CN109287021A - A kind of microwave heating temperature field intelligent control method based on on-line study - Google Patents
A kind of microwave heating temperature field intelligent control method based on on-line study Download PDFInfo
- Publication number
- CN109287021A CN109287021A CN201811197619.0A CN201811197619A CN109287021A CN 109287021 A CN109287021 A CN 109287021A CN 201811197619 A CN201811197619 A CN 201811197619A CN 109287021 A CN109287021 A CN 109287021A
- Authority
- CN
- China
- Prior art keywords
- microwave
- control strategy
- heating
- heating mode
- temperature
- 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.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H05—ELECTRIC TECHNIQUES NOT OTHERWISE PROVIDED FOR
- H05B—ELECTRIC HEATING; ELECTRIC LIGHT SOURCES NOT OTHERWISE PROVIDED FOR; CIRCUIT ARRANGEMENTS FOR ELECTRIC LIGHT SOURCES, IN GENERAL
- H05B6/00—Heating by electric, magnetic or electromagnetic fields
- H05B6/64—Heating using microwaves
- H05B6/66—Circuits
- H05B6/68—Circuits for monitoring or control
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Electromagnetism (AREA)
- Constitution Of High-Frequency Heating (AREA)
Abstract
A kind of microwave heating temperature field intelligent control method based on on-line study, using the dynamic associations in neural network model real-time learning part microwave heating process between heating mode and control strategy, and based on above-mentioned model according to the control strategy of the real-time predictive compensation Current Temperatures distribution of thought of heating mode complementation, accurate, intelligent compensation is carried out to non-uniform Temperature Distribution, realizes the accurate control to Part temperature uniformity in heating process.
Description
Technical field
The present invention relates to a kind of temperature field monitoring method, especially a kind of microwave heating temperature field monitoring method, specifically
Say it is a kind of microwave heating temperature field intelligent control method based on on-line study.
Background technique
Microwave is the electromagnetic wave that frequency is 300M to 300GHz.Microwave heating is material by absorbing microwave energy and by its turn
Thermal energy is changed into, to keep material whole while the heating method of heating.Since with high frequency characteristics, microwave electromagnetic field is with billions of
The surprising speed of secondary/second carries out cyclically-varying, and the polar molecule in material is (typical such as hydrone, protein, nucleic acid, rouge
Fat, carbohydrate etc.) under the action of electromagnetic field of high frequency polar movement is also done at a same speed, cause intermolecular frequently to touch
It hits and generates a large amount of frictional heats, temperature increases rapidly in a short time so as to cause material.Based on above-mentioned heating mechanism, microwave adds
Heat has the series of advantages such as heating speed is fast, part thickness direction temperature gradient is small, selectivity heats, is easily controllable, therefore
It is widely used in the various fields such as food processing, material processing, chemical synthesis.
However, there are the non-uniform problems in part same layer material temperature field for microwave heating technique.Its basic reason is
Electromagnetic field is distributed in standing wave state in microwave cavity.Near antinode, electric field or magnetic field strength are high, the vibration of inside parts polar molecule
Dynamic acutely heating is rapid, temperature is high, forms hot localised points；Near node, electric field or magnetic field strength are close to zero, in part
Polar molecule vibration in portion's is slight not to be vibrated even, and heating is slow, temperature is low, forms local cold spot.Non-uniform temperature distribution is serious
Threaten the Forming Quality of the safe and healthy and part processing of food processing.Existing method uses material rotary-tray and microwave mode
Blender etc. realizes microwave field and is heated the random relative motion between object to improve temperature uniformity.Material rotary-tray makes
It is material-to-be-heated to pass sequentially through the higher and lower region of microwave cavity internal electric field (or magnetic field) intensity, utilize in a period of time zero
Random cancellation effect on the same layer material of part between cold spot and hot spot improves temperature uniformity.Electromagnetic field mode blender is in cavity
A series of sheet metal of rotations is set at interior microwave feedback mouth, incident electromagnetic wave is dynamically dispersed to each region in cavity,
Improve the temperature uniformity of the same layer material of part using the random superposition effect of dynamic electromagnetic field in a period of time.But material revolves
The means such as disk, electromagnetic field mode blender that ask belong to the method for Temperature Distribution random back-off from principle, are inherently difficult to
Realize the accurate control being distributed to part same layer material temperature in microwave heating process.
Summary of the invention
The purpose of the present invention is heating the non-uniform problem in existing part same layer material temperature field for microwave current,
A kind of microwave heating temperature field intelligent control method based on on-line study is invented, microwave non-uniform heat flux is broken through from principle
Problem.
