CN114115380B - Temperature control method and system for 3D glass hot bending die - Google Patents
Temperature control method and system for 3D glass hot bending die Download PDFInfo
- Publication number
- CN114115380B CN114115380B CN202111408751.3A CN202111408751A CN114115380B CN 114115380 B CN114115380 B CN 114115380B CN 202111408751 A CN202111408751 A CN 202111408751A CN 114115380 B CN114115380 B CN 114115380B
- Authority
- CN
- China
- Prior art keywords
- temperature
- measurement point
- ith
- time
- instantaneous
- 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
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D23/00—Control of temperature
- G05D23/19—Control of temperature characterised by the use of electric means
- G05D23/20—Control of temperature characterised by the use of electric means with sensing elements having variation of electric or magnetic properties with change of temperature
-
- C—CHEMISTRY; METALLURGY
- C03—GLASS; MINERAL OR SLAG WOOL
- C03B—MANUFACTURE, SHAPING, OR SUPPLEMENTARY PROCESSES
- C03B23/00—Re-forming shaped glass
- C03B23/02—Re-forming glass sheets
- C03B23/023—Re-forming glass sheets by bending
- C03B23/03—Re-forming glass sheets by bending by press-bending between shaping moulds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Abstract
A3D glass hot bending die temperature control method and system, the method comprising: acquiring a temperature rise curve of each temperature measurement point of the 3D glass hot bending die, carrying out segmentation and segmented discrete control to obtain theoretical temperature SUi (tn) of the ith temperature measurement point at the future time of a time region [ tm, tn ], and detecting instantaneous temperature Vui (t); establishing an LSTM prediction model, inputting a plurality of instantaneous feature vectors into the trained LSTM prediction model, and obtaining the predicted temperature WUI (tn) of the ith temperature measurement point at the future moment of a time zone [ tm, tn ]; calculating the difference between the theoretical temperature SUi (tn) and the instantaneous temperature Vui (t) to obtain an instantaneous temperature difference value, and calculating to obtain a main control quantity corresponding to the ith temperature measurement point by utilizing a PID control algorithm; and calculating the power adjustment rate corresponding to the ith temperature measurement point by adopting a fuzzy control method, and finely adjusting the main control quantity by utilizing the power adjustment rate. The invention realizes the temperature cooperative control in time and space of the die in the die temperature control process, and improves the temperature regulation precision.
Description
Technical Field
The invention belongs to the technical field of molds, and particularly relates to a temperature control method and system for a 3D glass hot bending mold.
Background
With the development of electronic information technology, the application of transparent glass components in the industry is continuously increased, and particularly with the rapid advance of 5G technology, there is a great industrial demand for the wide application of wireless charging technology and flexible OLEDs. This requires that the terminal cover plate glass (e.g., front and back cover plates of a mobile phone, front cover plate of a smart watch, etc.) is designed to be a curved surface shape, also called 3D cover plate glass. From the current market application, 3D cover glass is well evaluated, and its demand is increasing. Compared with the traditional processing technology, the hot bending forming technology has the advantages of low cost and batch production, and is particularly suitable for the production of 3D cover plate glass. However, since the material such as glass is amorphous, if the temperature of the thermal bending is too low, the 3D glass member is easily broken or the size is not satisfactory; on the contrary, if the temperature of the hot bending is too high, the 3D glass component is easy to generate defects such as scalding, water ripples and the like; in addition, the 3D glass component is in a curved surface shape, so that heat transfer of the mold is uneven, and the hot bending yield of the 3D glass product is further reduced.
Generally speaking, the performance of the hot bending mold material requires that the material should have the characteristics of fine crystal grains, compact and uniform structure, high thermal stability, easy processing, good thermal conductivity, small thermal expansibility and the like. Alternative materials include: alloys, ceramics, graphite, and the like. Because of the excellent heat transfer performance and the precise processing characteristic of the graphite, the graphite better meets the requirements of a 3D cover plate glass hot bending forming die. At present, graphite is mostly used as a raw material of a hot bending die in the industry, and in order to improve the technological performance of products, graphite dies are generally used in pairs, namely, concave-convex dies are matched for use.
For the 3D glass hot bending forming process, the process parameters (temperature and pressure) have very important influence on the forming quality, particularly the temperature parameter. The traditional upper and lower die temperature control adopts a single PID (proportion integration differentiation) and other control methods, the deviation between the temperature at the current time and the expected temperature is based on, the temperature of the whole upper die or the whole lower die is regulated, and the accuracy of the temperature control is difficult to meet the application of actual conditions due to the fact that the die space structure is complex and a heating system is composed of a plurality of heating rods, so that the yield of 3D glass hot bending is reduced.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a temperature control method and a temperature control system for a 3D glass hot bending die, solves the problem of low precision in the die temperature control process, and improves the temperature regulation and control precision by realizing the temperature cooperative control in time and die space.
