logic interlocking control unit 1 and the data history storage unit 5, the prediction control processing unit 4 is respectively connected with the logic interlocking control unit 1 and the data history storage unit 5, the logic interlocking control unit 1 and the data history storage unit 5 are respectively connected with the operating state monitoring unit 2, and the operating state monitoring unit 2 is connected with the online fault processing unit 6;

the external auxiliary unit comprises a remote transmission monitoring unit 7 and an audible and visual alarm unit 8;

the mill monitoring system comprises a temperature sensor 9, an oil pressure sensor 10, a flow sensor 11 and a vibration sensor, a plurality of temperature measuring points, oil pressure monitoring points and oil quantity monitoring points are arranged on the mill, three temperature measuring points are respectively arranged on bearings 12 at a feeding end and a discharging end, three temperature measuring points are arranged on the surface of a pinion 13, two temperature measuring points are respectively arranged at two ends of the pinion 13, the temperature sensor 9 is arranged on each temperature measuring point, the temperature of the temperature measuring points is collected through the temperature sensor 9, and the temperature sensor 9 of the temperature measuring points on the surface of the pinion 13 adopts an infrared temperature sensor 9; flow sensors 11 are arranged on oil nozzles of bearings 12 at two ends of a pinion 13 and two ends of two cylinders, and an oil pressure sensor 10 is arranged on an oil supply pipeline 14; x, Y, Z vibration detection in three directions is established at the driving end, the non-driving end and the vertical direction of the pinion 13, each pinion 13 is provided with 6 vibration sensors, and two pinions 13 of one mill are provided with 12 vibration sensors; the mill monitoring system is connected with the screening control processing unit 3, and the obtained monitoring data are transmitted to the screening control processing unit 3;

the specific control process is as follows:

1) the discrimination control processing unit 3 acquires the infrared temperature difference of the pinion 13, the infrared temperature of the pinion 13, oil flow, oil pressure, the temperature of the bearing 12 and the vibration range, judges the distortion condition through fuzzy control, judges the distortion condition as a distortion signal, does not participate in logic linkage control, and enters a historical data recording unit for storage and is only used for inquiry; if the signal is judged to be a true signal, the data is transmitted to a logic linkage control unit 1 to participate in operation linkage control;

2) after receiving the data signal of the discrimination control processing unit 3, the logic interlocking control unit 1 analyzes the signal, judges the fault problem and searches the optimal use value set value by utilizing fuzzy control; firstly, each temperature sensor 9 sets the wire-breaking temperature of-200 ℃, directly transmits information to the online fault processing unit 6 when detecting the temperature, and sends out an alarm signal; setting the normal use temperature value as 60 ℃, obtaining the ultrahigh temperature as 90 ℃ by utilizing fuzzy control calculation, and sending a fault shutdown instruction of the detection point and sending alarm information when the ultrahigh temperature value is exceeded; setting the normal range of the temperature difference value at the two ends to be 15 and the ultra-large temperature difference value to be 30 according to the information of the infrared temperature sensor 9 in the temperature difference range at the small end of the pinion 13, sending an alarm signal when the temperature difference value exceeds 15, and sending a fault shutdown instruction when the temperature difference value exceeds 30; setting preset values of low pressure of 40kg/cm3 and ultra-low pressure of 15kg/cm3 in a normal use range for the oil pressure of an oil supply pipeline 14, and outputting alarm information or a fault shutdown instruction of a corresponding detection point when a signal reaches the preset values;

3) the online fault processing unit 6 is mainly used for processing two types of faults, namely lubricating oil fault processing and temperature detection fault processing, and the specific processing mode is as follows:

(2) the temperature detection fault processing is to monitor the line and the probe of each temperature sensor 9, when any one temperature sensor 9 sends out a disconnection fault, the disconnection information is displayed, and the maintainers are prompted to carry out targeted inspection and processing, so that the shutdown caused by the fault generated by high temperature is avoided;

4) the prediction control processing unit 4 sets an exceeding area limit value for normal operation according to the real-time detection data and the on-line trend record of the historical data and by combining the normal use range index value range of the equipment performance, and warns in advance when the detection data exceeds the area limit value, and informs maintenance personnel to perform performance analysis and maintenance on mechanical structures such as a bearing 12, a bearing bush, a gear and the like of the equipment in advance;

5) the acousto-optic alarm unit 8 is matched with an internal framework to carry out early warning on operation and maintenance personnel in time, so that the operation and maintenance personnel can take measures conveniently;

6) and the remote transmission monitoring unit 7 transmits the relevant information of the operation of the mill to the distributed control system, and performs real-time and historical storage, trend query, state information display and control functions on a server of the distributed control system, so that a central control operator can conveniently and timely master information of one hand of the operation of the mill and accurately issue corresponding instructions.

