CN115930594B - Shaft furnace production control method - Google Patents
Shaft furnace production control method Download PDFInfo
- Publication number
- CN115930594B CN115930594B CN202310137381.7A CN202310137381A CN115930594B CN 115930594 B CN115930594 B CN 115930594B CN 202310137381 A CN202310137381 A CN 202310137381A CN 115930594 B CN115930594 B CN 115930594B
- Authority
- CN
- China
- Prior art keywords
- temperature
- recommended
- gas flow
- combustion chamber
- flow
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Landscapes
- Heat Treatment Of Articles (AREA)
- Waste-Gas Treatment And Other Accessory Devices For Furnaces (AREA)
- Regulation And Control Of Combustion (AREA)
Abstract
The invention discloses a shaft furnace production control method, which is characterized in that the initial gas flow and the gas hood temperature are respectively predicted by a gas flow prediction model and a gas hood temperature prediction model and are input into a PID control model to obtain the valve opening of an air pipeline, a gas pipeline and a cooling air pipeline, the continuous combustion temperature of a shaft furnace combustion chamber can be controlled at the lower limit of temperature control, and the return rate of finished balls is ensured by controlling the temperature of the combustion chamber and the cooling air flow.
Description
Technical Field
The invention relates to the field of shaft furnace production processes, in particular to a shaft furnace production control method.
Background
The shaft furnace is a roasting furnace with a vertical furnace body, a vertical shaft is arranged on a furnace cover, and green pellets are preheated in the vertical shaft by utilizing high-temperature waste gas discharged by a combustion chamber.
The green pellets are continuously and uniformly distributed into the shaft furnace through a distributor, and are subjected to five stages of drying, preheating, roasting, soaking and cooling. The roasted pellets are uniformly discharged out of the furnace from the bottom of the shaft furnace, and the volleyball amount and the raw ball amount distributed in the operation process of the shaft furnace are required to be basically balanced, so that the shaft furnace production is a continuous operation process.
Raw balls are distributed into the shaft furnace by a shuttle distributor, move downwards at a certain speed, and are sprayed into the furnace from a flame path opening by high-temperature gas generated by combustion of gas combustion chambers arranged on two sides of the shaft furnace, so that the raw balls are roasted. And meanwhile, the lower cooling belt cools the heat generated by heat exchange and also provides a certain amount of heat for the roasting process. The green pellets are firstly dried and dehydrated on a baking bed at the upper part of the shaft furnace, preheated and oxidized, enter a roasting belt, are solidified at high temperature, pass through a heat equalizing belt, complete the whole solidification process, exchange heat between the solidified pellets and cooling air blown into the furnace from the lower part upwards, obtain cooled pellets (namely finished pellets) and are discharged from the bottom of the shaft furnace through a toothed roller. Most of the cooling air (hot air) after heat exchange is converged with hot waste gas of the combustion chamber under the fire grate through the air guide wall, green pellets are dried through the fire grate drying bed, then discharged from the top of the furnace, dedusted through the dust remover, and finally discharged into the chimney to the atmosphere.
The entire firing process of the pellets is thus essentially completed in a shaft furnace. Shaft furnaces are heat exchange devices operating on the countercurrent principle, i.e. the flow of material runs in the top-down direction and the flow of gas runs in the bottom-up direction.
However, the temperature of the combustion chamber, the temperature of the smoke hood and the gas flow of the combustion chamber are set and regulated manually, the manual regulation is highly dependent on operators, no small burden is caused to the operation of workers, meanwhile, the regulation methods of each worker are different, and the real-time accurate regulation is difficult to be carried out according to the actual technological condition all the time, so that the stability of the quality of the finished ball is not controlled.
Disclosure of Invention
In order to solve the technical problems, the invention provides a shaft furnace production control method.
In order to solve the technical problems, the invention adopts the following technical scheme:
a shaft furnace production control method comprising the steps of:
step A1: training a gas flow prediction model by utilizing historical production data D1 generated when the shaft furnace is operated manually; the production data D1 comprises gas flow, air flow, combustion chamber temperature and raw ball belt scale flow of a combustion chamber;
step A2: inputting production data D1 at the current moment into a coal gas flow prediction model which is trained, and predicting to obtain recommended coal gas flow of a combustion chamber at the current moment;
step A3: the control of the temperature of the combustion chamber is carried out through a PID control model of the temperature of the combustion chamber, and specifically comprises the following steps:
step A31: taking the recommended gas flow obtained in the step A2 as the initial recommended gas flowInputting the temperature of the combustion chamber into a PID control model;
step A32: based on the difference between the average value of the actual combustion chamber temperature and the target combustion chamber temperature in the previous T timeFor the initial recommended gas flow +.>Adjusting to obtain new recommended gas flow +.>;/>
Step A33: will recommend the gas flowRecommended gas flow predicted at the last moment of gas flow prediction modelFusion is carried out to obtain final recommended gas flow at the current moment>;
Step A34: setting an initial recommended airflow for the combustion chamberBy difference->The recommended air flow is adjusted as the final recommended air flow at the current moment +.>;
Step A35: setting the corresponding relation between different temperature levels and gas flow rate change valuesTying; when the actual temperature of the combustion chamber is lower than the set low temperature value, the gas flow is recommended according to the temperature level of the actual temperature of the combustion chamberIncreasing the corresponding gas flow on the basis of the above; when the actual temperature of the combustion chamber is higher than the set low temperature value, according to the temperature level of the actual temperature of the combustion chamber, the gas flow is recommended>On the basis of the above, reducing the corresponding gas flow;
step A36: recommended gas flow at the current momentConverting into the opening degree of a valve of a gas pipeline, and adding the recommended air flow rate at the current moment>The temperature of the combustion chamber is controlled by converting the temperature into the opening degree of the valve of the air pipeline.
