CN107015480A - A kind of intelligent greenhouse irrigation system based on generalized predictive control and Internet of Things - Google Patents
A kind of intelligent greenhouse irrigation system based on generalized predictive control and Internet of Things Download PDFInfo
- Publication number
- CN107015480A CN107015480A CN201710350659.3A CN201710350659A CN107015480A CN 107015480 A CN107015480 A CN 107015480A CN 201710350659 A CN201710350659 A CN 201710350659A CN 107015480 A CN107015480 A CN 107015480A
- Authority
- CN
- China
- Prior art keywords
- event
- control
- controller
- sampling
- irrigation system
- 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
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Greenhouses (AREA)
- Feedback Control In General (AREA)
Abstract
The invention discloses a kind of intelligent greenhouse irrigation system based on generalized predictive control and Internet of Things, first, using the generation of the transpiration model trigger event of crops；Then, using the output of soil humidity value and event generation module as Generalized Prediction model input；Finally, the output of zero-order holder module is controlled into the open and close of greenhouse irrigation system as the input of drive system.Event-based control device includes two parts：Event detector and controller, event detector decide whether that, by the event notification controller of kainogenesis, controller is made up of one group of generalized predictive controller, when detecting a new event, according to the one of generalized predictive controller of time point selection.Nantong Area attached-greenhouse test result indicates that, this control system reduces 20% water consumption on the premise of effective greenhouse irrigation effect is realized, and reduces the opening time of irrigation system, reduces the cost of control.
Description
Technical field
The present invention relates to a kind of intelligent greenhouse irrigation system based on generalized predictive control and Internet of Things.
Background technology
Precision agriculture is that field soil in a kind of new farming method, farmland, the characteristic of crop are not homogeneous
, typically over time, space and change, and traditional farmland management then thinks that these factors are all homogeneous, thus using
Unified fertilization time, dose and Irrigation regime.Precision agriculture has following 3 features compared with traditional agriculture, mainly：1.
Rational use of chemical fertilizer, reduces production cost；2. reduce and saving water resource；3. cost-saving synergistic, saving of work and time, good quality and high output.From
Dynamicization irrigation system is an important part in precision agriculture.
At present, many researchers devise effective automatic irrigation system：Shi Jianfei, Cao Hongjun, the refined beautiful grade of clothing is in " base
In PLC paddy rice in cold region water-fertilizer integral development of system " in devise using PLC as control core, by side wireless communication
Formula realizes the system that the data such as the soil moisture in paddy rice farmland, water level, equipment working state are acquired with analysis, improves filling
Uniformity, promptness and the simplicity irrigate and applied fertilizer.Wei Yichang, Wang Zhenying, Wang Tongchao etc. are in " soil moisture content monitoring in real time and essence
The soil moisture sensor for possessing independent intellectual property right, data acquisition and control module, data are utilized in the design of quasi- irrigation system "
The soil moisture content monitoring and Precision Irrigation control software of the hardware such as transport module, pipeline flowmeter and oneself establishment, are realized
Using water wisely, raising yield, the effect of improvement crop quality.Wei Kai, An Jinqiang, Jia Ye etc. are in " the perseverance based on ZigBee technology
ZigBee technology is based in pressure irrigation system research ", using HMI and PLC interaction platforms fuzzy closed-loop control system, is ground
A kind of drip irrigation automatic control system that can automatically realize and accurately control duty and constant ductwork pressure is made.Shin J
H, Park J S, Son J E are in Estimating the actual transpiration rate with compensated
1evels of accumulated radiation for the efficient irrigation of soilless
Recorded in cultures of paprika plants and devised the drip irrigation system based on PLC for soilless culture system,
The system considers influence of the factors such as illumination, transpiration and drainage system to soilless culture water consumption.Above-mentioned scheme
Mainly the humidity of soil is monitored in real time by technology of Internet of things mostly, according to the soil moisture value of detection in real time
Duty and time are controlled, what is had considers illumination, the influence of transpiration, devises more fine automatic irrigation system
System, its water-saving result is preferable.
Existing design is main to carry out appropriate compensation according to soil moisture to moisture, but it is rising to rarely have research to consider
Effect, influence of the change of intensity of illumination to crop water requirement, cause the water consumption of automatic irrigation system and control cost compared with
It is high.
The content of the invention
The technical problem to be solved in the present invention be overcome the defect of prior art there is provided one kind based on generalized predictive control with
The intelligent greenhouse irrigation system of Internet of Things, can be while common commercial irrigation system irrigation performance be realized, it will be apparent that reduction
Irrigation cost and control cost.