a deviation change rate ec (k) e (k) -e (k-1), (k) 0,1,2, … …);

basic domain of deviation e: [ -20, +20 ];

discrete domain of deviation linguistic value E: x { -6., 0 …, +6 };

the quantization factor for the deviation e is: k _e ＝2*6/[20-(-20)]＝0.3；

Linguistic values for the linguistic variable deviation E are selected as: PB, PM, PS, PO, NO, NS, NM, NB 8 linguistic values:

Assignment table for linguistic variable E

Basic domain of deviation rate of change EC: [ -15, +15 ];

determining membership functions mu (y) for describing fuzzy subsets PB, …, NB in the domain X, and establishing an assignment table of linguistic variables EC based thereon

Assignment table of language variable EC

Basic domain of control quantity change u: [ -10s, +10s ];

the scaling factor of the control quantity change u is then: k is _u ＝[10-(-10)]/2*7＝10/7；

Assignment table of linguistic variable U

3) Determination of fuzzy control rules:

when the error, namely the test value subtracted from the desired value, is negative and large, the test value is higher than the desired value, when the error change rate is also negative, the error has an increasing trend, and in order to eliminate the existing negative and large error and inhibit the increase of the existing negative and large error as soon as possible, the change of the control quantity is positive and large, namely the control quantity is increased, which means that the filtering time control setting is increased, so that the distortion degree is reduced; when the error is negative and the error change rate is positive, the system has the tendency of reducing the error, and a smaller control quantity is taken to eliminate the error as soon as possible without overshoot; when the error is negative or medium, the change of the control quantity can eliminate the error as soon as possible; based on the principle, the change of the control quantity is the same when the same error is negative; when the error is negative, the system approaches a steady state; when the error change is tiny, the control quantity is selected to be changed to be positive so as to restrain the error from changing to the negative direction; when the error changes to be positive, the system has the tendency of eliminating negative small errors, and the selection control quantity changes to be positive small; obtaining a fuzzy control table

Fuzzy control rule table

4) fuzzification and fuzzy reasoning:

Quantized to obtain e ^* ，ec ^*

And e ^* ∈X＝{-6,-5,...,-0,+0…,+5+6},ec ^* ∈Y＝{-6,-5,...,0,…,+5+6}

E is to be ^* 、ec ^* Fuzzification to obtain fuzzy subset E ^* 、EC ^*

Then

Namely, the output of the fuzzy controller is the synthesis of the error vector, the error change rate vector and the fuzzy relation;

5) defuzzification of the control quantity:

and (3) generating a deblurring and fuzzy control table, wherein the maximum membership method is adopted to deblur fuzzy control increments to obtain control increments u which take values in a universe Z { -7 ^* The above-described blurring, fuzzy inference and fuzzy inference are performed on all combinations of elements of the domain X (X { -6.,. 0, +0, …, +6}) and the domain Y (Y { -6.,. 0, …, +6})Fuzzy decision operation can obtain the one-to-one correspondence between the elements in the universe of discourse Z (Z { -7., 0, …, +7}), and the correspondence is made into a table to obtain the fuzzy control table

Fuzzy control look-up table

The fuzzy control table is calculated off-line according to

e(k)、ec(k),(k＝0,1,2,……)

Respectively quantizing to obtain e required by looking up fuzzy control table ^* 、ec ^* Then, the corresponding control amount change u is looked up from the table ^* And a scale factor k _u The change value u of the actual control quantity for the controlled object is obtained by multiplying,

i.e. u-u ^* *k _u ；

The purpose of the fuzzy control is to remove the distortion degree, and all u with e less than or equal to 0 are included ^* In addition k _u 10/7; the setting of the filtering time is not less than 5s when being combined with the trend oscillogram of the actual operation historical data, and u is equal to u ^* *k _u Calculating for more than or equal to 5s to obtain u ^* Not less than 3.5, so, under the fuzzy controller, all u ^* The control quantity more than 3 can meet the requirement of removing the distortion degree; but u ^* If the selection is too large, the change value u of the actual control quantity is also too large, namely the filtering time is increased, so that the control system enters a logic stateThe normal signals of the units such as the edit control unit are seriously lagged, the system processes the external normal signals untimely, the overshoot is serious, the superiority of the system is lost, and therefore the control quantity u of the optimal distortion degree ^* It should satisfy: u is more than 3 ^* 4, so the optimal filtering time length for controlling the system signal processing unit to remove the signal distortion is set as 5 s.