Specifically, when the gas flow prediction model is trained in the step A1, the method comprises the following steps:
step A11: collecting historical production data D1 generated when the shaft furnace is operated manually;
step A12: the historical production data D1 is divided periodically with fixed duration to obtain historical production cycle data D1', and the historical production cycle data D1' with the first 30% of the cycle with the minimum average value of the gas flow of the combustion chamber and the flow of the green belt scale is selected as training data;
step A13: inputting training data into different machine learning models, wherein the target value of the machine learning model is the gas flow at the current moment, and the variables are the raw ball belt scale flow and the gas flow at the first five sampling moments;
step A14: adopting cross verification contrast, and selecting a machine learning model with the best fitting as a coal gas flow prediction model;
and (A2) when the recommended gas flow of the combustion chamber at the current moment is predicted, taking the current moment as a reference, inputting the raw ball belt scale flow at the first five sampling moments and the gas flow at the first five sampling moments into a gas flow prediction model to obtain the recommended gas flow of the combustion chamber at the current moment.
Specifically, the combustion chamber target temperature is equal to an average of the set combustion chamber temperature upper limit value and the set combustion chamber temperature lower limit value.
Specifically, in step A32, according to the differenceFor the initial recommended gas flow +.>When the adjustment is performed, the difference value is obtainedMultiplying a fixed scaling factor as an adjustment factor +.>By adjusting the coefficient->Recommended gas flow->And (5) adjusting.
Specifically, in step A33, the recommended gas flow rate is setRecommended gas flow predicted at the previous time with gas flow prediction model +.>Fusion is carried out to obtain recommended gas flow at the current moment>When (I)>。
Further, the method also comprises a fume hood temperature prediction control process at the top of the shaft furnace, and specifically comprises the following steps:
step B1: training a hood temperature prediction model using historical production data D2 generated when the shaft furnace is operated manually; the historical production data D2 comprises gas flow, air flow, combustion chamber temperature, raw ball belt scale flow and smoke hood temperature of the combustion chamber;
step B2: inputting historical production data D2 at the current moment into a trained hood temperature prediction model, and predicting to obtain a target hood temperature at the current moment;
the temperature of the fume hood is controlled by a fume hood temperature PID control model, and the method specifically comprises the following steps:
step B31: taking the target hood temperature obtained in the step B2 as the initial target hood temperatureInputting the temperature of the fume hood into a PID control model;
step B32: according to the difference between the average value of the actual hood temperature and the target hood temperature in the previous T timeFor the initial recommended cooling air flow set +.>Adjusting to obtain new recommended cooling air flow +.>;
Step B33: will recommend the cooling air flowRecommended cooling air flow rate +.A recommended cooling air flow rate obtained from a target hood temperature predicted at a previous time on a hood temperature prediction model>Fusion is carried out to obtain final recommended cooling air flow at the current moment>;
Step B34: setting the corresponding relation between different temperature levels and cooling air flow change values; when the actual temperature of the fume hood is lower than the set low temperature value, recommending cooling air flow according to the temperature level of the actual temperature of the fume hoodOn the basis of the above, increasing the corresponding cooling air flow; when the actual temperature of the fume hood is higher than the set low temperature value, the flow of the cooling air is recommended according to the temperature level of the actual temperature of the fume hood>On the basis of the above, the corresponding cooling air flow is reduced;
step B35: recommended cooling air flow rate at current momentThe temperature of the fume hood is controlled by converting the temperature into the opening of a valve of the cooling pipeline.
Specifically, when the hood temperature prediction model is trained in the step B1, the method comprises the following steps:
step B11: collecting historical production data D2 generated when the shaft furnace is operated manually;
step B12: the historical production data D2 is divided periodically with fixed duration to obtain historical production cycle data D2', and the historical production cycle data D2' with the first 30% of the cycle with the smallest average value of the gas flow of the combustion chamber and the flow of the green belt scale is selected as training data;
step B13: inputting training data into different machine learning models, wherein the target value of the machine learning model is the temperature of a smoke hood at the current moment, and the variables are the flow of the green-ball belt scale, the flow of cooling air and the average temperature of a combustion chamber at the first five sampling moments;
step B14: adopting cross verification contrast, and selecting a machine learning model with the best fitting as a smoke hood temperature prediction model;
and B2, when the target smoke hood temperature at the current moment is predicted, taking the current moment as a reference, inputting the flow of the green-ball belt scale, the flow of cooling air and the average temperature of the combustion chamber at the first five sampling moments into a smoke hood temperature prediction model to obtain the target smoke hood temperature at the current moment.