The invention firstly provides a temperature control method for a 3D glass hot bending die, which comprises the following steps:
step S1, obtaining a temperature rise curve of each temperature measurement point of the 3D glass hot bending mould, segmenting the temperature rise curve according to time intervals, sequentially carrying out segmented discrete control on each obtained time region, regarding the time region [ tm, tn ] and the tn time as the future time, obtaining theoretical temperature SUi (tn) of the ith temperature measurement point at the future time of the time region [ tm, tn ], and detecting instant temperature Vui (t) corresponding to the heated ith temperature measurement point at the time region [ tm, tn ];
step S2, establishing an LSTM prediction model of the mold temperature, collecting an instantaneous feature vector of the ith temperature measurement point of the 3D glass hot bending mold in a time region [ tm, tn ], inputting the instantaneous feature vector into the trained LSTM prediction model, and obtaining the predicted temperature Wui (tn) of the ith temperature measurement point at the future moment of the time region [ tm, tn ];
step S3, calculating the difference between the theoretical temperature SUi (tn) and the instantaneous temperature Vui (t) to obtain an instantaneous temperature difference value, and calculating to obtain a main control quantity corresponding to the ith temperature measurement point by using a PID control algorithm;
step S4, the predicted temperature Wui (tn) and the theoretical temperature SUi (tn) of the ith temperature measuring point at the future time tn are used as input, a fuzzy control method is adopted to calculate the power adjustment rate corresponding to the ith temperature measuring point, and the power adjustment rate is used for fine adjustment of the main control quantity.
Further, the temperature rise curve and the LSTM prediction model are obtained through simulation experiments and/or processing experiments.
Further, the 3D glass hot bending mould comprises an upper mould and a lower mould which are matched with each other.
Further, in step S2, the instantaneous feature vector includes a current position temperature, a current position first neighboring point temperature, a current position second neighboring point temperature, a current position third neighboring point temperature, a hot bender internal cavity temperature, a hot bender external temperature, a mold working time, and a mold type.
Further, in step S3, the calculation formula of the instantaneous temperature difference value eui (t) of the ith temperature measurement point at the future time of the time zone [ tm, tn ] is:
eUi(t)＝SUi(tn)-VUi(t)。
further, in step S3, the relationship between the ith temperature measurement point input eui (t) and the output xui (t) is:
in the formula, eUi(t)、
Respectively, an error integral and an error differential term, k, corresponding to the ith temperature measurement pointiUP,kiUI,kiUDRespectively, a proportionality coefficient, an integral coefficient and a differential coefficient corresponding to the ith temperature measurement point.
Further, in step S4, when SUi (tn)) epsilon (0 ℃ -300 ℃ C.) of the i-th temperature measurement point at the future time tn and ABS (SUi (tn) -WUi (tn)) ≦ a ℃ C., the power adjustment rate is (100+ a)%, if SUi (tn) is greater than WUi (tn), and (100-a)%, if SUi (tn) is less than WUi (tn).
Further, after step S4, the method further includes training the LSTM prediction model in real time, and replacing the current LSTM prediction model when the training of the LSTM prediction model is completed.
The invention also provides a system of the temperature control method for the 3D glass hot bending die, which comprises the following steps:
the theoretical model module is used for obtaining a temperature rise curve of each temperature measurement point of the 3D glass hot bending mould, segmenting the temperature rise curve according to time intervals, sequentially carrying out segmented discrete control on each obtained time region, and obtaining theoretical temperature SUi (tn) of the ith temperature measurement point at the future time of the time region [ tm, tn ] for the time region [ tm, tn ] with the time tn as the future time;
the instantaneous detection module is used for detecting the instantaneous temperature Vui (t) corresponding to the heated ith temperature measurement point in the time region [ tm, tn ];
the prediction model module is used for establishing an LSTM prediction model of the mold temperature, acquiring an instantaneous feature vector of the ith temperature measurement point of the 3D glass hot bending mold in a time region [ tm, tn ], inputting the instantaneous feature vector into the trained LSTM prediction model, and obtaining the predicted temperature Wui (tn) of the ith temperature measurement point at the future moment of the time region [ tm, tn ];
the main control module is used for calculating the difference between the theoretical temperature SUi (tn) and the instantaneous temperature Vui (t) to obtain an instantaneous temperature difference value, and calculating to obtain a main control quantity corresponding to the ith temperature measuring point by utilizing a PID control algorithm;
and the auxiliary control module is used for taking the predicted temperature Wui (tn) of the future time tn of the ith temperature measurement point and the theoretical temperature SUi (tn) as input, calculating the power adjustment rate corresponding to the ith temperature measurement point by adopting a fuzzy control method, and finely adjusting the main control quantity by utilizing the power adjustment rate.
Furthermore, the main control module comprises a PID controller and a heating device, the auxiliary control module comprises a fuzzy controller, and the output ends of the PID controller and the fuzzy controller are respectively connected with the input end of the heating device.