In the control step 2), the specific steps of finding the optimal use value setting value by using fuzzy control are as follows:

inputting linguistic variables: the deviation linguistic variable is E, and the deviation change rate EC;

a deviation change rate ec (k) e (k) -e (k-1), (k) 0,1,2, … …);

3) determination of fuzzy control rules:

Fuzzy control rule table

4) fuzzification and fuzzy reasoning:

Quantized to obtain e ^* 、ec ^*

And e ^* ∈X＝{-6,-5,...,-0,+0…,+5+6}

ec ^* ∈Y＝{-6,-5,...,0,…,+5+6}

E is to be ^* And ec ^* Fuzzification to obtain fuzzy subset E ^* And EC ^*

Then the

5) defuzzification of the control quantity:

and (3) deblurring the fuzzy control increment by adopting a maximum membership method to obtain a control increment u which takes a value on a universe Z { -7 ^* (ii) a Performing the above-mentioned fuzzification, fuzzy inference and fuzzy decision operation on all combinations of all elements of a domain X (X { -6.,. 0, +0, …, + 6) and a domain Y (Y { -6.,. 0, …, +6}) to obtain one-to-one correspondence between elements in a domain Z (Z { -7.,. 0, …, +7}), and making the correspondence into a fuzzy control lookup table:

fuzzy control look-up table

The fuzzy control lookup table is obtained by off-line calculation:

according to

At E ^* ＝(PB,PM，PS,PO，NO,NS,NM，NB),EC ^* When the value is (PB, PM, PS, ZO, NS, NM, NB),calculating an output fuzzy vector U of a controller ^* Then performing defuzzification according to a maximum membership method, and selecting the control quantity as U ^* The maximum value calculated on (PB, PM, PS, ZO, NS, NM, NB) is U ^* _max Filling the values into the fuzzy control lookup table;

in the case of on-line control, the measured deviations e (k), ec (k), and (k ═ 0,1,2, and … …) are quantized in each control cycle, and e required for searching the fuzzy control table is obtained ^* 、ec ^* Then, the corresponding control amount change u is looked up from the table ^* With a scale factor k _u Multiplying the control values to obtain a change value u of the actual control quantity for the controlled object, i.e. u-u ^* *k _u ；

(1) searching the optimal set value of infrared temperature difference alarm:

the scaling factor of the control quantity change u is then: k _u ＝[10-(-10)]/2*7＝30/7；

In the specific control process, a lubricating oil constant temperature control process is also arranged, the stable and efficient operation of the mill is required to be kept according to the field use condition, the stable and constant temperature operation of a lubricating control system is required, and the temperature of the lubricating system of the mill is stable at 28 ℃ and is in an optimal state through long-time experience summary; the control process utilizes fuzzy calculation to control, and the specific control steps are as follows:

a temperature deviation change rate ec (k) e (k) -e (k-1), (k) 0,1,2, … …);

outputting language variables: controlling the variation U of the action time of the heater in the constant temperature keeping process;

basic domain of deviation e: [ -10 ℃ and +10 ℃ C ];

discrete domain of bias linguistic value E: x { -6., 0 …, +6 };

the quantization factor for the deviation e is: k _e ＝2*6/[10-(-10)]＝0.6；

Assignment table for linguistic variable E

Basic domain of deviation rate of change EC: [ -15, +15 ];

discrete domain of variance rate linguistic values EC: y { -6, -5, ·,0 …, +5, +6 };

The linguistic values of the linguistic variable deviation EC are selected as: PB, PM, PS, ZO, NS, NM and NB7 language values;

Assignment table of language variable EC

Basic domain of control quantity change u: [ -60min, +60min ];

the scaling factor of the control quantity change u is then: k _u ＝[60-(-60)]/2*7＝60/7；

Assignment table of linguistic variable U

3) Determination of fuzzy control rules:

when the error, namely the desired value minus the detection value, is negative and large, the detection value is higher than the desired value, and when the error change rate is also negative, the error has an increasing trend, so that the existing negative and large error is eliminated and the increase of the error is restrained as soon as possible, and therefore, the change of the control quantity is positive and large; when the error is negative and the error change rate is positive, the system has the tendency of reducing the error, so that a smaller control quantity is taken to eliminate the error as soon as possible without overshoot; when the error is negative, the change of the control quantity can eliminate the error as soon as possible, and based on the principle, the change of the control quantity is the same when the error is negative; when the error is negative, the system is close to a steady state, and if the error changes slightly, the control quantity is selected to be changed to be positive so as to inhibit the error from changing to the negative direction; when the error changes to be positive, the system has the tendency of eliminating small negative errors, and the selected control quantity changes to be positive or negative;

obtaining a fuzzy control rule table according to the rules

Fuzzy control rule table

4) fuzzification and fuzzy reasoning:

Quantized to obtain e ^* And ec ^* And is and

e ^* ∈X＝{-6,-5,...,-0,+0…,+5+6},

ec ^* ∈Y＝{-6,-5,...,0,…,+5+6}

e is to be ^* And ec ^* Fuzzification to obtain fuzzy subset E ^* And EC ^*

Then

5) defuzzification of the control quantity:

the fuzzy control increment is deblurred by adopting a maximum membership method to obtain a control increment u which takes a value on a discourse domain Z { -7 ^* ；

Fuzzy control look-up table

And (3) fuzzy control lookup table off-line calculation:

according to

in the case of on-line control, the measured deviations e (k), ec (k), and (k ═ 0,1,2, and … …) are quantized in each control cycle, and e required for searching the fuzzy control table is obtained ^* 、ec ^* Then, the corresponding control variable change u is looked up from the table ^* And a scale factor k _u The multiplication results in a change u of the actual controlled variable for the controlled object, i.e. u-u ^* *k _u ；

And (3) final control process: the fuzzy control variable quantity is obtained by combining the temperature deviation detected in real time and the deviation change rate, and the required actual control quantity is output through a series of conversion of anti-fuzzy control: when the actual measured temperature is 14 ℃, the target temperature is 28 ℃, the temperature difference is 14 ℃, the change rate of the temperature difference during heating is positive and has a decreasing trend, and a fuzzy control range is found by calculating and looking up a table, and u can be selected ^* 2, u-2 × 60/7-17.14, so from the current detection situation,and continuously heating for about 17 minutes, simultaneously performing real-time detection, inputting the detection result into a fuzzy control system, performing real-time correction control output, and finally stabilizing the temperature at a target value of 28 ℃ to ensure that a mill lubricating system reaches an optimal lubricating operation state.

Claims

1. The utility model provides a large-scale mill process control system, installs on the mill, its characterized in that this control system includes inside framework, outside auxiliary unit and mill monitoring system, and inside framework, outside auxiliary unit and mill monitoring system pass through the wire and connect, wherein:

the internal architecture comprises a logic linkage control unit (1), an operation state monitoring unit (2), a screening control processing unit (3), a prediction control processing unit (4), a data history storage unit (5) and an online fault processing unit (6), wherein the screening control processing unit (3) is respectively connected with the logic linkage control unit (1) and the data history storage unit (5), the prediction control processing unit (4) is respectively connected with the logic linkage control unit (1) and the data history storage unit (5), the logic linkage control unit (1) and the data history storage unit (5) are respectively connected with the operation state monitoring unit (2), and the operation state monitoring unit (2) is connected with the online fault processing unit (6);

the external auxiliary unit comprises a remote transmission monitoring unit (7) and an audible and visual alarm unit (8);

the mill monitoring system comprises a temperature sensor (9), an oil pressure sensor (10), a flow sensor (11) and a vibration sensor, a plurality of temperature measuring points, an oil pressure monitoring point and an oil amount monitoring point are arranged on the mill, three temperature measuring points are respectively arranged on bearings (12) at a feeding end and a discharging end, three temperature measuring points are arranged on the surface of a pinion (13), two temperature measuring points are respectively arranged at two ends of the pinion (13), the temperature sensor (9) is arranged on each temperature measuring point, the temperature of the temperature measuring points is collected through the temperature sensor (9), and the temperature sensor (9) of the temperature measuring points on the surface of the pinion (13) adopts an infrared temperature sensor; flow sensors (11) are arranged on oil nozzles of bearings (12) at two ends of a pinion (13) and two ends of two cylinders, and an oil pressure sensor (10) is arranged on an oil supply pipeline (14); x, Y, Z vibration detection in three directions is established at the driving end, the non-driving end and the vertical direction of the pinion (13), each pinion (13) is provided with 6 vibration sensors, and two pinions (13) of one mill are provided with 12 vibration sensors in total; the mill monitoring system is connected with the screening control processing unit (3), and the obtained monitoring data are transmitted to the screening control processing unit (3);

the specific control process is as follows:

1) the discrimination control processing unit (3) acquires the infrared temperature difference of the pinion (13), the infrared temperature of the pinion (13), oil flow, oil pressure, the temperature of the bearing (12) and the vibration range, judges the distortion condition through fuzzy control, judges the distortion condition as a distortion signal, does not participate in logic interlocking control, enters the historical data recording unit for storage and is only used for inquiry; if the signal is judged to be a true signal, the data is transmitted to a logic linkage control unit (1) to participate in logic linkage control;