Specifically, in step B32, the difference is usedFor the initial recommended cooling air flow +.>When the adjustment is performed, the difference value is obtainedMultiplying a fixed scaling factor as an adjustment factor +.>By adjusting the coefficient->Recommended cooling air flow->And (5) adjusting.
Specifically, in step B33, the cooling air flow rate will be recommendedRecommended cooling air flow rate +.A recommended cooling air flow rate obtained from a target hood temperature predicted at a previous time on a hood temperature prediction model>Fusion is carried out to obtain recommended cooling air flow at the current moment>In the time-course of which the first and second contact surfaces,。
compared with the prior art, the invention has the beneficial technical effects that:
according to the invention, the minimum gas distribution amount can be realized according to the amount and the components of the incoming green ball materials, and the gas is saved by more than 2 percent; the temperature of continuous combustion in the combustion chamber of the shaft furnace is kept at the lower limit of temperature control, and the self-control rate is more than 95%; the return rate of the finished ball can be ensured to be below 7% by controlling the temperature of the combustion chamber and the flow of cooling air.
Drawings
FIG. 1 is an overall flow chart of the present invention;
FIG. 2 is a schematic diagram of a gas flow prediction model construction and prediction flow;
FIG. 3 is a schematic diagram of a construction and prediction flow of a smoke hood temperature prediction model according to the present invention.
Detailed Description
A preferred embodiment of the present invention will be described in detail with reference to the accompanying drawings.
1. Predicting recommended gas flow for a combustor
Predicting the gas flow of the combustion chamber at the current moment by using a machine learning model according to the historical production data of the shaft furnace and the current furnace condition; the operation interval time of the gas flow prediction model is one minute, and the recommended gas flow predicted by the gas flow prediction model is fed back to the combustor temperature PID control model and is used for carrying out initial assignment on the recommended gas flow of the combustor.
In this embodiment, the combustion chamber includes an east combustion chamber and a west combustion chamber, and the gas flow prediction model includes an east combustion chamber gas flow prediction model and a west combustion chamber gas flow prediction model. The overall idea of predicting recommended gas flows for the east and west combustors is as follows.
(1) And collecting historical production data D1 of the manual furnace to generate a training set. The historical production data D1 includes gas flow rate of the combustion chamber, air flow rate, temperature of the combustion chamber, and raw ball belt scale flow rate.
(2) The shaft furnace is a continuous process, the invention takes one hour as a period, divides the historical production data D1 of the normal stage of the process into the historical production period data D1', and selects the historical production period data D1' of the first 30% period with the smallest average value of the gas flow of the combustion chamber and the flow of the green-ball belt scale as training data. And respectively obtaining training data for training the east combustor gas flow prediction model and training data for training the west combustor gas flow prediction model according to the screening conditions.
(3) The target value of the gas flow prediction model is the gas flow at the current moment, and the variables are the raw ball belt scale flow at the first five sampling moments and the gas flow at the first five sampling moments. In this embodiment, the training data is minute-level data, and one sampling time is one minute.
(4) The invention adopts a common machine learning model of cross verification comparison, and takes average absolute error, mean square error, root mean square error, decision coefficient or average absolute percentage error and the like as indexes for selecting the model, wherein the model selection standard is average absolute error, mean square error, root mean square error or average absolute percentage error as small as possible, but the decision coefficient is as large as possible. In the embodiment, the compatibility_models of the pycaret module of machine learning are adopted, and a machine learning model with the best fitting is selected as a coal gas flow prediction model.
(5) When the recommended gas flow of the combustion chamber at the current moment is predicted, the raw ball belt scale flow at the first five sampling moments and the gas flow at the first five sampling moments are input into a gas flow prediction model to obtain the recommended gas flow of the combustion chamber at the current moment. The recommended gas flow of the combustion chamber at the current moment comprises the recommended gas flow of the combustion chamber at the east at the current moment and the recommended gas flow of the combustion chamber at the west at the current moment.
2. Predicting fume hood temperature at top of shaft furnace
Predicting the temperature of a fume hood at the top of the shaft furnace at the current moment by using a machine learning model according to historical production data of the shaft furnace and the current furnace condition; the operation interval time of the smoke hood temperature prediction model is one minute, and the target smoke hood temperature predicted by the smoke hood temperature prediction model is fed back to the smoke hood temperature PID control model and is used for carrying out initial assignment on the recommended smoke hood temperature. The overall idea for predicting the temperature of the fume hood at the top of the shaft furnace is as follows.
(1) And collecting historical production data D2 of the manual furnace, and generating a training set. The historical production data D2 includes gas flow to the combustor, air flow, combustor temperature, green-ball belt scale flow, and hood temperature.