In order to solve the above-mentioned technical problem, the invention provides following technical scheme：
The invention discloses a kind of intelligent greenhouse irrigation system based on generalized predictive control and Internet of Things, it is by following step
It is rapid to produce：
S1, using crops transpiration model trigger event generation；
S2, using the output of soil humidity value and event generation module as Generalized Prediction model input；
S3, using the output of zero-order holder module as drive system input, control greenhouse irrigation system unlatching with
Close.
Further, the transpiration model in step S1 wanes including solar irradiance with vapour pressure, the structure of model
It is as follows：
A(z-1) y (t)=B (z-1)u(t-nk)+C(z-1)e(t) (1)
The transpiration rate estimated when y (t) is time discretization t in model in formula, A (z-1) and C (z-1) it is n respectivelyaWith ncRank
Monic polynomial, wherein z-1Represent backward shift operator, B (z-1)=[B1(z-1)B2(z-1)] it is 1 × 2 vector tieed up, include two
Multinomial, both ranks are nb, u (t-nk) is the column vector of 2 × 1 dimensions, contains the input parameter of modelWhereinu2(t)=VVPD(t).When being the discretization of each input correlation
Between postpone, e (t) is evaluated error；Therefore it can obtain：
Further, described irrigation system controls the supply of water using on/off controllers.
Further, event generation module includes two parts：Event detector and event controller, event detector are determined
Fixed whether by the event notification controller of kainogenesis, event controller is made up of one group of generalized predictive controller, when detecting one
During individual new event, according to the one of generalized predictive controller of time point selection.
Further, the output of event generation module sampling greenhouse process, the sampling time is Tbase, if sampling control
Behavior processed, the then frequency occurred according to event adjusts the sampling time Tf of change；Tf is Tbase multiple (Tf=fTbase, f
∈ [1, nmax]), and Tf≤Tmax, Tmax=nmaxTbase are maximum sampling time values；Event generation module is according to each
The basic sampling period monitors the output of control process.
Further, whether event detector meets some specified requirements using the output of the Information Authentication greenhouse process,
If meeting these conditions, an event is produced using sampling period Tf, so as to save secondary control activity；Otherwise, only
A control activity is produced in t=t+tmax；Sampling time Tf based on change calculates control activity, therefore uses one group
GPC controllers, each GPC controllers one sampling time Tf=fTbase of correspondence, f ∈ [1, nmax].
Further, control process uses GPC algorithms as feedback controller, using one group of GPC controller, each control
One sampling time T of device correspondencef, f ∈ [1, nmax], in the controller set each controller by using it is corresponding discrete when
Between the classical GPC algorithms of model realization；The target of GPC controllers is the cost function for minimizing multiple stages：
In formulaIt is j steps optimal before system exports prediction, t is current time, Δ uf(t+j-1) it is not
Carry out controlling increment, w (t+j) is future reference track, includes sampling time Tf(t=kTf, k ∈ Z+) in all signals；In addition,
Tuning parameter be respectively minimum prediction level, maximum predicted level,Controlled level,Following error, control add
Weight factor δfWith λf；Prediction level is defined asWithWeighted factor δf=1；GPC target is to calculate
Following control sequenceI.e. by minimizing J, the crops close to w (t+j) are derived
Future output yf(t+j)。
Further, by the sampling of event generator module management event, the module is produced using two different conditions
Raw new event, if a condition is changed into TRUE, produces a new event, the current demand signal of process is transmitted to control
Module, according to one new control activity of the signal of change；
First condition determines the time of event using rising model, and uses asynchronous-sampling；If crops are rising
Total amount y (t) be more than the threshold value beta specified, then produce a new event, y (t) computational methods are as follows：
In formulaIt is last event eiThe time of generation；If its value is set to 0 by y (t) more than β；If lost
Total Water reaches specified threshold β, then generates new control system event, it is clear that β value determines the frequency that the system event is produced；
Second condition is a time-based condition, and the condition is used to improve stability；The conditional definition is two
The calculating time difference of control signal, is set to Tmax：
Use minimum sampling time slot TbaseSecond condition is checked, the sampling time T of change is usedf=fTbase, f ∈
[1, nmax] all events are checked.