2) after the logic linkage control unit (1) receives the data signals of the discrimination control processing unit (3), the signals are analyzed, the fault problem is judged, and the set value of the optimal use value is searched by utilizing fuzzy control; firstly, each temperature sensor (9) sets the disconnection temperature of-200 ℃, directly transmits information to an online fault processing unit (6) when detecting the disconnection temperature, and sends out an alarm signal; setting a normal use temperature value as 60 ℃, obtaining an ultrahigh temperature of 90 ℃ by utilizing fuzzy control calculation, and sending a fault shutdown instruction of the detection point and sending alarm information when the ultrahigh temperature value is exceeded; setting the normal range of the temperature difference value at two ends of a pinion (13) to be 15 and the ultra-large temperature difference value to be 30 according to the information of an infrared temperature sensor, sending an alarm signal when the temperature difference value exceeds 15, and sending a fault shutdown instruction when the temperature difference value exceeds 30; the oil pressure of the oil supply pipe (14) is set to 40kg/cm at a low pressure within a normal use range ³ The ultra-low pressure is 15kg/cm ³ When the signal reaches a low-voltage value, outputting alarm information of a corresponding detection point, and when the signal reaches an ultra-low-voltage value, sending a fault shutdown instruction;

3) the online fault processing unit (6) mainly aims at two types of fault processing, namely lubricating oil fault processing and temperature detection fault processing, and the specific processing mode is as follows:

(1) the lubricating oil fault processing is carried out on-line processing without stopping, and comprises two functions of lubricating oil fault processing waiting intermittence and timing zero setting, wherein the waiting intermittence time is preset after the lubricating oil injection period is finished, the timing zero setting function is added into the waiting intermittence time function, the fault processing of a lubricating oil system is carried out in the waiting intermittence time, the timing zero setting function is started in the waiting intermittence time, the waiting intermittence time is restarted, the fault processing is ensured to have sufficient time, and in the processing process, the lubricating oil is supplemented in time in a manual mode by manually observing the condition of the lubricating oil until the fault is eliminated;

(2) the temperature detection fault processing is to monitor the circuit and the probe of each temperature sensor (9), when any one temperature sensor (9) sends out a disconnection fault, disconnection information is displayed, and maintainers are prompted to conduct targeted inspection and processing, so that the shutdown caused by the fault generated by high temperature is avoided;

4) the prediction control processing unit (4) is used for setting an exceeding area limit value for normal operation according to real-time detection data and historical data online trend records and in combination with a normal use range index value range of equipment performance, and when the detection data exceeds the area limit value, early warning is carried out in advance to inform maintenance personnel to carry out performance analysis and maintenance on a bearing (12), a bearing bush and a gear of the equipment in advance;

5) the sound and light alarm unit (8) is matched with an internal framework to carry out early warning on operation and maintenance personnel in time, so that the operation and maintenance personnel can take measures conveniently;

6) and the remote transmission monitoring unit (7) transmits the relevant information of the operation of the mill to the distributed control system, and performs real-time and historical storage, trend query, state information display and control functions on a server of the distributed control system, so that a central control operator can conveniently and timely master information of one hand of the operation of the mill and accurately issue corresponding instructions.

2. A large mill process control system as claimed in claim 1, characterized in that in step 1) of the control process, the fuzzy control distortion condition determining step is as follows:

the bias linguistic variable E ═ r ₂ –y ₂ ，r ₂ Set value for distortion degree, y ₂ Is a distortion test value;

a rate of change of deviation ec (k) ═ E (k) — E (k-1), where k ═ 0,1,2, … …;

outputting language variables: controlling an output language variable U of the filtering time length in the distortion degree testing process;

2) determination of linguistic values of input deviation linguistic variable E, deviation change rate EC, and output linguistic variable U:

basic domain of discordance linguistic variable E: [ -20, +20 ];

discrete domain of the bias linguistic variable E: x _E ＝{-6,...,0…,+6}；

The quantization factor for the bias linguistic variable E is: k _e ＝2*6/[20-(-20)]＝0.3；

The linguistic values of the biased linguistic variable E are selected as: PB, PM, PS, PO, NO, NS, NM, NB 8 linguistic values:

determined at discrete discourse domain X _E The membership functions μ (x) of the fuzzy subsets PB, …, NB are described above, and the assignment table of the biased linguistic variables E is built accordingly:

assignment table for bias linguistic variable E

Basic domains of variation rate of deviation EC: [ -15, +15 ];

discrete domains of variation rate EC: y is _EC ＝{-6，-5,...,0…,+5，+6}；

The linguistic values of the deviation change rate EC are selected as: PB, PM, PS, ZO, NS, NM, NB7 language values;

determined in discrete universe of discourse Y _EC The membership functions μ (y) of the fuzzy subsets PB, …, NB are described above, and the assignment table of the deviation change rates EC is established accordingly:

assignment table of deviation change rate EC

Basic domain of output linguistic variable U: [ -10s, +10s ];

discrete domain of output linguistic variable U: z _U ＝{-7,-6,...,0,…,+6,+7}；

The scale factor of the output linguistic variable U is: k _u ＝[10-(-10)]/2*7＝10/7；