(2) The shaft furnace is a continuous process, the invention takes one hour as a period, divides the historical production data D2 of the normal stage of the process into the historical production period data D2', and selects the historical production period data D2' of the first 30% period with the smallest average value of the gas flow of the combustion chamber and the flow of the green-ball belt scale as training data.
(3) The target value of the smoke hood temperature prediction model is the smoke hood temperature at the current moment, and the variables are the flow of the green-ball belt balance and the flow of cooling air at the first five sampling moments and the average temperature of the combustion chamber at the first five sampling moments; in this embodiment, there are two east and west combustion chambers, and the average temperature of the combustion chambers is the average of four combustion chamber temperatures of the two east combustion chambers and the two west combustion chambers.
(4) The invention adopts a common machine learning model of cross verification comparison, and takes average absolute error, mean square error, root mean square error, decision coefficient or average absolute percentage error and the like as indexes for selecting the model, wherein the model selection standard is average absolute error, mean square error, root mean square error or average absolute percentage error as small as possible, but the decision coefficient is as large as possible. In this embodiment, the compatibility_models of the machine-learned pycaret module are used, and the best-fitting machine learning model is selected as the hood temperature prediction model.
(5) When the target smoke hood temperature at the current moment is predicted, the current moment is taken as a reference, and the raw ball belt scale flow, the cooling air flow and the average temperature of the combustion chamber at the first five sampling moments are input into a smoke hood temperature prediction model to obtain the target smoke hood temperature at the current moment.
PID control model
The PID control model comprises a combustion chamber temperature PID control model and a smoke hood temperature PID control model
3.1 The control flow of the combustion chamber temperature PID control model comprises the following steps:
(1) Taking the recommended gas flow obtained in the gas flow prediction model as an initial recommended gas flowMeasuring amountThe temperature control model is input to a combustion chamber temperature PID control model, and an upper limit value and a lower limit value of the combustion chamber temperature are set.
(2) Calculating the difference between the average value of the actual combustion chamber temperature and the target combustion chamber temperature in the previous T timeDifference valueMultiplying a scaling factor as an adjustment factor +.>According to adjustment coefficient->Adjusting recommended gas flow +.>To obtain a new recommended gas flow +.>The method comprises the steps of carrying out a first treatment on the surface of the The recommended gas flow is reduced when the actual temperature is higher, and the gas flow is improved when the actual temperature is lower. In this example, T is one minute. The combustion chamber target temperature is an average of the combustion chamber temperature upper limit value and the combustion chamber temperature lower limit value.
In this embodiment, the target temperature of the combustion chamber includes the target temperature of the east combustion chamber and the target temperature of the west combustion chamber, according to the difference value corresponding to the east combustion chamberThe recommended gas flow of the east combustion chamber is regulated according to the corresponding difference value of the west combustion chamberAnd adjusting the recommended gas flow of the Western combustion chamber.
(3) Will recommend the gas flowRecommended gas flow predicted at the previous time with gas flow prediction model +.>Fusion is carried out to obtain recommended gas flow at the current moment>The method comprises the steps of carrying out a first treatment on the surface of the At fusion +.>And->The weight of (2) is 0.5 respectively, and the gas flow is recommended。
(4) Setting initial recommended air flow rates for east and west combustors based on historical production data。
(5) By difference valueThe recommended air flow is adjusted as the final recommended air flow at the current moment +.>The method comprises the steps of carrying out a first treatment on the surface of the The recommended gas flow is improved when the actual temperature is higher, and the gas flow is reduced when the actual temperature is lower. When the variable frequency fan is arranged on the operation site and air flow is regulated, the frequency of the variable frequency fan can be regulated besides air valve regulation.
(6) The upper limit value and the lower limit value of the air-coal ratio are controlled, a fixed value is set according to historical experience, the air-coal ratio refers to air flow divided by gas flow, and the specific control mode is to adjust the opening of an air pipeline valve and a gas pipeline valve.
(7) The temperature wind control of the combustion chamber is divided into low Wen Fengkong and high temperature wind control, and the corresponding relation between different temperature levels and gas flow change values is preset. When the actual temperature of the combustion chamber is lower than a set low-temperature value, increasing the corresponding gas flow according to the temperature level of the actual temperature of the combustion chamber; when the actual temperature of the combustion chamber is higher than the set low temperature value, the corresponding gas flow is reduced according to the temperature level of the actual temperature of the combustion chamber.