The technical scheme is that
A kind of microwave heating temperature field intelligent control method based on on-line study, it is characterised in that: use neural network
Dynamic associations in model real-time learning part microwave heating process between heating mode and control strategy, and it is based on above-mentioned mould
Type carries out uniform microwave to part according to the control strategy of the real-time predictive compensation Current Temperatures distribution of thought of heating mode complementation
Heating.
The heating mode control strategy prediction model for establishing part based on neural network algorithm refers to be added in microwave
In thermal process, using the Temperature Distribution of the same layer material of temperature sensor real-time monitoring part, for any time k (k >=p),
Using in heating mode-control strategy database apart from the preceding p group heating mode HP and control strategy U data that current time is nearest
It is right:
{(HPk-1,Uk-1),(HPk-2,Uk-2),…,(HPk-p,Uk-p)}
It exercises supervision training to above-mentioned prediction model；
As k < p, using preceding k-1 group heating mode HP and control strategy U data pair:
{(HP1,U1),(HP2,U2),…,(HPk-1,Uk-1)}
It exercises supervision training to above-mentioned prediction model；Forgetting Mechanism is used in the training process: in each training, distance
The contribution that current time remoter data update Model Weight is smaller, to learn the dynamic of microwave heating system more accurately
State feature；
After the completion of training, the thought based on heating mode complementation is quickly calculated for compensating Current Temperatures distribution Tk-1's
Target heating mode HP 'k, and be input in the prediction model for completing training, quick predict target heating mode HP 'kControl
Make strategy Uk；
In control strategy UkOn the basis of, power controller adjusts the function of each in running order microwave source in real time
Rate carries out uniform microwave heating to part according to scenario earthquake；
By control strategy UkAfter running the Δ t time, T is distributed based on Current TemperatureskQuickly calculate control strategy UkLower reality
Heating mode HPk, and by newest heating mode-control strategy data to (HPk,Uk) it is saved in heating mode-control strategy
In database；
It repeats the above process, until completing the entire microwave heating process of part.
The heating mode of the part is any control strategy UkHeating rate at the lower same layer material each point of part(c is constant).
The control strategy of the heating mode is the assembled state U=[δ of multiple microwave sources1,δ2,…,δl], wherein δ is represented
The switch state (value is 0 or 1) of microwave source, l represents the number of each microwave source on micro-wave oven.
The heating mode complementation thought applies biggish microwave to the lower region of part same layer material temperature
Power or heating rate apply lesser microwave power or heating rate to the higher region of part same layer material temperature.
Power controller calculates the intracorporal general power of microwave cavity based on pid algorithm according to the temperature curve of setting in real time, and
In control strategy UkOn the basis of power increment is averagely allocated to currently running microwave source.
The beneficial effects of the present invention are:
By the dynamic associations in on-line study Composite Microwave heating process between heating mode and control strategy,
It realizes and accurate, intelligent compensation is carried out to the uneven temperature distribution monitored in any part microwave heating process, from principle
The problem for breaching microwave non-uniform heat flux significantly improves the temperature uniformity for being heated object in microwave heating process.
Detailed description of the invention
Fig. 1 is the microwave heating temperature field intelligent monitoring flow chart based on on-line study.
Fig. 2 is model on-line study and the time series chart predicted in real time.
Specific embodiment
The present invention is further illustrated with reference to the accompanying drawings and examples.
As shown in Figs. 1-2.
The present embodiment uses chopped carbon fiber felt/epoxy resin composite material flat part (length 300mm, width
300mm, thickness 2mm) it is heating target, use the octagon High performance industrial micro-wave oven with No. 16 microwave sources for heating dress
It is standby.Using the Temperature Distribution of 30 channel fiber fluorescence temperature measurement systems monitoring composite material surface, and composite material surface is divided equally
For 10 (length direction m) × 6 (a thermometric regions width direction n).It is combined using different microwave sources as composite material parts and is added
The control strategy of heat pattern.Different microwave source combinations mainly include the letter such as different microwave source quantity or different microwave source distributing positions
Breath, can be described as following formula:
U=[δ1,δ2,…,δl]
Wherein, U is the control strategy of composite material parts heating mode, and δ is the switch of some microwave source in microwave cavity
(value is 0 or 1 to state, and is 0, is opened as 1), l is the number of certain specific microwave source in microwave cavity (value is less than or equal to 16).