Further, in addition to signal reconstruction and resampling technology：Use the sampling time section T of changefDetermine one
New control activity, therefore, in order to realize GPC control algolithms, past value and the control signal of process variable must be available (adopt
Sample time interval is Tf), so needing to rebuild corresponding signal；
Assuming that a control signal is ub, every TbaseTime preserves control signal value using the variable；First, calculate
Required past information, more new signal ub；Assuming that producing a new event, cause a new sampling period Tf=fTbase,
Because using TbaseSample ubValue, so using ubThe average value of middle f value in the past is rebuildPast value；
I=P in formulau..., 1, j=k-1- (Pu- i) f,With PuIt is u respectivelyfPast value and required past value number
Amount；First, according to ubIn past value calculate new sampling time TfPast information, be stored in a variable and (be set to), make
(y is set to the information and past the output of process dataf) new control activity is calculated, pass through the normal of two continuous events of holding
Value updates ubSignal, uf(Tf)=ub(k)；
According to the above-mentioned GPC work rules based on event, the output of process variable monitored using asynchronous-sampling, in order to recover
Information between two continuous events, the letter is rebuild using lagrange formula method.
The beneficial effect that is reached of the present invention is：
The present invention manages GPC controllers using the event generator based on the rising model of crops, according to control process
The activation frequency of dynamic regulation control system.It is under the conditions of different weather test result indicates that, this control system reduces control
Cost, improve control accuracy.Compared with common ON/OFF control device, this control system can keep close irrigation
In the case of performance, reduce by 20% water consumption.
Brief description of the drawings
Accompanying drawing is used for providing a further understanding of the present invention, and constitutes a part for specification, the reality with the present invention
Applying example is used to explain the present invention together, is not construed as limiting the invention.In the accompanying drawings：
Fig. 1 is the structural representation of controller；
Experimental result of Fig. 2 control systems of the present invention in sunny weather；
Fig. 3 is experimental result of the control system of the present invention in cloudy weather；
Fig. 4 is the result of this ordinary greenhouse irrigation system；
Fig. 5 is this control system experimental result of the 6th day.
Embodiment
The preferred embodiments of the present invention are illustrated below in conjunction with accompanying drawing, it will be appreciated that preferred reality described herein
Apply example to be merely to illustrate and explain the present invention, be not intended to limit the present invention.
A kind of intelligent greenhouse irrigation system based on generalized predictive control and Internet of Things, it is produced by following steps：
S1, using crops transpiration model trigger event generation；
S2, using the output of soil humidity value and event generation module as Generalized Prediction model input；
S3, using the output of zero-order holder module as drive system input, control greenhouse irrigation system unlatching with
Close.
Greenhouse irrigation system
Greenhouse
The data of this research come from the agricultural greenhouse laboratory of Nantong City, and crop growth is in 700 square metres of company
In greenhouse, greenhouse ceiling be vinyon top, greenhouse it is Height Adjustable.Greenhouse direction is east-west direction, and crops are south
Northern directional spreding.
In the seedling of Luo Kewei rock wool-steel plywood culture tomatoes, tomato cropping intensity be 2 crops/square metre,
Tomato is transplanted on flat board by tomato before blooming.Tomato soilless culture is conveyed by a drip irrigation system and marked
Accurate nutrient solution, each flat board sets three drip irrigation transmitters.
Rising model
Parametrization black-box model can represent arbitrary system, represent transpiration using 32MISO ARMAX models families herein
Dynamic, the model comprising two input：Solar irradiance wanes with vapour pressure, and the structure of model is as follows：
A(z-1) y (t)=B (z-1)u(t-nk)+C(z-1)e(t) (1)
The transpiration rate estimated when y (t) is time discretization t in model in formula, A (z-1) and C (z-1) it is n respectivelyaWith ncRank
Monic polynomial, wherein z-1Represent backward shift operator, B (z-1)=[B1(z-1)B2(z-1)] it is 1 × 2 vector tieed up, include two
Multinomial, both ranks are nb, u (t-nk) is the column vector of 2 × 1 dimensions, contains the input parameter of modelWhereinu2(t)=VVPD(t).When being the discretization of each input correlation
Between postpone, e (t) is evaluated error.Therefore it can obtain：
It is this paper black-box model shown in table 1[9], rising dynamic is non-linear process, can be approximately linear model, use
Leaf area index (LAI) XLAIBe divided into different intervals the crops cycle：(0~0.7), (0.7~1.5), (1.5 ,+∞).
The black-box model of the tomato Crop transpirstion of table 1 effect
The On/Off controllers of irrigation system
Greenhouse irrigation process typically uses on/off controllers[10]To control the supply of water.
Irrigation system based on generalized predictive control
Controller architecture
Event-based control device includes two parts：Event detector and controller.Event detector decide whether by
The event notification controller of kainogenesis, controller is made up of one group of GPC (generalized predictive control) controller, new when detecting one
Event when, according to the one of GPC controllers of time point selection.It is the complete structure of controller shown in Fig. 1, contains processing
Device, actuator, controller and event generator.The design philosophy of this controller is as follows：
The output of greenhouse process 1. event generator module is sampled, the sampling time is TbaseIf sampling controlling behavior,
The frequency then occurred according to event adjusts the sampling time T of changef。
2.TfIt is TbaseMultiple (Tf=fTbase, f ∈ [1, nmax]), and Tf≤Tmax, Tmax=nmaxTbaseIt is maximum adopt
Sample time value.