(8) The air pipeline comprises an air main pipe and air branch pipes, wherein the east combustion chamber and the west combustion chamber correspond to one air branch pipe respectively; the air branch pipe flow is converted into an air branch pipe valve opening degree:
a) The calculation formula of the opening degree of the air branch pipe valve is as follows: actual air flow +a1 (recommended air flow rateIs +.>Difference) +b1 (recommended air flow +.>Is +.>Difference of) +c1 (recommended air flow +.>Is +.>Difference) +d1 (recommended air flow +.>Is +.>Is a difference of (c). The actual air flow is fed back by the opening degree of the valve of the air branch pipe, < >>And->、/>、/>The actual air flow rates at the current time and the first three sampling times, a1, b1, c1, d1, are coefficients, respectively.
b) Setting the opening of the air branch pipe valve to change the upper limit and the lower limit each time, limiting the air branch pipe valve to change too fast each time, and preventing the swing frequency of the air branch pipe valve from being too high.
c) The maximum adjustment amplitude of the air branch valve is set, so that other conditions such as equipment damage and the like caused by excessive adjustment of the air branch valve are prevented.
d) And setting an upper limit value and a lower limit value of the air branch pipe valve, wherein the upper limit value and the lower limit value of the air branch pipe valve are set according to historical experience.
(9) The gas pipeline comprises a gas main pipe and a gas branch pipe; the east combustion chamber and the west combustion chamber are respectively corresponding to one gas branch pipe; the flow of the gas branch pipe is converted into the opening of the valve of the gas branch pipe:
a) The calculation formula of the opening of the gas branch pipe valve is as follows: actual gas flow +a2 (recommended gas flow)Is>Difference of (2) +b2 (recommended gas flow +.>Is>Difference of (c 2) + (recommended gas flow +.>Is>D2 (recommended gas flow +.>Is>Is a difference of (c). The actual gas flow is fed back by a gas branch pipe valve and is ∈10->And->、/>、/>The actual gas flow rates at the current time and the first three sampling times are respectively, and a2, b2, c2 and d2 are coefficients.
b) Setting the opening of the gas branch pipe valve to change the upper limit and the lower limit each time, limiting the too fast range of each time of changing the gas branch pipe valve, and preventing the too high swing frequency of the gas branch pipe valve.
c) The opening of the gas branch pipe valve is further adjusted according to the actual gas flow, and if the recommended gas flow and the actual gas flow are too large in difference, the gas branch pipe valve is further adjusted according to the gradient.
d) And the maximum adjustment amplitude of the gas branch pipe valve is set, so that other conditions such as equipment damage and the like caused by excessive adjustment of the gas branch pipe valve are prevented.
e) Setting an upper limit value and a lower limit value of a gas branch pipe valve, wherein the upper limit value and the lower limit value of the gas branch pipe valve are set according to historical experience.
(10) The air manifold valve opening takes the average of the air manifold valve openings over one minute.
(11) The flow of the gas main pipe is converted into the opening of a valve of the gas main pipe:
a) Coal gas main flow = east combustion chamber recommended coal gas flow + west combustion chamber recommended coal gas flow
b) And (8) converting the flow of the gas main pipe into the opening degree of the valve of the gas main pipe in the same step (8).
(12) Interaction of gas main pipe valve adjustment and gas branch pipe valve adjustment:
a) When the average temperature of the east combustion chamber and the west combustion chamber is higher or lower than the set target temperature of the combustion chamber, the opening degree of the gas main pipe valve is regulated, and the opening degree of the two gas branch pipe valves is kept unchanged.
b) When the average temperature of the east combustion chamber is higher than the target temperature of the combustion chamber and the average temperature of the west combustion chamber is lower than the target temperature of the combustion chamber, the opening of the gas main pipe valve is kept unchanged, the opening of the east combustion chamber gas branch pipe valve is reduced, and the opening of the west combustion chamber gas branch pipe valve is increased.
c) When the average temperature of the east combustion chamber is lower than the target temperature of the combustion chamber and the average temperature of the west combustion chamber is higher than the target temperature of the combustion chamber, the opening degree of the gas main pipe valve is kept unchanged, the opening degree of the gas branch pipe valve of the east combustion chamber is increased, and the opening degree of the gas branch pipe valve of the west combustion chamber is reduced.
3.2 control flow of the hood temperature PID control model, comprising the following steps:
(1) And leading in the target smoke hood temperature output by the smoke hood temperature prediction model.
(2) According to the difference between the average value of the actual hood temperature and the target hood temperature in the previous T timeFor the initial recommended cooling air flow set +.>Adjusting to obtain new recommended cooling air flow +.>The method comprises the steps of carrying out a first treatment on the surface of the I.e. the actual temperatureThe recommended cooling air flow is reduced when the flow is higher, and the recommended cooling air flow is increased when the flow is lower. In this example, T is one minute.
(3) Will recommend the cooling air flowRecommended cooling air flow rate +.A recommended cooling air flow rate obtained from a target hood temperature predicted at a previous time on a hood temperature prediction model>Fusion is carried out to obtain final recommended cooling air flow at the current moment>The method comprises the steps of carrying out a first treatment on the surface of the At fusion +.>Andweight of 0.5 each, recommended cooling air flow +.>。
(4) The temperature wind control of the fume hood is divided into low Wen Fengkong and high temperature wind control, and the corresponding relation between different temperature levels and cooling wind flow change values is preset. When the actual temperature of the fume hood is lower than a set low-temperature value, increasing the corresponding cooling air flow according to the temperature level of the actual temperature of the fume hood; when the actual temperature of the fume hood is higher than the set low-temperature value, the corresponding cooling air flow is reduced according to the temperature level of the actual temperature of the fume hood.