Heating rate distribution of the same layer material of part in a period of time after normalized is defined as heating mode:
Wherein
In above-mentioned formula, c indicates normaliztion constant (value 10)；ThIndicate the temperature of the same layer material of h moment part
Distribution；Th-1For the Temperature Distribution of the same layer material of h-1 moment part；Δ t is this heating time；Pij rIndicate length on part
Direction is i, and width direction is the material interior heating rate after normalization when heated of the position j；When indicating h
Carving length direction on part is i, and width direction is the temperature value of the material of the position j；Indicate that the h moment passes through temperature sensor
The mean temperature of the same layer material of the part measured；Indicate the same layer material institute of interior part when heated
The maximum value for thering is temperature measuring point temperature to rise.
Before beginning to warm up, the weighting parameter of prediction model is initialized, temperature process curve is set.Define loss function:
Wherein, δiFor the control strategy history tab data generated in part microwave heating process, δi' it is model according to defeated
The control strategy of the heating mode prediction entered.Any time k during heating, prediction model use heating mode-control
Heating mode HP and control strategy U data are organized to { (HP apart from current time nearest preceding p (p=50) in policy databasek-1,
Uk-1),(HPk-2,Uk-2),…,(HPk-p,Uk-p) exercise supervision training.Forgetting Mechanism is taken to get over apart from current time simultaneously
The contribution that remote data update Model Weight is smaller: the time order and function that 50 groups of data for being used for training are generated according to data is suitable
Sequence is divided into 5 groups, and every group of 10 data, each group of data input neural network updates a weight later, according to gradient decline
Thought is trained to weight with new formula every time are as follows:
Wherein η is constant, and λ is Forgetting coefficient, and value is smaller to illustrate that Forgetting Mechanism effect is more obvious.N expression group number, n=0 table
Show most newly generated 10 groups of data, n=4 indicates the 10 group data farthest away from current time.If the total data pair currently generated
Less than 50 groups, is then exercised supervision training using all data for currently generating to model, Forgetting Mechanism is not used at this time, based on ladder
Spend the changing ideas weight of decline:
The variation of concern microwave heating system behavioral characteristics in real time in this way.
Prediction model is constantly trained during heating, while the heating mode based on input predicts corresponding control plan
Slightly.The k moment after heating starts, according to present material surface temperature distribution
Target heating mode is calculated using following formula in thought based on pattern compensation
Wherein:
Using target heating mode as the input of prediction model, one group of control strategy is obtained:
U (k)=[δ1,δ2,…,δ16],δi=0 or 1
Wherein in running order magnetron quantity is m, and m≤16, power controller is based on pid algorithm, according to real-time
Program control temperature difference information in the mean temperature of 30 temperature spots monitored and corresponding moment target process curve obtains micro-
General power P required for Wave heating systemm, therefore, the operating power of microwave heating system are as follows:
P=U × Pm÷ m=[δ1,δ2,…,δ16]×Pm÷m
This group of control strategy is applied in microwave heating system and is heated Δ t seconds, is heated by temperature sensor
The Temperature Distribution T (k) of same layer material afterwards is equally normalized to calculate the heating rate distribution in this period
Processing obtains actual heating rate distribution i.e. heating mode:
Wherein:
Meanwhile above-mentioned heating mode control strategy data are to (HPk,Uk) will be as a group of labels data storage to heating
In mode-control strategy database.
Above procedure is repeated until completing heating.
The present invention uses in neural network model real-time learning part microwave heating process between heating mode and control strategy
Dynamic associations, and based on above-mentioned model according to the real-time predictive compensation Current Temperatures distribution of the thought of heating mode complementation
Control strategy carries out accurate, intelligent compensation to non-uniform Temperature Distribution, realizes to Part temperature uniformity in heating process
Accurate control.
The above is only specific application examples of the invention, are not limited in any way to protection scope of the present invention.All uses
Equivalents or equivalence replacement and the technical solution formed, all fall within rights protection scope of the present invention.
Part that the present invention does not relate to is the same as those in the prior art or can be realized by using the prior art.
Claims (6)
1. a kind of microwave heating temperature field intelligent control method based on on-line study, it is characterised in that: use neural network mould
Dynamic associations in type real-time learning part microwave heating process between heating mode and control strategy, and it is based on above-mentioned model
According to the control strategy of the real-time predictive compensation Current Temperatures distribution of the thought of heating mode complementation, uniform microwave is carried out to part and is added
Heat.