The temperature control method and the temperature control system for the 3D glass hot bending die provided by the invention have the following beneficial effects:
1) the problem that the control method only aims at the temperature deviation at the current moment in the mold temperature control process is low in precision is solved, a cooperative control method of the mold temperature in time integrating the temperature deviation at the current moment and the temperature deviation at the future moment is provided, and the temperature regulation and control precision is improved;
2) because the space structure of the die is complex, the temperature collected by a single temperature sensor is less than the conduction rule of heat in the heating process of the die, and the data of a plurality of adjacent temperature sensors are collected in the space, so that accurate data are provided for predicting the temperature value at the future moment;
3) the plurality of independent temperature regulation and control modules are arranged in the die, so that different areas can be independently controlled and the temperature can be regulated and controlled, and the precise temperature control, namely the spatial cooperative control, can be realized for the specific area to be thermally bent of the 3C glass component by fusing PID control and fuzzy control according to the thermal bending process requirement of the 3C glass component.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
FIG. 1 is a schematic flow chart of a 3D glass hot bending mold temperature control method according to an embodiment of the invention;
FIG. 2 is a schematic connection diagram of a temperature control system for a 3D glass hot bending mold according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of the distribution of mold temperature measurement points for an embodiment of the present invention;
FIG. 4 is a sample vector diagram of a prediction model module according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art based on the embodiments of the present invention without inventive step, are within the scope of protection of the present invention.
In addition, the following description of the various embodiments refers to the accompanying drawings that illustrate specific embodiments in which the invention may be practiced. Directional phrases used in this disclosure, such as, for example, "upper," "lower," "front," "rear," "left," "right," "inner," "outer," "side," and the like, refer only to the orientation of the appended drawings and are, therefore, used herein for better and clearer illustration and understanding of the invention, and do not indicate or imply that the device or element so referred to must have a particular orientation, be constructed and operated in a particular orientation, and are therefore not to be considered limiting of the invention.
The embodiment of the invention discloses a temperature control method for a 3D glass hot bending die, which comprises the following steps with reference to the attached figure 1:
step S110, obtaining a temperature rise curve of each temperature measurement point of the 3D glass hot bending mould, segmenting the temperature rise curve according to time intervals, sequentially carrying out segmented discrete control on each obtained time region, regarding the time region [ tm, tn ], and regarding the tn moment as the future moment, obtaining the theoretical temperature SUi (tn) of the ith temperature measurement point at the future moment of the time region [ tm, tn ];
step S120, detecting the instantaneous temperature Vui (t) of the heated ith temperature measuring point corresponding to the time zone [ tm, tn ];
step S130, establishing an LSTM prediction model of the mold temperature, collecting an instantaneous characteristic vector of the ith temperature measurement point of the 3D glass hot bending mold in a time region [ tm, tn ], inputting the instantaneous characteristic vector into the trained LSTM prediction model, and obtaining the predicted temperature Wui (tn) of the ith temperature measurement point at the future time of the time region [ tm, tn ];
step S140, calculating the difference between the theoretical temperature SUi (tn) and the instantaneous temperature Vui (t) to obtain an instantaneous temperature difference value, and calculating to obtain a main control quantity corresponding to the ith temperature measurement point by utilizing a PID control algorithm;
step S150, the predicted temperature Wui (tn) and the theoretical temperature SUi (tn) of the ith temperature measuring point at the future time tn are used as input, a fuzzy control method is adopted to calculate the power adjustment rate corresponding to the ith temperature measuring point, and the power adjustment rate is utilized to finely adjust the main control quantity.
According to the temperature control method of the 3D glass hot bending mold, the difference is made between the predicted temperature of the ith temperature measuring point of the mold at the next future moment of a time region [ tm, tn ] and the theoretical temperature, then the obtained temperature deviation signal is input into a mold sub-controller and used as a power compensation signal of a mold heating wire to adjust the power of a heating rod in real time, and therefore the main controller of the mold is finely adjusted; therefore, the control precision of the system is improved, and the dynamic performance of the temperature control process of the system is also improved.
The temperature rise curve and the LSTM prediction model are obtained through simulation experiments and/or processing experiments. In step S110, after a simulation experiment and a machining experiment, an optimal temperature rise curve of the mold is determined, and the temperature rise curve is segmented at intervals of 1 to 3 seconds. Taking the first measurement point of the mold as an example, the temperature rise curve after cutting is recorded as { SU1t1,SU1t2,…,SU1tnFor each time zone in turn (e.g. [ t1, t2 ]],…,[tn-1,tn]) Implementing piecewise discrete control for [ ti, tj]For temperature control in the time period of (d), the time tj is the future time corresponding to the temperature of the optimal mold temperature rise curve (SU 1)tj) Is the theoretical temperature for the next future time instant of the time period.
In an embodiment of the present invention, the 3D glass hot bending mold includes an upper mold 100 and a lower mold 200 that are fitted to each other.