(5) Converting the recommended cooling air flow into a cooling air pipeline valve opening degree:
a) The calculation formula of the opening of the cooling air pipeline valve is as follows: actual cooling air flow +a3 (recommended cooling air flowAnd the actual cooling air flow->Difference of (d) +b3 (recommended cooling air flow +.>And the actual cooling air flow->And c3 (recommended cooling air flow +.>And the actual cooling air flow->D3 (recommended cooling air flow +.>And the actual cooling air flow->Is a difference of (c). The actual cooling air flow is fed back by a cooling air pipeline valve to obtain +.>And->、、/>The actual cooling air flow rates at the current time and the first three sampling times are respectively represented by a3, b3, c3 and d3 as coefficients.
b) Setting the opening of the cooling air pipeline valve to change the upper limit and the lower limit each time, limiting the too fast range of each time of changing of the cooling air pipeline valve, and preventing the too high swinging frequency of the cooling air pipeline valve.
c) The opening of the cooling air pipeline valve is further adjusted according to the actual cooling air flow, and if the difference between the recommended cooling air flow and the actual cooling air flow is too large, the cooling air pipeline valve is further adjusted according to the gradient.
d) And the maximum adjustment amplitude of the cooling air pipeline valve is set, so that other conditions such as equipment damage and the like caused by excessive adjustment of the cooling air pipeline valve are prevented.
e) And setting an upper limit value and a lower limit value of the cooling air pipeline valve, wherein the upper limit value and the lower limit value of the cooling air pipeline valve are set according to historical experience.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.
Furthermore, it should be understood that although the present disclosure describes embodiments, not every embodiment is provided with a single embodiment, and that this description is provided for clarity only, and that the disclosure is not limited to specific embodiments, and that the embodiments may be combined appropriately to form other embodiments that will be understood by those skilled in the art.
Claims (9)
1. A shaft furnace production control method comprising the steps of:
step A1: training a gas flow prediction model by utilizing historical production data D1 generated when the shaft furnace is operated manually; the production data D1 comprises gas flow, air flow, combustion chamber temperature and raw ball belt scale flow of a combustion chamber;
step A2: inputting production data D1 at the current moment into a coal gas flow prediction model which is trained, and predicting to obtain recommended coal gas flow of a combustion chamber at the current moment;
step A3: the control of the temperature of the combustion chamber is carried out through a PID control model of the temperature of the combustion chamber, and specifically comprises the following steps:
step A31: taking the recommended gas flow obtained in the step A2 as the initial recommended gas flowInputting the temperature of the combustion chamber into a PID control model;
step A32: based on the difference between the average value of the actual combustion chamber temperature and the target combustion chamber temperature in the previous T timeFor the initial recommended gas flow +.>Adjusting to obtain new recommended gas flow +.>;
Step A33: will recommend the gas flowRecommended gas flow predicted at the previous time with gas flow prediction model +.>Fusion is carried out to obtain final recommended gas flow at the current moment>;
Step A34: setting an initial recommended airflow for the combustion chamberBy difference->The recommended air flow is adjusted as the final recommended air flow at the current moment +.>;
Step A35: setting the corresponding relation between different temperature levels and gas flow change values; when the actual temperature of the combustion chamber is lower than the set low temperature value, the gas flow is recommended according to the temperature level of the actual temperature of the combustion chamberIncreasing the corresponding gas flow on the basis of the above; when the actual temperature of the combustion chamber is higher than the set low temperature value, according to the temperature level of the actual temperature of the combustion chamber, the gas flow is recommended>On the basis of the above, reducing the corresponding gas flow;
step A36: recommended gas flow at the current momentConverting into the opening degree of a valve of a gas pipeline, and adding the recommended air flow rate at the current moment>The temperature of the combustion chamber is controlled by converting the temperature into the opening degree of the valve of the air pipeline.
2. The shaft furnace production control method according to claim 1, characterized in that: when training the gas flow prediction model in the step A1, the method comprises the following steps:
step A11: collecting historical production data D1 generated when the shaft furnace is operated manually;
step A12: the historical production data D1 is divided periodically with fixed duration to obtain historical production cycle data D1', and the historical production cycle data D1' with the first 30% of the cycle with the minimum average value of the gas flow of the combustion chamber and the flow of the green belt scale is selected as training data;
step A13: inputting training data into different machine learning models, wherein the target value of the machine learning model is the gas flow at the current moment, and the variables are the raw ball belt scale flow and the gas flow at the first five sampling moments;
step A14: adopting cross verification contrast, and selecting a machine learning model with the best fitting as a coal gas flow prediction model;
and (A2) when the recommended gas flow of the combustion chamber at the current moment is predicted, taking the current moment as a reference, inputting the raw ball belt scale flow at the first five sampling moments and the gas flow at the first five sampling moments into a gas flow prediction model to obtain the recommended gas flow of the combustion chamber at the current moment.