2. according to the method described in claim 1, it is characterized by: establishing the heating mode control of part based on neural network algorithm
Make tactful prediction model；In microwave heating process, using the Temperature Distribution of the same layer material of temperature sensor real-time monitoring part,
For any time k (k >=p), heated using in heating mode-control strategy database apart from current time nearest preceding p group
Mode HP and control strategy U data pair:
{(HPk-1,Uk-1),(HPk-2,Uk-2),…,(HPk-p,Uk-p)}
It exercises supervision training to above-mentioned prediction model；
As k < p, using preceding k-1 group heating mode HP and control strategy U data pair
{(HP1,U1),(HP2,U2),…,(HPk-1,Uk-1)}
It exercises supervision training to above-mentioned prediction model；
After the completion of training, the thought based on heating mode complementation is quickly calculated for compensating Current Temperatures distribution Tk-1Target
Heating mode HP 'k, and be input in the prediction model for completing training, quick predict target heating mode HP 'kControl plan
Slightly Uk；
In control strategy UkOn the basis of, power controller adjusts the power of each in running order microwave source in real time, according to
Scenario earthquake carries out uniform microwave heating to part；
By control strategy UkAfter running the Δ t time, T is distributed based on Current TemperatureskQuickly calculate control strategy UkUnder it is actual plus
Heat pattern HPk, and by newest heating mode-control strategy data to (HPk,Uk) it is saved in heating mode-control strategy data
In library；
It repeats the above process, until completing the entire microwave heating process of part.
3. according to the method described in claim 2, it is characterized by: the heating mode of the part is any control strategy UkUnder
Heating rate at the same layer material each point of partC is constant.
4. according to the method described in claim 2, it is characterized by: the control strategy of the heating mode is multiple microwave sources
Assembled state U=[δ1,δ2,…,δl], wherein δ represents the switch state of microwave source, and opening value is 1, and closing value is that 0), l is represented
The number of each microwave source on micro-wave oven.
5. according to the method described in claim 2, it is characterized by: the heating mode complementation thought is i.e. to part same layer
The lower region of material temperature applies biggish microwave power or heating rate, to the higher region of part same layer material temperature
Apply lesser microwave power or heating rate.
6. according to the method described in claim 2, it is characterized by: power controller is based on PID according to the temperature curve of setting
Algorithm calculates the intracorporal general power of microwave cavity in real time, and in control strategy UkOn the basis of power increment is averagely allocated to currently
The microwave source of operation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811197619.0A CN109287021B (en) | 2018-10-15 | 2018-10-15 | Online learning-based intelligent monitoring method for microwave heating temperature field |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811197619.0A CN109287021B (en) | 2018-10-15 | 2018-10-15 | Online learning-based intelligent monitoring method for microwave heating temperature field |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109287021A true CN109287021A (en) | 2019-01-29 |
| CN109287021B CN109287021B (en) | 2021-01-12 |
Family
ID=65176486
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811197619.0A Active CN109287021B (en) | 2018-10-15 | 2018-10-15 | Online learning-based intelligent monitoring method for microwave heating temperature field |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109287021B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110213844A (en) * | 2019-06-28 | 2019-09-06 | 厦门艾美森新材料科技股份有限公司 | A kind of automatic compensating method of air cushioning machine and its heater strip heating power |
| CN114621778A (en) * | 2020-12-11 | 2022-06-14 | 中国石油化工股份有限公司 | Memory, temperature control method, device and equipment for biomass microwave pyrolysis process |
| CN114679806A (en) * | 2022-04-01 | 2022-06-28 | 昆明理工大学 | Self-switching control method and system for improving microwave heating uniformity |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH05113220A (en) * | 1991-10-21 | 1993-05-07 | Matsushita Electric Ind Co Ltd | Cooking equipment |
| GB2293027A (en) * | 1994-09-07 | 1996-03-13 | Sharp Kk | Apparatus for and method of controlling a microwave oven |
| CN103561497A (en) * | 2013-11-18 | 2014-02-05 | 四川大学 | Distributed type microwave heating and drying control device and method |
| CN105142408A (en) * | 2012-12-04 | 2015-12-09 | 英戈·施托克格南特韦斯伯格 | Heat treatment monitoring system |
| CN106037448A (en) * | 2016-07-29 | 2016-10-26 | 广东美的厨房电器制造有限公司 | Cooking control method and equipment and cooking device |