3. event generation module is according to each basic sampling period (Tbase) monitoring control process output.Detection module makes
Some specified requirements whether are met with the output of the Information Authentication greenhouse process, if meeting these conditions, sampling are used
Cycle TfAn event is produced, so as to save secondary control activity；Otherwise, only in t=t+tmaxWhen produce a control activity.
4. the sampling time T based on changefControl activity is calculated, therefore uses one group of GPC controller, each GPC controls
One sampling time T of device correspondencef=fTbase, f ∈ [1, nmax]。
GPC algorithm
This control process uses GPC algorithms as feedback controller.Use one group of GPC controller, each controller correspondence
One sampling time Tf, f ∈ [1, nmax], each controller is by using corresponding discrete time model in the controller set
Realize classical GPC algorithms.The target of GPC controllers is the cost function for minimizing multiple stages：
In formulaIt is j steps optimal before system exports prediction, t is current time, Δ uf(t+j-1) it is not
Carry out controlling increment, w (t+j) is future reference track, includes sampling time Tf(t=kTf, k ∈ z+) in all signals.In addition,
Tuning parameter be respectively minimum prediction level, maximum predicted level,Controlled level,Following error, control add
Weight factor δfWith λf.Prediction level is defined asWithWeighted factor δf=1.GPC target is to calculate
Following control sequenceI.e. by minimizing J, the crops close to w (t+j) are derived
Future output yf(t+j)。
Signal sampling based on event
As can be seen from Figure 1 by the sampling of event generator module management event, the module uses two different conditions
New event is produced, if a condition is changed into TRUE, a new event is produced, the current demand signal of process is transmitted to control
Molding block, according to one new control activity of the signal of change.
First condition determines the time of event using rising model, and uses asynchronous-sampling.If crops are rising
Total amount y (t) be more than the threshold value beta specified, then produce a new event, y (t) computational methods are as follows：
In formulaIt is last event eiThe time of generation.If its value is set to 0 by y (t) more than β.If lost
Total Water reaches specified threshold β, then generates new control system event, it is clear that β value determines the frequency that the system event is produced.
Second condition is a time-based condition, and the condition is used to improve stability.The conditional definition is two
The calculating time difference of control signal, is set to Tmax：
Use minimum sampling time slot TbaseSecond condition is checked, the sampling time T of change is usedf=fTbase, f ∈
[1, nmax] all events are checked.
Signal reconstruction and resampling technology
As described above, using the sampling time section T of changefA new control activity is determined, therefore, in order to realize GPC
Control algolithm, the past value of process variable and control signal must be that available (sampling time interval is Tf), so needing weight
Build corresponding signal.
The reconstruction of past control signal
Assuming that a control signal is nb, every TbaseTime preserves control signal value using the variable.First, calculate
Required past information, more new signal ub.Assuming that producing a new event, cause a new sampling period Tf=fTbase,
Because using TbaseSample ubValue, so using ubThe average value of middle f value in the past is rebuildPast value.
I=P in formulau..., 1, j=k-1- (Pu- i) f,With PuIt is u respectivelyfPast value and required past value number
Amount.First, according to ubIn past value calculate new sampling time TfPast information, be stored in a variable and (be set to), make
(y is set to the information and past the output of process dataf) new control activity is calculated, pass through the normal of two continuous events of holding
Value updates ubSignal, uf(Tf)=ub(k)。
The reconstruction of the output of process
According to the above-mentioned GPC work rules based on event, the output of process variable monitored using asynchronous-sampling, in order to recover
Information between two continuous events, herein using lagrange formula method[11]Rebuild the information.
As a result
Emulation experiment
Experiment condition
Simulation study is using Fig. 1 greenhouse irrigation system and the meteorological data in 2016 spring, in order to obtain reliable data,
All control systems are tested in different 10 days in weather condition.Using PWM technology humanized electromagnetism ON/OFF, by controller
Continuous signal is converted to the pulse of change width, and the width of pulse is determined by control signal, and scope is 0%~100%, by PWM
Modulating frequency be set to 0.02Hz.
The setting of model parameter
The dynamic of acquisition procedure is first had in the development process of control system, so the variable near selection desired operation point
Carry out several groups of experiments.Irrigation process description is integral process, as follows：G (s)=0.005/s.This irrigation process control
GPC parameters are set to following value：Controlled level is set to Nu=5, prediction level is set to N2=15, pass through experiment statisticses control signal
Weighting parameters λ, λ is set to 5 to obtain desired control system performance.The minimum sampling gap of GPC controllers is set to 5 points
Clock.
For the performance of analysis and Control system, the common ON/OFF control device in 5 minute sampling time is tested.Using
Absolute integral error (IAE) determines the control performance of every group of configuration parameter,：
The formula calculates setting value and the error of control variable.The index is widely used in the evaluation of control system performance
In, water consumption WU is defined as：1m2The total water consumption of regional irrigation, event index definition is：The event that every group of configuration parameter is produced
Quantity.
First, influence of the sampling time slot of analysis and Control system to performance, following value is set to by β variables：β=0,
0.1,0.5,0.75,1,1,5,2,2.5 }.The configuration of β=0 is corresponding to classical system, when its output of process is fixed sampling
Gap (5 minutes).It is the system based on event because asynchronously trigger controller for other β value, affair logic
The event number that width value is triggered on event generation module has direct influence, and determines the performance of control system.It is wide
Angle value is smaller, and event number is higher, and control performance is better；Otherwise, event number is lower, and the performance of control system is poorer.β value is determined
Control cost and control performance are determined, because ON/OFF control device is the most widely used controller of greenhouse irrigation system, institute
To be emulated to ON/OFF control device simultaneously.
It is different configuration of Control performance standard shown in table 2, it can be seen that β=0 obtains optimal control performance, but this is matched somebody with somebody
The control cost put is wherein highest.The system has six to assemble the performance put to be reduced better than ON/OFF control device, and WU values
10%, it will be apparent that reduce water consumption.Table 2 shows the cost that control system is can significantly reduce using larger width value,
But the precision of control is reduced simultaneously, so must be weighed therebetween.For greenhouse irrigation system, this control system
Control accuracy be higher than ON/OFF control device, and control cost it is also relatively low, the system be based on event sampling realize dynamic tune
Section, so the system is better than ON/OFF control device.
This controller using β=1.5 obtain with ON/OFF control device identical control performance, both IAE are respectively etc.
In 14.5 and 14.6.The system controls cost to be WU=27.11, and to reduce about 20% than ON/OFF control system use water
Amount.Because ON/OFF control device has fully met the demand of current greenhouse irrigation application, the system selection β=1.5
Carry out follow-up analysis.
The different configuration of Control performance standard of table 2
Fig. 2 and Fig. 3 is result of this control system (β=1.5) for experiment the 2nd and 9 day respectively.Fig. 2 shows this control
Device processed for fine day control performance, fine day cause transpiration acutely, the water demand of crop improve, cause the control of feed water valve to be believed
Number change is violent, and event-generator continually trigger event.It can be seen that the humidity of soil remains setting value in figure, when
Controller is just activated when transpiration changes.
Fig. 3 shows control performance of this controller (β=1.5) for cloudy weather.Because transpiration is reduced,
The water requirement reduction of this day, causes the control signal change frequency of feed water valve relatively low, event generator produces the frequency of event
Substantially reduction.Compare the control signal of second and the 9th day, it can be seen that what this controller can be according to needed for being adjusted actual demand
Control cost, the feature is especially important in greenhouse irrigation system, the feature can reduce water consumption, reduce the maintenance in greenhouse into
This.
Experimental result of this control system of Fig. 2 in sunny weather：(a) the rising water of the change curve (b) of light radiation intensity
Change curve (c) humidity change curve (d) water control valve change curve (e) control system event change curve
Experimental result of this control system of Fig. 3 in cloudy weather：(a) the rising water of the change curve (b) of light radiation intensity
Change curve (c) humidity change curve (d) water control valve change curve (e) control system event change curve
Experimental evaluation
The experimental result in 4 days to No. 10 May is counted, because weather is relatively warmed up, the yield of crop is higher with transpiration rate, leads
Cause output higher.
Realize in a distributed fashion and sensor and actuator is provided with the configuration of control system, greenhouse, use NI (states
Family instrument) compatibility-FieldPoint hardware come carry out perceive with activation task.Each compatibility-FieldPoint units equipment
AD (analog-to-digital conversion) and DA (digital-to-analogue conversion) module.Controller node is set up in the PC of a standard, controller node makes
This controller is performed with the software based on LabVIEW, programmed environment is Matlab2011b.For the ease of realizing, control system
All nodes are connected by a special Ethernet.