3. The shaft furnace production control method according to claim 1, characterized in that: the combustion chamber target temperature is equal to an average of the set combustion chamber temperature upper limit value and the set combustion chamber temperature lower limit value.
4. The shaft furnace production control method according to claim 1, characterized in that: in step A32, according to the differenceFor the initial recommended gas flow +.>When the adjustment is performed, the difference value is->Multiplying a fixed scaling factor as an adjustment factor +.>By adjusting the coefficient->Recommended gas flow->And (5) adjusting.
5. The shaft furnace production control method according to claim 1, characterized in that: in step A33, the recommended gas flow rateRecommended gas flow predicted at the previous time with gas flow prediction model +.>Fusion is carried out to obtain recommended gas flow at the current moment>When (I)>。
6. The shaft furnace production control method according to claim 1, characterized in that: the method also comprises a fume hood temperature prediction control process at the top of the shaft furnace, and specifically comprises the following steps:
step B1: training a hood temperature prediction model using historical production data D2 generated when the shaft furnace is operated manually; the historical production data D2 comprises gas flow, air flow, combustion chamber temperature, raw ball belt scale flow and smoke hood temperature of the combustion chamber;
step B2: inputting historical production data D2 at the current moment into a trained hood temperature prediction model, and predicting to obtain a target hood temperature at the current moment;
the temperature of the fume hood is controlled by a fume hood temperature PID control model, and the method specifically comprises the following steps:
step B31: taking the target hood temperature obtained in the step B2 as the initial target hood temperatureInputting the temperature of the fume hood into a PID control model;
step B32: according to the difference between the average value of the actual hood temperature and the target hood temperature in the previous T timeFor the initial recommended cooling air flow set +.>Adjusting to obtain new recommended cooling air flow +.>;
Step B33: will recommend the cooling air flowRecommended cooling air flow rate +.A recommended cooling air flow rate obtained from a target hood temperature predicted at a previous time on a hood temperature prediction model>Fusion is carried out to obtain final recommended cooling air flow at the current moment>;
Step B34: setting the corresponding relation between different temperature levels and cooling air flow change values; when the actual temperature of the fume hood is lower than the set low temperature value, recommending cooling air flow according to the temperature level of the actual temperature of the fume hoodOn the basis of the above, increasing the corresponding cooling air flow; when the actual temperature of the fume hood is higher than the set low temperature value, the flow of the cooling air is recommended according to the temperature level of the actual temperature of the fume hood>On the basis of the above, the corresponding cooling air flow is reduced;
7. The shaft furnace production control method according to claim 6, characterized in that: when the smoke hood temperature prediction model is trained in the step B1, the method comprises the following steps of:
step B11: collecting historical production data D2 generated when the shaft furnace is operated manually;
step B12: the historical production data D2 is divided periodically with fixed duration to obtain historical production cycle data D2', and the historical production cycle data D2' with the first 30% of the cycle with the smallest average value of the gas flow of the combustion chamber and the flow of the green belt scale is selected as training data;
step B13: inputting training data into different machine learning models, wherein the target value of the machine learning model is the temperature of a smoke hood at the current moment, and the variables are the flow of the green-ball belt scale, the flow of cooling air and the average temperature of a combustion chamber at the first five sampling moments;
step B14: adopting cross verification contrast, and selecting a machine learning model with the best fitting as a smoke hood temperature prediction model;
and B2, when the target smoke hood temperature at the current moment is predicted, taking the current moment as a reference, inputting the flow of the green-ball belt scale, the flow of cooling air and the average temperature of the combustion chamber at the first five sampling moments into a smoke hood temperature prediction model to obtain the target smoke hood temperature at the current moment.
8. The shaft furnace production control method according to claim 6, characterized in that: in step B32, according to the differenceFor the initial recommended cooling air flow +.>When the adjustment is performed, the difference value is->Multiplying a fixed scaling factor as an adjustment factor +.>By adjusting the coefficient->Recommended cooling air flow->And (5) adjusting.