The linguistic value of the output linguistic variable U is selected as: PB, PM, PS, ZO, NS, NM, NB7 language values;

determined at discrete universe of discourse Z _U The above describes the membership functions μ (z) of the fuzzy subsets PB, …, NB, and accordingly establishes the assignment table of the output linguistic variables U:

assignment table for output linguistic variable U

3) Determination of fuzzy control rules:

when the error, namely the expected value minus the test value, is negative and large, the test value is higher than the expected value, when the error change rate is negative, the error has an increasing trend, and in order to eliminate the existing negative and large error and restrain the increase of the error as soon as possible, the change of the control quantity is positive and large, namely the control quantity is increased, which means that the filtering time control setting is increased, so that the distortion degree is reduced; when the error is negative and the error change rate is positive, the system has the tendency of reducing the error, and a smaller control quantity is taken for eliminating the error as soon as possible without overshoot; when the error is negative or medium, the change of the control quantity can eliminate the error as soon as possible; based on the principle, the change of the control quantity is the same when the same error is negative; when the error is negative, the system approaches a steady state; when the error change is tiny, the control quantity is selected to be changed to be positive so as to restrain the error from changing to the negative direction; when the error changes to be positive, the system has the tendency of eliminating negative small errors, and the selection control quantity changes to be positive small; obtaining a fuzzy control rule table:

fuzzy control rule table

R＝(E×U)○(EC×U)＝E×EC×U

＝(NB _E ×NB _EC ×PB _U )∪(NM _E ×NB _EC ×PB _U )∪(NS _E ×NB _EC ×PM _U )∪……

4) fuzzification and fuzzy reasoning:

let the measured deviations be e (k), ec (k) ═ e (k) — e (k-1), where k ═ 0,1,2, … …

Quantized to obtain e ^* ，ec ^*

And e ^* ∈X _e* ＝{-6,-5,...,-0,+0…,+5，+6},ec ^* ∈Y _ec* ＝{-6,-5,...,0,…,+5，+6}

E is to be ^* 、ec ^* Fuzzification to obtain fuzzy subset E ^* 、EC ^*

Then the

I.e. the output of the fuzzy controller is the fuzzy subset E of error vectors ^* Error rate of change vector fuzzy subset EC ^* Synthesizing fuzzy relation;

5) defuzzification of the control quantity:

generating a fuzzy control solving query table, and solving the fuzzy control increment by adopting a maximum membership method to obtain a Z domain of discourse _u* A control increment u taken on { -7., 0, …, +7} ^* To discourse field X _u* And discourse domain Y _u* All the elements are combined to carry out the fuzzification, fuzzy inference and fuzzy decision operation, and the universe of discourse Z can be obtained _u* Is in one-to-one correspondence with the element(s) above, wherein X _u* ＝{-6,...,-0,+0，…,+6}，Y _u* ＝{-6,...,0，…,+6}，Z _u* The corresponding relations are made into a table, namely { -7., 0, …, +7}, and then the fuzzy control lookup table is obtained:

fuzzy control look-up table

The fuzzy control look-up table is obtained by off-line calculation according to

At E ^* ＝(PB,PM，PS,PO，NO,NS,NM，NB),EC ^* When the output fuzzy vector U of the controller is (PB, PM, PS, ZO, NS, NM, NB), the output fuzzy vector U of the controller is calculated ^* Then performing defuzzification according to a maximum membership method, and selecting the control quantity as U ^* The maximum value calculated on (PB, PM, PS, ZO, NS, NM, NB) is U ^* _max Filling the values into the fuzzy control lookup table; in the case of on-line control, the deviation will be measured in each control period

e (k), ec (k) and quantization are respectively carried out, wherein k is 0,1,2, … …

Obtaining e required for searching fuzzy control lookup table ^* 、ec ^* Then look up the corresponding control increment u from the table ^* With a scale factor k _u Multiplying the output speech variable U by the actual control variable U of the controlled object ^* *k _u ；

The purpose of the fuzzy control is to remove the distortion degree, and then all u with e less than or equal to 0 is included ^* In addition k _u 10/7; the filtering time is set to be not less than 5s, and U is ^* *k _u Calculating for more than or equal to 5s to obtain u ^* ≧ 3.5, so, under the fuzzy controller, all u ^* The control quantity more than 3 can meet the requirement of removing the distortion degree; but u ^* If the selection is too large, the output linguistic variable U of the actual control quantity is also too large, namely the filtering time is increased, so that the normal signal entering the logic linkage control unit (1) in the control system is seriously lagged, the system cannot process the external normal signal in time, the overshoot is serious, the superiority of the system is lost, and the control quantity U of the optimal distortion degree ^* It should satisfy: u is more than 3 ^* 4, the optimal filtering time length for controlling the system signal processing unit to remove the signal distortion degree is set to be 5 s.

3. A large mill process control system as claimed in claim 1, characterized in that in step 2) of the control process, the specific steps of finding the set value of the optimum use value by using fuzzy control are as follows:

the bias linguistic variable E ═ r ₃ –y ₃ ，r ₃ Is a target value of the set value, y ₃ A plurality of test values;

a rate of change of deviation ec (k) E (k) -E (k-1), where k is 0,1,2, … …;

outputting language variables: setting an output language variable U for controlling the length of a set value in the target stable operation process;

2) determining linguistic values of an input deviation linguistic variable E, a deviation change rate EC and an output linguistic variable U; forming a two-dimensional coordinate system, and performing fuzzy calculation by adopting a triangular or semi-trapezoidal membership function relationship;

3) determination of fuzzy control rules:

when the error is large, the control quantity is selected to eliminate the error as soon as possible; when the error is small, the control quantity is selected to be careful to prevent overshoot, and according to the above rule, the following fuzzy control rule table is obtained:

fuzzy control rule table

Obtaining a fuzzy relation R between the deviation of the set value, the deviation change rate and the length of the set value according to the control rule:

R＝(E×U)○(EC×U)＝E×EC×U

4) fuzzification and fuzzy reasoning:

Quantized to obtain e ^* 、ec ^*

And e ^* ∈X _e* ＝{-6,-5,...,-0,+0…,+5，+6}

ec ^* ∈Y _ec* ＝{-6,-5,...,0,…,+5，+6}

E is to be ^* And ec ^* Fuzzification to obtain fuzzy subset E ^* And EC ^*

Then

5) defuzzification of the control quantity:

the fuzzy control increment is deblurred by adopting a maximum membership method to obtain a domain of discourse Z _u* A control increment u taken on { -7., 0, …, +7} ^* (ii) a Discourse domain X _u* And discourse domain Y _u* All the elements are combined to carry out the fuzzification, fuzzy inference and fuzzy decision operation to obtain a universe of discourseZ _u* The elements above correspond to one another, where X { -6., -0, +0, …, +6}, Y _u* ＝{-6,...,0，…,+6}，Z _u* The corresponding relations are made into a fuzzy control lookup table:

fuzzy control look-up table

The fuzzy control lookup table is obtained by off-line calculation:

according to

in the case of online control, the measured deviations e (k), ec (k) are quantified in each control cycle, where k is 0,1,2, … …

E required for obtaining the look-up fuzzy control table ^* 、ec ^* Then look up the corresponding control increment u from the table ^* With a scale factor k _u Multiplying the output speech variable U by the actual control variable U of the controlled object ^* *k _u ；

(1) searching the optimal set value of infrared temperature difference alarm:

selecting: basic domain of output linguistic variable U of alarm value: [ -15, +15 ];

the scale factor of the output linguistic variable U is: k _u ＝[10-(-10)]/2*7＝15/7；

By U-U ^* *k _u Calculated to obtain u ^* 7; the maximum membership value-taking principle is met, so the optimal setting of the alarm value is 15 ℃;

selecting: basic domain of output linguistic variable U of alarm value: [ -30, +30 ];

the scale factor of the output linguistic variable U is: k is _u ＝[10-(-10)]/2*7＝30/7；

From U ═ U ^* *k _u Calculated to obtain u ^* And 7, the maximum membership value principle is satisfied, so the optimal fault shutdown value is set to be 30 ℃.

4. A large mill process control system as claimed in claim 1, wherein the specific control process is further provided with a lubricating oil constant temperature control process, according to the field use condition, stable and efficient operation of the mill is required to be maintained, the lubricating control system must stably operate at constant temperature, and the temperature of the mill lubricating system is stable at 28 ℃ and is in an optimal state; the control process utilizes fuzzy calculation to control, and the specific control steps are as follows:

temperature deviation linguistic variable E ═ r ₄ –y ₄ ，r ₄ For lubricating system temperature set point, y ₄ The temperature detection value of the lubricating system is obtained;

a rate of change in temperature deviation ec (k) ═ E (k) — E (k-1), where k is 0,1,2, … …;

outputting language variables: an output language variable U for controlling the action duration of the heater in the constant temperature keeping process;

basic domain of the bias linguistic variable E: [ -10 ℃ and +10 ℃ C ];

discrete domain of deviation linguistic variable E: x _E ＝{-6,...,0…,+6}；

The quantization factor for the bias linguistic variable E is: k _e ＝2*6/[10-(-10)]＝0.6；

The linguistic values of the biased linguistic variable E are selected as: PB, PM, PS, PO, NO, NS, NM and NB 8 language values;

determined at discrete discourse domain X _E The membership function μ (x) used to describe the fuzzy subsets PB, …, NB is a triangular membership function, and a table of assignments of the biased linguistic variables E is established accordingly:

assignment table for bias linguistic variable E

Basic domains of variation rate of deviation EC: [ -15, +15 ];

The quantization factor for the rate of change of deviation EC is then: k is _ec ＝2*6/[15-(-15)]＝6/15；

The linguistic values of the deviation change rate EC are selected as: PB, PM, PS, ZO, NS, NM, and NB7 linguistic values;

determined in discrete universe of discourse Y _EC The membership function μ (y) for the fuzzy subsets PB, …, NB is the triangular membership function, and the assignment table of the variance rate EC is established accordingly:

assignment table of deviation change rate EC

Basic domain of output linguistic variable U: [ -60min, +60min ];

The scale factor of the output linguistic variable U is: k _u ＝[60-(-60)]/2*7＝60/7；

The linguistic value of the output linguistic variable U is selected as: PB, PM, PS, ZO, NS, NM and NB7 language values;

determined in discrete universe of discourse Z _U The membership function μ (z) for describing the fuzzy subsets PB, …, NB is a triangular membership function, and accordingly, an assignment table of the output linguistic variable U is established:

assignment table for output linguistic variable U

3) Determination of fuzzy control rules:

when the error, namely the desired value minus the detection value is negative and large, the detection value is higher than the desired value, when the error change rate is also negative, the error has an increasing trend, and the change of the control quantity is positive and large in order to eliminate the existing negative and large error and restrain the increase of the existing negative and large error as soon as possible; when the error is negative and the error change rate is positive, the system has the tendency of reducing the error, so that a smaller control quantity is taken to eliminate the error as soon as possible without overshoot; when the error is negative, the change of the control quantity can eliminate the error as soon as possible, and based on the principle, the change of the control quantity is the same when the error is negative; when the error is negative, the system is close to a steady state, and if the error changes slightly, the control quantity is selected to be changed to be positive so as to inhibit the error from changing to the negative direction; when the error changes to be positive, the system has the tendency of eliminating negative small errors, and the selection control quantity changes to be positive small;

according to the rules, obtaining a fuzzy control rule table:

fuzzy control rule table

According to the control rule, obtaining a fuzzy relation R between the temperature deviation, the deviation change rate and the heater action time length:

R＝(E×U)○(EC×U)＝E×EC×U

4) fuzzification and fuzzy reasoning:

Quantized to obtain e ^* And ec ^* And is and

e ^* ∈X _e* ＝{-6,-5,...,-0,+0…,+5+6},

ec ^* ∈Y _ec* ＝{-6,-5,...,0,…,+5+6}

e is to be ^* And ec ^* Fuzzification to obtain fuzzy subset E ^* And EC ^*

Then

5) defuzzification of the control quantity:

the fuzzy control increment is subjected to ambiguity resolution by adopting a maximum membership method to obtain a domain Z _u* A control increment u taken on { -7., 0, …, +7} ^* ；

To domain X _u* And discourse domain Y _u* All the elements of the system are combined to carry out fuzzification, fuzzy reasoning and fuzzy decision operation, and a universe of discourse Z can be obtained _u* Is in one-to-one correspondence with the element(s) above, wherein X _u* ＝{-6,...,-0,+0，…,+6}，Y _u* ＝{-6,...,0，…,+6}，Z _u* And (7) { -7., 0, …, +7}, and making the corresponding relations into a table to obtain a fuzzy control lookup table:

fuzzy control look-up table

And (3) fuzzy control lookup table offline calculation:

according to

At E ^* ＝(PB,PM，PS,PO，NO,NS,NM，NB),EC ^* When the output fuzzy vector U of the controller is (PB, PM, PS, ZO, NS, NM, NB), the fuzzy vector U of the output of the controller is calculated ^* Then carrying out defuzzification according to a maximum membership method, and selecting the control quantity as U ^* The maximum value calculated on (PB, PM, PS, ZO, NS, NM, NB) is U ^* _max Filling the values into the fuzzy control lookup table;

E required for obtaining lookup fuzzy control table ^* 、ec ^* Then look up the corresponding control increment u from the table ^* With a scale factor k _u Multiplying to obtain an output linguistic variable U for the actual controlled variable of the controlled object, i.e. U ═ U ^* *k _u ；

And (3) final control process: the fuzzy control variable quantity is obtained by combining the temperature deviation detected in real time with the deviation change rate, and the required actual control quantity is output through a series of conversion of anti-fuzzy control.