The first step of exploitation control system is the dynamic for capturing control process, in order to obtain dynamic process response, to the phase
The independent variable hoped around operating point carries out several groups of experiments.It is linear programming form that process model building, which will be irrigated,：G (s)=0.0005/
Second.Because process is dynamic change, the GPC parameters for irrigating process control are set to following value：Controlled level is set
For Nu=5, prediction level is set to N2=10 (the main process dynamics of crawl), the weight factor λ of control signal is set to 2, and glug is bright
The degree of day formula is set to 2.GPC controllers set the sampling time of 7 minutes, and the setting value of soil moisture is set in all analyses
60%.According to the experiment of upper one section, β=1.5 are set.
First, using widely used commercial greenhouse irrigation system[12]Tested, be representative one shown in Fig. 4
It can be seen that, successfully maintain soil moisture near desired value in it greenhouse irrigation process, figure, use a fixed mould
Formula realizes the injection of water.In the irrigation system of business, WD values are 20% of total water consumption or so, and this control program is significantly reduced
WD values.It must be pointed out：By adjusting the relation that β value is adjustably controlled between performance and control cost.
It is this paper control systems experimental result of the 6th day in Fig. 5.It can be seen that transpiration causes from the change of control signal
This controller adjusts the supply of water, and transpiration is higher, and output is higher.Because this control system is according to the true need of crop
Regulation output is sought, so total water consumption is reduced, meanwhile, the mechanism humidity that this control system is safeguarded is closer to expect
Value, so the control performance of this control system is more preferable.In general, this control system improves control accuracy, reduces control
Cost.
Finally it should be noted that：The preferred embodiments of the present invention are the foregoing is only, are not intended to limit the invention,
Although the present invention is described in detail with reference to the foregoing embodiments, for those skilled in the art, it still may be used
To be modified to the technical scheme described in foregoing embodiments, or equivalent substitution is carried out to which part technical characteristic.
Within the spirit and principles of the invention, any modification, equivalent substitution and improvements made etc., should be included in the present invention's
Within protection domain.
Claims (9)
1. a kind of intelligent greenhouse irrigation system based on generalized predictive control and Internet of Things, it is characterised in that produced by following steps
It is raw：
S1, using crops transpiration model trigger event generation；
S2, using the output of soil humidity value and event generation module as Generalized Prediction model input；
S3, using the output of zero-order holder module be used as drive system input, control greenhouse irrigation system open and close.
2. the intelligent greenhouse irrigation system according to claim 1 based on generalized predictive control and Internet of Things, its feature exists
In the transpiration model in step S1 wanes including solar irradiance with vapour pressure, and the structure of model is as follows：
A(z-1) y (t)=B (z-1)u(t-nk)+C(z-1)e(t) (1)
The transpiration rate estimated when y (t) is time discretization t in model in formula, A (z-1) and C (z-1) it is n respectivelyaWith ncThe monic of rank is multinomial
Formula, wherein z-1Represent backward shift operator, B (z-1)=[B1(z-1)B2(z-1)] it is 1 × 2 vector tieed up, include two multinomials, both ranks
It is nb, u (t-nk) is the column vector of 2 × 1 dimensions, contains the input parameter of model
Whereinu2(t)=VVPD(t),It is the related time discretization delay of each input, e (t) is evaluated error；
Therefore it can obtain：
3. the intelligent greenhouse irrigation system according to claim 1 based on generalized predictive control and Internet of Things, its feature exists
In described irrigation system controls the supply of water using on/off controllers.
4. the intelligent greenhouse irrigation system according to claim 3 based on generalized predictive control and Internet of Things, its feature exists
In event generation module includes two parts：Event detector and event controller, event detector decide whether kainogenesis
Event notification controller, event controller is made up of one group of generalized predictive controller, when detecting a new event, root
According to the one of generalized predictive controller of time point selection.
5. the intelligent greenhouse irrigation system according to claim 4 based on generalized predictive control and Internet of Things, its feature exists
In, the output of event generation module sampling greenhouse process, the sampling time is Tbase, if sampling controlling behavior, according to thing
The sampling time Tf for the frequency regulation change that part occurs；Tf is Tbase multiple (Tf=fTbase, f ∈ [1, nmax]), and Tf
≤ Tmax, Tmax=nmaxTbase are maximum sampling time values；Event generation module is monitored according to each basic sampling period
The output of control process.
6. the intelligent greenhouse irrigation system according to claim 5 based on generalized predictive control and Internet of Things, its feature exists
In whether event detector meets some specified requirements using the output of the Information Authentication greenhouse process, if meeting these
Condition, then produce an event using sampling period Tf, so as to save secondary control activity；Otherwise, only produced in t=t+tmax
A raw control activity；Sampling time Tf based on change calculates control activity, therefore uses one group of GPC controller, each GPC
Controller one sampling time Tf=fTbase of correspondence, f ∈ [1, nmax].
7. the intelligent greenhouse irrigation system according to claim 1 based on generalized predictive control and Internet of Things, its feature exists
In control process uses GPC algorithms as feedback controller, uses one group of GPC controller, each one sampling of controller correspondence
Time Tf, f ∈ [1, nmax], each controller is realized classical by using corresponding discrete time model in the controller set
GPC algorithms；The target of GPC controllers is the cost function for minimizing multiple stages：
In formulaIt is j steps optimal before system exports prediction, t is current time, Δ uf(t+j-1) it is following control
Increment processed, w (t+j) is future reference track, includes sampling time Tf(t=kTf, k ∈ z+) in all signals；In addition, tuning
Parameter be respectively minimum prediction level, maximum predicted level,Controlled level,Following error, control weighting because
Sub- δfWith λf；Prediction level is defined asWithWeighted factor δf=1；GPC target is to calculate future
Control sequenceI.e. by minimizing J, the crops future close to w (t+j) is derived
Export yf(t+j)。
8. the intelligent greenhouse irrigation system according to claim 1 based on generalized predictive control and Internet of Things, its feature exists
In, by the sampling of event generator module management event, the module produces new event using two different conditions, if
One condition is changed into TRUE, then produces a new event, the current demand signal of process is transmitted to control module, according to the signal
Calculate a new control activity；
First condition determines the time of event using rising model, and uses asynchronous-sampling；If crops transpiration is total
Amount y (t) is more than the threshold value beta specified, then produces a new event, and y (t) computational methods are as follows：
In formulaIt is last event eiThe time of generation；If its value is set to 0 by y (t) more than β；If the total water lost
Amount reaches specified threshold β, then generates new control system event, it is clear that β value determines the frequency that the system event is produced；
Second condition is a time-based condition, and the condition is used to improve stability；The conditional definition is two controls
The calculating time difference of signal, is set to Tmax：
Use minimum sampling time slot TbaseSecond condition is checked, the sampling time T of change is usedf=fTbase, f ∈ [1,
nmax] all events are checked.
9. the intelligent greenhouse irrigation system according to claim 1 based on generalized predictive control and Internet of Things, its feature exists
In, in addition to signal reconstruction and resampling technology：Use the sampling time section T of changefA new control activity is determined, because
This, in order to realize GPC control algolithms, past value and the control signal of process variable must be that available (sampling time interval is
Tf), so needing to rebuild corresponding signal；
Assuming that a control signal is ub, every TbaseTime preserves control signal value using the variable；First, needed for calculating
Past information, more new signal ub；Assuming that producing a new event, cause a new sampling period Tf=fTbase, because
Use TbaseSample ubValue, so using ubThe average value of middle f value in the past is rebuildPast value；
I=P in formulau..., 1, j=k-1- (Pu- i) f,With PuIt is u respectivelyfPast value and required past value quantity；It is first
First, according to ubIn past value calculate new sampling time TfPast information, be stored in a variable and (be set to), use this
Information (is set to y with past the output of process dataf) new control activity is calculated, by the constant value for keeping two continuous events
To update ubSignal, uf(Tf)=ub(k)；
According to the above-mentioned GPC work rules based on event, the output of process variable monitored using asynchronous-sampling, in order to recover two
Information between continuous events, the information is rebuild using lagrange formula method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710350659.3A CN107015480B (en) | 2017-05-17 | 2017-05-17 | Intelligent greenhouse irrigation system based on generalized predictive control and Internet of things |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710350659.3A CN107015480B (en) | 2017-05-17 | 2017-05-17 | Intelligent greenhouse irrigation system based on generalized predictive control and Internet of things |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107015480A true CN107015480A (en) | 2017-08-04 |
| CN107015480B CN107015480B (en) | 2020-12-01 |
Family
ID=59450664
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710350659.3A Active CN107015480B (en) | 2017-05-17 | 2017-05-17 | Intelligent greenhouse irrigation system based on generalized predictive control and Internet of things |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107015480B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110432046A (en) * | 2019-09-17 | 2019-11-12 | 华北水利水电大学 | A kind of indoor Intelligent irrigation system of temperature |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH10116105A (en) * | 1996-10-11 | 1998-05-06 | Tokyo Electric Power Co Inc:The | Generalized predictive control system and denitration controller |
| CN1193784A (en) * | 1997-03-21 | 1998-09-23 | 陈志森 | Anti-faking label and manufacture thereof and identifier therefor |
| CN102998974A (en) * | 2012-11-28 | 2013-03-27 | 上海交通大学 | Multi-model generalized predictive control system and performance evaluation method thereof |
| CN105353656A (en) * | 2015-11-05 | 2016-02-24 | 宁波熵联信息技术有限公司 | Intelligent greenhouse irrigation control device based on fuzzy inference |
| CN106325071A (en) * | 2016-09-28 | 2017-01-11 | 哈尔滨工程大学 | Event driving based generalized predictive and self-adaptive supply ship navigation direction control method |
-
2017
- 2017-05-17 CN CN201710350659.3A patent/CN107015480B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH10116105A (en) * | 1996-10-11 | 1998-05-06 | Tokyo Electric Power Co Inc:The | Generalized predictive control system and denitration controller |
| CN1193784A (en) * | 1997-03-21 | 1998-09-23 | 陈志森 | Anti-faking label and manufacture thereof and identifier therefor |
| CN102998974A (en) * | 2012-11-28 | 2013-03-27 | 上海交通大学 | Multi-model generalized predictive control system and performance evaluation method thereof |
| CN105353656A (en) * | 2015-11-05 | 2016-02-24 | 宁波熵联信息技术有限公司 | Intelligent greenhouse irrigation control device based on fuzzy inference |
| CN106325071A (en) * | 2016-09-28 | 2017-01-11 | 哈尔滨工程大学 | Event driving based generalized predictive and self-adaptive supply ship navigation direction control method |
Non-Patent Citations (2)
| Title |
|---|
| ANDRZEJ PAWLOWSKI等: "Generalized Predictive Control With Actuator Deadband for Event-Based Approaches", 《IEEE TRANSACTIONS ON INDUSTRAL INFORMATICS》 * |
| 李富强: "基于事件触发的网络化系统H∞/L∞控制研究", 《中国博士学位论文全文数据库 信息科技辑》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110432046A (en) * | 2019-09-17 | 2019-11-12 | 华北水利水电大学 | A kind of indoor Intelligent irrigation system of temperature |
| CN110432046B (en) * | 2019-09-17 | 2022-03-01 | 华北水利水电大学 | Intelligent irrigation system in greenhouse |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107015480B (en) | 2020-12-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| González-Amarillo et al. | An IoT-based traceability system for greenhouse seedling crops | |
| Lea-Cox et al. | Advancing wireless sensor networks for irrigation management of ornamental crops: An overview | |
| CN108781926B (en) | Greenhouse irrigation system and method based on neural network prediction | |
| Kuijpers et al. | Lighting systems and strategies compared in an optimally controlled greenhouse | |
| CN110069032B (en) | Eggplant greenhouse environment intelligent detection system based on wavelet neural network | |
| Suciu et al. | Efficient IoT system for precision agriculture | |
| CN113994868B (en) | Automatic irrigation method and system based on plant growth cycle | |
| Sivagami et al. | Automated irrigation system for greenhouse monitoring | |
| Van Straten et al. | The significance of crop co-states for receding horizon optimal control of greenhouse climate | |
| CN110119767A (en) | A kind of cucumber green house temperature intelligent detection device based on LVQ neural network | |
| CN107329420B (en) | A kind of intelligence potting house keeper's control system | |
| CN106773720A (en) | A kind of warmhouse booth environment automatic control system | |
| Zapata et al. | Field test of an automatic controller for solid-set sprinkler irrigation | |
| CN107121927A (en) | A kind of irrigation system based on generalized predictive control | |
| CN107015480A (en) | A kind of intelligent greenhouse irrigation system based on generalized predictive control and Internet of Things | |
| CN116310815B (en) | Disease and pest early warning and self-learning method suitable for intelligent tea garden | |
| CN116449897A (en) | Greenhouse environment optimal control method, server and system | |
| Murugan et al. | Arduino Based Wireless Sensor Network Systems for Agriculture | |
| Daniels et al. | Optimal Control for Indoor Vertical Farms Based on Crop Growth | |
| Pawłowski et al. | Event-based control for a greenhouse irrigation system | |
| Gu et al. | Decision support system for irrigation scheduling based on Raspberry-Pi embedded with neural network | |
| Ghazouani et al. | A drip irrigation prediction system in a greenhouse based on long short-term memory and connected objects | |
| Senavirathne et al. | Greenhouse Automation with Artificial Intelligence and Industry 4.0 Integration | |
| Singhal et al. | Discretization schemes Comparison for the Greenhouse Temperature Model | |
| CN116224870B (en) | Intelligent planting environment control system and method based on multi-network fusion and plant photosynthesis feedback |
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 |