In step S130, the instantaneous feature vector includes a current position temperature, a current position first neighboring point temperature, a current position second neighboring point temperature, a current position third neighboring point temperature, a thermal bender inner cavity temperature, a thermal bender outer temperature, a mold operating time, and a mold type. The upper die LSTM prediction module collects multiple groups of characteristic vectors, wherein the multiple groups of characteristic vectors comprise the temperature of the current position of an upper die, the temperature of a first adjacent point of the current position of the upper die, the temperature of a second adjacent point of the current position of the upper die, the temperature of a third adjacent point of the current position of the upper die, the temperature of an inner cavity of a hot bending machine, the temperature of the outer part of the hot bending machine, the working time of the die and the type of the die. As shown in fig. 3, for the UD1 temperature measurement point of the upper die 100, the upper die current position first adjacent point, the upper die current position second adjacent point and the upper die current position third adjacent point are an upper die UD2 temperature measurement point, a lower die DD1 temperature measurement point and a lower die DD2 temperature measurement point, respectively; for the UD2 temperature measurement point of the upper die 100, the first adjacent point of the current position of the upper die, the second adjacent point of the current position of the upper die and the third adjacent point of the current position of the upper die are respectively an UD1 temperature measurement point of the upper die, a DD3 temperature measurement point of the upper die and a DD2 temperature measurement point of the lower die. Inputting the instantaneous feature vector collected from the [ tm, tn ] time zone of the ith temperature measurement point into the trained upper module LSTM prediction module to obtain the temperature output of the measurement point at the time tn in the future, and marking the temperature output as Wui (tn).
In step S140, the calculation formula of the instantaneous temperature difference value eui (t) of the ith temperature measurement point at the future time of the time zone [ tm, tn ] is:
eUi(t)＝SUi(tn)-VUi(t)。
for the upper die, after the eui (t) is input through upper die PID control, the main module control quantity XUi (t) of the ith temperature measurement point of the upper die is obtained.
In step S140, taking the mold as an example, the main control quantity xui (t) for the ith temperature measurement point is controlled by the PID controller and calculated by using the PID control algorithm. The PID controller consists of a proportional unit (P), an integral unit (I) and a differential unit (D), and the relation between the input eUi (t) and the output XUi (t) of the ith temperature measuring point is as follows:
in the formula, eUi(t)、
Respectively, the error integral and the error differential term, k, of the corresponding first temperature measuring pointiUP,kiUI,kiUDProportional coefficient, integral coefficient and differential coefficient corresponding to the ith temperature measurement point are respectively.
For the following mold as an example, the control quantity XDi (t) of the ith temperature measurement point is also controlled by a PID controller and calculated by adopting a PID control algorithm: the PID controller consists of a proportional unit (P), an integral unit (I) and a differential unit (D), and the relation between the ith temperature measuring point input eDi (t) and the output XDi (t) is as follows:
in the formula, eDi(t)，
Respectively an error, an error integral and an error differential term; k is a radical of formulaiDP,kiDI,kiDDThe proportional coefficient, the integral coefficient and the differential coefficient of the ith temperature measuring point of the lower die are respectively.
In step S150, the predicted temperature wui (tn) at the time tn in the future of the ith temperature measurement point and the desired temperature sui (tn) at the time tn in the future are input, the power adjustment rate of the upper mold heating rod corresponding to the temperature measurement point is controlled by a fuzzy control method, and the heat generation amount of the upper mold heating rod is controlled after merging with the upper mold main control amount xui (t). The control method of the lower die heating rod is similar and will not be described again.
In step S150, when SUi (tn) epsilon (0 ℃ -300 ℃) is expected at the ith temperature measurement point at the future time tn and ABS (SUi (tn) -WUi (tn) ≦ a ℃, the power adjustment rate is (100+ a)%, if SUi (tn) is greater than WUi (tn), and (100-a)%, if SUi (tn) is less than WUi (tn).
Wherein, for the upper mold, the theoretical temperature and the predicted temperature at the time tn in the future are used as input, and the power adjustment rate (increase or decrease) of the upper mold heating rod is used as output, as shown in table 1; taking one example of the above, when the expected temperature sui (tn) epsilon (0-300 ℃) at the time tn in the future and ABS (sui (tn) -wui (tn)) ≦ 1 ℃, if sui (tn) is greater than wui (tn), the power adjustment rate of the upper mold heating rod is increased by 1%, which is 101% of the standard power (preferably 800W/heating rod), and if sui (tn) is less than wui (tn), the power adjustment rate of the upper mold heating rod is decreased by 1%, which is 99% of the standard power (preferably 800W/heating rod).
Fuzzy control method for die power adjustment rate in table 1
For the lower die, the expected temperature and predicted temperature at time tn in the future are taken as inputs, and the power adjustment rate (increase or decrease) of the upper die heating rod is taken as an output, as shown in table 2; taking one example of the above, when the expected temperature SDi (tn) epsilon (0-300 ℃) at the time tn in the future and ABS (SDi (tn) -WDi (tn)) ≦ 3 ℃, if SDi (tn) is greater than WDi (tn), the power adjustment rate of the upper mold heating rod is increased by 3%, which is 103% of the standard power (preferably 800W/heating rod), and if SDi (tn) is less than Wui (tn), the power adjustment rate of the upper mold heating rod is decreased by 3%, which is 97% of the standard power (preferably 800W/heating rod).
TABLE 2 fuzzy control method for die power regulation rate
Further, after step S150, the method further includes training the LSTM prediction model in real time, and replacing the current LSTM prediction model after the LSTM prediction model training is completed.
The training of the upper model LSTM prediction model comprises a pre-training part and a post-training part, samples of the pre-training part and the post-training part are shown in FIG. 4 and comprise duration, characteristics and batches; the time length takes a sampling period of 40-100ms as an interval, discretization processing is carried out on the whole heating time, and quantitative data on a time scale are obtained; the characteristics comprise an upper die characteristic and a lower die characteristic, wherein the upper die characteristic comprises an upper die current position temperature, an upper die current position first adjacent point temperature, an upper die current position second adjacent point temperature, an upper die current position third adjacent point temperature, a hot bender inner cavity temperature, a hot bender outer temperature, a die working time and a die type; the batches were 32 batches, with 16 pre-training batches, and 16 post-training batches. The pre-training data of the upper model LSTM prediction model is from simulation or experimental data; when the post-training data batch of the upper mould LSTM prediction model is less than 16 batches, the feature vector at the current moment is randomly extracted from the pre-training 16 batches of data to replace, so that the total batch is 32 batches, and the upper mould LSTM prediction model can be normally trained. The post-training 16 batches of data of the upper model LSTM prediction model are divided into 3 updating areas which are respectively an optimal prediction area, a worst prediction area and a latest prediction area, the optimal prediction area is 6 batches of data, the worst prediction area is 6 batches of data, and the latest prediction area is 4 batches of data. And after one round of 3D glass hot bending forming is finished, the upper mold LSTM prediction model evaluates the prediction temperature of each discrete time point, calculates the prediction error of the temperature, and stores the prediction error in the upper mold LSTM prediction model. The optimal prediction area is a batch with the best prediction error at the current time after the past hot bending forming, the worst prediction area is a batch with the worst prediction error at the current time after the past hot bending forming, and the latest prediction area is the last batch. The upper module LSTM prediction model is realized by an ARM processor and a related data storage, and sample data of pre-training and post-training and the prediction model thereof are stored. The upper mold LSTM prediction model trains the prediction model in real time in the background, and after the prediction model is trained, the current prediction model is replaced when the next round of 3D glass hot bending forming is finished. The lower LSTM prediction model is similar to the upper LSTM prediction model and is not described again.
By the above, the output of the control quantity of the upper die main loop and the output of the control quantity of the lower die main loop can be obtained, the conventional error of the upper die main loop and the conventional error of the lower die main loop are adjusted, meanwhile, the auxiliary loop is utilized, the temperature deviation of the upper die heating rod and the lower die heating rod at the future moment is adjusted, and therefore the problem that temperature regulation is not accurate due to the fact that the heat flow density is not uniform under the influence of the die space structure can be effectively solved.
In addition, in order to avoid system steady-state oscillation, the system adopts a control strategy of automatic switching of partitioned PID parameters, and in order to improve the sensitivity and precision of control, the optimal sampling period of the system is 40-100 ms. PID parameters of each sub-region are determined in advance by simulation or experiment and are stored in a control chip (preferably, the realization chip is a low-energy ARM processor) in the upper die main loop control quantity output and the lower die main loop.
The invention also provides a system of the temperature control method for the 3D glass hot-bending die, referring to fig. 2, including:
the theoretical model module 10 is configured to obtain a temperature rise curve of each temperature measurement point of the 3D glass hot bending mold, segment the temperature rise curve according to a time interval, sequentially perform segmented discrete control on each obtained time region, and obtain a theoretical temperature sui (tn) of an ith temperature measurement point at a future time of the time region [ tm, tn ] for the time region [ tm, tn ] with the time tn as the future time;
an instantaneous detection module 20, configured to detect an instantaneous temperature vui (t) corresponding to the heated ith temperature measurement point in the time region [ tm, tn ];
the prediction model module 30 is used for establishing an LSTM prediction model of the mold temperature, acquiring an instantaneous feature vector of the ith temperature measurement point of the 3D glass hot bending mold in a time region [ tm, tn ], inputting the instantaneous feature vector into the trained LSTM prediction model, and obtaining a predicted temperature Wui (tn) of the ith temperature measurement point at the future time of the time region [ tm, tn ];
the main control module 40 is configured to calculate a difference between the theoretical temperature sui (tn) and the instantaneous temperature vui (t) to obtain an instantaneous temperature difference value, and calculate, by using a PID control algorithm, to obtain a main control quantity corresponding to the ith temperature measurement point;
and the auxiliary control module 50 is configured to use the predicted temperature wui (tn) at the future time tn of the ith temperature measurement point and the theoretical temperature sui (tn) as inputs, calculate a power adjustment rate corresponding to the ith temperature measurement point by using a fuzzy control method, and finely adjust the main control quantity by using the power adjustment rate, so as to control the power of the mold heating rod 60.
In practical application of the temperature control system for the 3D glass hot bending die provided by the embodiment of the invention, for small and medium 3C glass components (such as a 4-6 inch smart phone glass cover plate), the number of the preferable upper die heating rods (built-in resistance wires) is 4, the corresponding positions in the middle of the upper die are respectively provided with one temperature sensor, the number of the lower die heating rods (built-in resistance wires) is 5, and the corresponding positions in the middle of the lower die are also respectively provided with one temperature sensor.
The prediction model module comprises an upper module LSTM prediction module and a lower module LSTM prediction module. The upper mould LSTM prediction module has a temperature prediction function and predicts the temperature of the position where each sensor of the upper mould is arranged, and a plurality of groups (the preferred group number is 80-120) of feature vectors (the current time and the historical time) input by the upper mould LSTM prediction module comprise: the temperature of the current position of the upper die, the temperature of a first adjacent point of the current position of the upper die, the temperature of a second adjacent point of the current position of the upper die, the temperature of a third adjacent point of the current position of the upper die, the temperature of an inner cavity of a hot bending machine, the temperature of the outside of the hot bending machine, the working time of the die and the type of the die. The lower die LSTM prediction module also has a temperature prediction function and predicts the temperature of the position where each sensor of the lower die is arranged, and the lower die LSTM prediction module inputs a plurality of groups (the preferred group number is 80-120) of feature vectors (the current time and the historical time) including: the temperature of the current position of the lower die, the temperature of a first adjacent point of the current position of the lower die, the temperature of a second adjacent point of the current position of the lower die, the temperature of a third adjacent point of the current position of the lower die, the temperature of an inner cavity of a hot bending machine, the temperature of the outside of the hot bending machine, the working time of the die and the type of the die.
The main control module comprises a PID controller and a heating device, the auxiliary control module comprises a fuzzy controller, and the output ends of the PID controller and the fuzzy controller are respectively connected with the input end of the heating device. Specifically, it has last mould main control unit and last mould heating rod to go up mould main control module, it has last mould sub-control unit to go up mould sub-control module, will go up mould main control unit and last mould sub-control unit's output and be connected with the input of last mould heating rod respectively, and it needs to go up the mould heating rod and carry out independent control to last mould main control module very much. The lower mould main control module is provided with a lower mould main controller and a lower mould heating rod, the lower mould auxiliary control module is provided with a lower mould auxiliary controller, the output ends of the lower mould main controller and the lower mould auxiliary controller are respectively connected with the input end of the lower mould heating rod, and particularly, the lower mould main control module needs to independently control the lower mould heating rod. The upper die main controller and the lower die main controller are PID controllers; and the upper die sub-controller and the lower die sub-controller are fuzzy controllers.
The above description is not intended to limit the embodiments of the present invention, and the description of the above embodiments is intended to describe and illustrate the technical solutions of the present invention, and the above embodiments are merely illustrative and not restrictive. All technical equivalents and modifications which are obvious to those skilled in the art and which fall within the scope of the claims of the present invention are equally possible within the scope of the invention.
Claims (9)
1. A temperature control method for a 3D glass hot bending die is characterized by comprising the following steps:
step S1, obtaining a temperature rise curve of each temperature measurement point of the 3D glass hot bending mould, segmenting the temperature rise curve according to time intervals, sequentially carrying out segmented discrete control on each obtained time region, regarding the time region [ tm, tn ] and the tn time as the future time, obtaining theoretical temperature SUi (tn) of the ith temperature measurement point at the future time of the time region [ tm, tn ], and detecting instant temperature Vui (t) corresponding to the heated ith temperature measurement point at the time region [ tm, tn ];
step S2, establishing an LSTM prediction model of the mold temperature, collecting an instantaneous feature vector of the ith temperature measurement point of the 3D glass hot bending mold in a time region [ tm, tn ], wherein the instantaneous feature vector comprises the current position temperature, the current position first adjacent point temperature, the current position second adjacent point temperature, the current position third adjacent point temperature, the hot bending machine inner cavity temperature, the hot bending machine outer temperature, the mold working time and the mold type, and inputting the instantaneous feature vector into the trained LSTM prediction model to obtain the predicted temperature Wui (tn) of the ith temperature measurement point at the future moment of the time region [ tm, tn ];
step S3, calculating the difference between the theoretical temperature SUi (tn) and the instantaneous temperature Vui (t) to obtain an instantaneous temperature difference value, and calculating to obtain a main control quantity corresponding to the ith temperature measurement point by using a PID control algorithm;
step S4, using the predicted temperature wui (tn) and the theoretical temperature sui (tn) at the future time tn of the ith temperature measurement point as inputs, calculating the power adjustment rate corresponding to the ith temperature measurement point by using a fuzzy control method, and fine-tuning the main control quantity by using the power adjustment rate.
2. The 3D glass hot-bending mold temperature control method according to claim 1, wherein: the temperature rise curve and the LSTM prediction model are obtained through simulation experiments and/or processing experiments.
3. The method for controlling the temperature of the 3D glass hot-bending mold according to claim 1, wherein the 3D glass hot-bending mold comprises an upper mold and a lower mold which are matched with each other.
4. The method for controlling the temperature of a 3D glass hot-bending mold according to claim 1, wherein in step S3, the calculation formula of the instantaneous temperature difference value eUi (t) of the ith temperature measuring point at the future time of the time zone [ tm, tn ] is:
eUi(t)＝SUi(tn)-VUi(t)。
5. the method of claim 4, wherein in step S3, the i-th temperature measurement point input eUi (t) is related to the output XUi (t) by:
in the formula, eUi(t)、
Respectively, an error integral and an error differential term, k, corresponding to the ith temperature measurement pointiUP,kiUI,kiUDRespectively, a proportionality coefficient, an integral coefficient and a differential coefficient corresponding to the ith temperature measurement point.
6. The method for controlling the temperature of a 3D glass hot-bending mold according to claim 1, wherein in step S4, when SUi (tn) epsilon (0 ℃ -300 ℃) which is the expected temperature of the ith temperature measurement point at the future time tn and ABS (SUi (tn) -WUi (tn)) ≦ a ℃, if SUi (tn) is greater than WUi (tn), the power adjustment rate is (100+ a)%, and if SUi (tn) is less than WUi (tn), the power adjustment rate is (100-a)%.
7. The method for controlling the temperature of a 3D glass hot-bending mold according to claim 1, further comprising training the LSTM prediction model in real time after step S4, and replacing the current LSTM prediction model when the training of the LSTM prediction model is completed.
8. A system for applying the temperature control method for the 3D glass hot bending mold according to any one of claims 1 to 7, comprising:
the theoretical model module is used for obtaining a temperature rise curve of each temperature measurement point of the 3D glass hot bending mould, segmenting the temperature rise curve according to time intervals, sequentially carrying out segmented discrete control on each obtained time region, and obtaining theoretical temperature SUi (tn) of the ith temperature measurement point at the future time of the time region [ tm, tn ] for the time region [ tm, tn ] with the time tn as the future time;
the instantaneous detection module is used for detecting the instantaneous temperature Vui (t) corresponding to the heated ith temperature measurement point in the time region [ tm, tn ];
the prediction model module is used for establishing an LSTM prediction model of the mold temperature, acquiring an instantaneous feature vector of the ith temperature measurement point of the 3D glass hot bending mold in a time region [ tm, tn ], inputting the instantaneous feature vector into the trained LSTM prediction model, and obtaining the predicted temperature Wui (tn) of the ith temperature measurement point at the future moment of the time region [ tm, tn ];
the main control module is used for calculating the difference between the theoretical temperature SUi (tn) and the instantaneous temperature Vui (t) to obtain an instantaneous temperature difference value, and calculating to obtain a main control quantity corresponding to the ith temperature measuring point by utilizing a PID control algorithm;
and the auxiliary control module is used for taking the predicted temperature Wui (tn) and the theoretical temperature SUi (tn) of the ith temperature measuring point at the future time tn as input, calculating the power adjustment rate corresponding to the ith temperature measuring point by adopting a fuzzy control method, and finely adjusting the main control quantity by utilizing the power adjustment rate.
9. The 3D glass hot-bending mold temperature control system according to claim 8, wherein the main control module comprises a PID controller and a heating device, the secondary control module comprises a fuzzy controller, and output ends of the PID controller and the fuzzy controller are respectively connected with an input end of the heating device.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111408751.3A CN114115380B (en) | 2021-11-24 | 2021-11-24 | Temperature control method and system for 3D glass hot bending die |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111408751.3A CN114115380B (en) | 2021-11-24 | 2021-11-24 | Temperature control method and system for 3D glass hot bending die |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114115380A CN114115380A (en) | 2022-03-01 |
| CN114115380B true CN114115380B (en) | 2022-07-22 |
Family
ID=80372639
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111408751.3A Active CN114115380B (en) | 2021-11-24 | 2021-11-24 | Temperature control method and system for 3D glass hot bending die |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114115380B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115138309B (en) * | 2022-09-05 | 2022-12-06 | 济南明鑫制药股份有限公司 | Intelligent production monitoring method for crude product preparation in methyl ester production |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5475610A (en) * | 1990-11-29 | 1995-12-12 | The Perkin-Elmer Corporation | Thermal cycler for automatic performance of the polymerase chain reaction with close temperature control |
| EP3082010A1 (en) * | 2015-04-17 | 2016-10-19 | Fourdeg Oy | A system for dynamically balancing a heat load and a method thereof |
| EP3474112A2 (en) * | 2017-10-19 | 2019-04-24 | Robert Bosch GmbH | Method for determining an expected temperature line and method for determining a heating and/or cooling program of a heating and/or cooling system |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3000806B2 (en) * | 1992-10-12 | 2000-01-17 | トヨタ自動車株式会社 | Continuous differential pressure casting |
| JP3537215B2 (en) * | 1995-05-17 | 2004-06-14 | 株式会社東芝 | Heating furnace temperature controller |
| FR2782375B1 (en) * | 1998-08-13 | 2000-10-06 | Suisse Electronique Microtech | SYSTEM FOR REGULATING THE HEATING OF A BUILDING |
| US6855916B1 (en) * | 2003-12-10 | 2005-02-15 | Axcelis Technologies, Inc. | Wafer temperature trajectory control method for high temperature ramp rate applications using dynamic predictive thermal modeling |
| US7925389B2 (en) * | 2006-06-15 | 2011-04-12 | International Business Machines Corporation | Method of detecting recirculation of heated air within a rack enclosure |
| US20080099570A1 (en) * | 2006-10-04 | 2008-05-01 | Steve Krebs | System and method for estimating temperature drift and drive curves |
| CA2606067A1 (en) * | 2006-10-04 | 2008-04-04 | Telkonet Inc. | System and method for estimating temperature drift and drive curves |
| EP2866117B1 (en) * | 2013-10-25 | 2017-03-15 | Fourdeg Oy | Distributed adaptive and predictive heating control system and method |
| US11166344B2 (en) * | 2018-01-25 | 2021-11-02 | University Of Massachusetts | Electrically-heated fiber, fabric, or textile for heated apparel |
-
2021
- 2021-11-24 CN CN202111408751.3A patent/CN114115380B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5475610A (en) * | 1990-11-29 | 1995-12-12 | The Perkin-Elmer Corporation | Thermal cycler for automatic performance of the polymerase chain reaction with close temperature control |
| EP3082010A1 (en) * | 2015-04-17 | 2016-10-19 | Fourdeg Oy | A system for dynamically balancing a heat load and a method thereof |
| EP3474112A2 (en) * | 2017-10-19 | 2019-04-24 | Robert Bosch GmbH | Method for determining an expected temperature line and method for determining a heating and/or cooling program of a heating and/or cooling system |
Non-Patent Citations (1)
| Title |
|---|
| 基于模糊预测控制的燃料电池温度控制系统的研究;程珍等;《仪表技术》;20120815(第08期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114115380A (en) | 2022-03-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110027139B (en) | Mold preparation method, mold heating system and heating control method | |
| CN114115380B (en) | Temperature control method and system for 3D glass hot bending die | |
| CN101347822B (en) | Method for testing on-line temperature field of bloom continuous casting and method for controlling secondary cooling water | |
| CN101428301B (en) | Coiling temperature control device and control method | |
| CN105290380B (en) | A kind of method for designing of directional solidification blade running gate system internal baffle | |
| CN104959606A (en) | Partial temperature control system for metal material 3D printing | |
| CN106125790B (en) | Die face and the drop stamping temperature control system of plate Multi-Point Closed-Loop observing and controlling | |
| CN107052292B (en) | A kind of Hot Charging of Continuous Casting Slab tracking and calculating method calculated based on thermal physical property parameter distribution | |
| TW200806984A (en) | Rolling line material quality prediction and control apparatus | |
| CN114510092B (en) | Transition packet internal temperature control system and method based on fuzzy PID (proportion integration differentiation) of prediction variable universe | |
| CN107457378A (en) | The method that continuous casting steel billet three-dimensional dynamic water distribution water is set and controllable slighter compress is realized | |
| CN108710729A (en) | A kind of aluminum alloy heat forming technology formulating method based on organization and performance control | |
| CN108772539A (en) | A kind of casting mould cavity quickly heats and is quickly cooled down system | |
| CN110154303A (en) | A kind of temperature-controlling system and temperature control method based on heat pressing forming machines | |
| CN110687941A (en) | System and method for synchronously controlling temperatures of upper die and lower die of 3D cover plate glass | |
| CN109977442A (en) | A kind of method for numerical simulation of super thick slab multi- pass rolling technique | |
| CN104289532B (en) | Strip steel watermark point temperature-controlled process | |
| CN104294020A (en) | Heat stamping apparatus used for quenching control | |
| CN114178504B (en) | Intelligent temperature control method for low-pressure casting aluminum alloy melt | |
| CN115595667B (en) | Intelligent growth method, system, equipment and storage medium for tellurium-zinc-cadmium crystals | |
| CN208583955U (en) | A kind of casting mould type chamber quickly heats and is quickly cooled down system | |
| CN107399072B (en) | BOPP extruder die head bolt positioning method based on Thickness sensitivity | |
| CN107322905B (en) | A kind of film thickness monitoring method self-positioning based on bolt | |
| CN105695728A (en) | Devices and method of online steel plate solution treatment process control system | |
| CN109158569A (en) | Water-cooling type ductile iron pipe centrifuge and its control system |
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 |