| CN107071953A (en) * | 2017-04-10 | 2017-08-18 | 南京航空航天大学 | Based on the complementary microwave heating temperature uniformity Active Control Method of heating mode |
-
2018
- 2018-10-15 CN CN201811197619.0A patent/CN109287021B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH05113220A (en) * | 1991-10-21 | 1993-05-07 | Matsushita Electric Ind Co Ltd | Cooking equipment |
| GB2293027A (en) * | 1994-09-07 | 1996-03-13 | Sharp Kk | Apparatus for and method of controlling a microwave oven |
| CN105142408A (en) * | 2012-12-04 | 2015-12-09 | 英戈·施托克格南特韦斯伯格 | Heat treatment monitoring system |
| CN103561497A (en) * | 2013-11-18 | 2014-02-05 | 四川大学 | Distributed type microwave heating and drying control device and method |
| CN106037448A (en) * | 2016-07-29 | 2016-10-26 | 广东美的厨房电器制造有限公司 | Cooking control method and equipment and cooking device |
| CN107071953A (en) * | 2017-04-10 | 2017-08-18 | 南京航空航天大学 | Based on the complementary microwave heating temperature uniformity Active Control Method of heating mode |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110213844A (en) * | 2019-06-28 | 2019-09-06 | 厦门艾美森新材料科技股份有限公司 | A kind of automatic compensating method of air cushioning machine and its heater strip heating power |
| CN114621778A (en) * | 2020-12-11 | 2022-06-14 | 中国石油化工股份有限公司 | Memory, temperature control method, device and equipment for biomass microwave pyrolysis process |
| CN114621778B (en) * | 2020-12-11 | 2023-09-01 | 中国石油化工股份有限公司 | Memory, biomass microwave pyrolysis process temperature control method, device and equipment |
| CN114679806A (en) * | 2022-04-01 | 2022-06-28 | 昆明理工大学 | Self-switching control method and system for improving microwave heating uniformity |
| CN114679806B (en) * | 2022-04-01 | 2024-04-12 | 昆明理工大学 | Self-switching control method and system for improving microwave heating uniformity |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109287021B (en) | 2021-01-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109287021A (en) | A kind of microwave heating temperature field intelligent control method based on on-line study | |
| CN108563121B (en) | Intelligent microwave heating temperature field monitoring method based on historical data | |
| CN108196607B (en) | The temprature control method of the electromagnetic induction heating system of uncertain stochastic Time-Delay | |
| Sun | Adaptive and intelligent temperature control of microwave heating systems with multiple sources | |
| WO2018188513A1 (en) | Method for active control of microwave heating temperature uniformity based on heating mode complementarity | |
| Damour et al. | Energy analysis and optimization of a food defrosting system | |
| Wang et al. | Research on variable universe fuzzy PID control strategy of pipe lining induction heating system | |
| Li et al. | A novel method to improve temperature uniformity in polymer composites microwave curing process through deep learning with historical data | |
| Pradeep et al. | A finite horizon Markov decision process based reinforcement learning control of a rapid thermal processing system | |
| Ma et al. | Using deep reinforcement learning for zero defect smart forging | |
| Akkari et al. | Global linearizing control of MIMO microwave-assisted thawing | |
| Domingo et al. | Fast power–frequency function estimation for induction heating appliances | |
| Zeng et al. | Estimation and coordinated control for distributed parameter processes with a moving radiant actuator | |
| Lee et al. | Microwave thawing of cylinders | |
| Kutasi et al. | Heat transport modelling and adaptive model predictive temperature control of the direct current plasma nitriding process performed in a linear non-isotherm plasma reactor | |
| Vasilyev et al. | The functional dependencies of the drying coefficient for the use in modeling of heat and moisture-exchange processes | |
| Li et al. | Combining sliding mode neural network with Cuckoo Search to make a uniform microwave heating process | |
| CN108366446B (en) | Microwave heating temperature field control method based on time-varying electromagnetic field | |
| Kim | Control-oriented dynamic model of an inductively coupled plasma torch by artificial intelligence methodology | |
| US20230239966A1 (en) | Model-based control method, model-based control system, and storage medium | |
| Winkler et al. | Flatness-based control of a continuous furnace | |
| Jiang et al. | Particle Swarm Optimization Algorithm Based on Natural Selection and Simulated Annealing for PID Controller Parameters | |
| Voutetakis et al. | Model-based control of temperature and energy requirements in a fluidised furnace reactor | |
| CN117192996A (en) | Microwave metallurgy process multi-intelligent model predictive control method based on priority negotiation | |
| Yahi et al. | Machine Learning Assisted Resonant Inverter Control for Temperature Regulation in Rubber Manufacturing |
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 |