9. The shaft furnace production control method according to claim 6, characterized in that: in step B33, the flow rate of the cooling air is recommendedRecommended cooling air flow rate +.A recommended cooling air flow rate obtained from a target hood temperature predicted at a previous time on a hood temperature prediction model>Fusion is carried out to obtain recommended cooling air flow at the current moment>When (I)>。/>
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310137381.7A CN115930594B (en) | 2023-02-20 | 2023-02-20 | Shaft furnace production control method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310137381.7A CN115930594B (en) | 2023-02-20 | 2023-02-20 | Shaft furnace production control method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN115930594A CN115930594A (en) | 2023-04-07 |
CN115930594B true CN115930594B (en) | 2023-05-30 |
Family
ID=85820280
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310137381.7A Active CN115930594B (en) | 2023-02-20 | 2023-02-20 | Shaft furnace production control method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115930594B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116300666A (en) * | 2023-05-24 | 2023-06-23 | 科大智能物联技术股份有限公司 | Power plant boiler operation control method based on XGBoost optimization |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH10226809A (en) * | 1996-12-11 | 1998-08-25 | Nkk Corp | Method for controlling hot stove |
CN102692124A (en) * | 2012-05-24 | 2012-09-26 | 北京首钢自动化信息技术有限公司 | Automatic control method for improving uniformity of temperature of sleeve kiln |
CN103570258A (en) * | 2012-07-30 | 2014-02-12 | 西安众智惠泽光电科技有限公司 | Production process automatic control system of gas burning lime shaft kiln |
CN105159235A (en) * | 2015-01-08 | 2015-12-16 | 北方工业大学 | Comprehensive coordination control method and system for rotary kiln during calcining process |
CN109492335A (en) * | 2018-12-12 | 2019-03-19 | 中国地质大学(武汉) | A kind of annealing furnace furnace temperature prediction technique and system |
CN114622048A (en) * | 2022-03-15 | 2022-06-14 | 恒创数字科技(江苏)有限公司 | Combustion optimization system and method for hot blast stove |
CN115186582A (en) * | 2022-07-05 | 2022-10-14 | 科大智能物联技术股份有限公司 | Steel rolling heating furnace control method based on machine learning model |
-
2023
- 2023-02-20 CN CN202310137381.7A patent/CN115930594B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH10226809A (en) * | 1996-12-11 | 1998-08-25 | Nkk Corp | Method for controlling hot stove |
CN102692124A (en) * | 2012-05-24 | 2012-09-26 | 北京首钢自动化信息技术有限公司 | Automatic control method for improving uniformity of temperature of sleeve kiln |
CN103570258A (en) * | 2012-07-30 | 2014-02-12 | 西安众智惠泽光电科技有限公司 | Production process automatic control system of gas burning lime shaft kiln |
CN105159235A (en) * | 2015-01-08 | 2015-12-16 | 北方工业大学 | Comprehensive coordination control method and system for rotary kiln during calcining process |
CN109492335A (en) * | 2018-12-12 | 2019-03-19 | 中国地质大学(武汉) | A kind of annealing furnace furnace temperature prediction technique and system |
CN114622048A (en) * | 2022-03-15 | 2022-06-14 | 恒创数字科技(江苏)有限公司 | Combustion optimization system and method for hot blast stove |
CN115186582A (en) * | 2022-07-05 | 2022-10-14 | 科大智能物联技术股份有限公司 | Steel rolling heating furnace control method based on machine learning model |
Also Published As
Publication number | Publication date |
---|---|
CN115930594A (en) | 2023-04-07 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11788012B2 (en) | Integrated coke plant automation and optimization using advanced control and optimization techniques | |
CN115930594B (en) | Shaft furnace production control method | |
US11060032B2 (en) | Integrated coke plant automation and optimization using advanced control and optimization techniques | |
CN101892338B (en) | Constant wind temperature control system for hot blast stove | |
CN106839746B (en) | A kind of bf coal injection system combustion gas furnace, which automatically optimizes, burns furnace control method | |
KR101956365B1 (en) | System for control temperature pattern of strip in continuous annealing line and the method of the same | |
CN106011352A (en) | Optimized control system for coal gas flow of blast-furnace hot blast stove | |
CN110205427A (en) | A kind of intelligence hot-blast stove Optimal Control System and method | |
CN110699502B (en) | Method for high-precision prediction of gas consumption of blast furnace hot blast stove | |
JP2018103067A (en) | Processing unit, method and program in pulverization plant | |
CN109086949B (en) | Blast furnace gas generation amount and heat value prediction method based on gas component change | |
CN103616815A (en) | Control method for waste plastic oil refining cracking furnace chamber temperature based on dynamic matrix control optimization | |
CN109710976B (en) | Method for predicting gas consumption of steel rolling heating furnace | |
JP5286729B2 (en) | Combustion control method and combustion control apparatus for hot stove | |
CN102297451B (en) | Method for regulating and controlling air-flowing environment of heating furnace | |
CN106011353B (en) | A kind of blast funnace hot blast stove air-fuel ratio self-optimization method | |
CN204369760U (en) | A kind of annular lime kiln firing system | |
CN104061586A (en) | Controlling method and controlling device for combustion air flow | |
JP2009084636A (en) | Method and apparatus for controlling combustion in hot blast stove | |
JPH10226809A (en) | Method for controlling hot stove | |
CN114721253A (en) | Heating furnace temperature fractional order PID control system and method based on artificial bee colony algorithm | |
CN109141015A (en) | Double-hearth lime kiln two close cycles temperature control equipment and its application method | |
CN114967793B (en) | Hot-blast stove energy-saving control method and system based on exhaust gas temperature rise curve | |
CN104266500A (en) | Cooler heat source parameter forecasting and regulation method based on heat condition in sintering process | |
CN207659471U (en) | Blast-furnace hot-air wind-warm syndrome monitoring device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |