CN105807811A - Remote greenhouse temperature control system based on WI-FI - Google Patents

Abstract

The invention provides a remote greenhouse temperature control system composed of a management unit, a wireless transmission unit and a control unit. The management unit accesses the control unit through the wireless transmission unit and monitors the real-time temperature state of a greenhouse. The control unit adjusts the temperature of the greenhouse in real time through a control algorithm based on single neural element PID. The greenhouse temperature object generally fluctuates in a crop tolerable temperature range, and at the moment, a heating element is controlled to realize a temperature adjustment function; and only when the greenhouse temperature object exceeds a highest crop tolerable temperature, the temperature reducing element is controlled to carry out water spraying for temperature reduction. According to the invention, a WI-FI network is used as a transmission medium, the monitoring range is flexibly enlarged or shortened by the multi-layer distributed wireless network, the number of required greenhouse temperature monitoring nodes can be conveniently increased or reduced, and the greenhouse temperature control efficiency is simultaneously improved; in addition, a neural network calculation principle and a routine PID control algorithm are combined, the control efficiency is improved, and the system performance is optimized.

Description

A kind of long-range greenhouse temperature monitoring system based on WI-FI

Technical field

The present invention relates to a kind of greenhouse temperature monitoring system, particularly relate to a kind of based on WI-FI radio data transmission method With greenhouse temperature monitoring system based on Research on algorithm of single neuron adaptive PID.

Background technology

Along with the development of Chinese national economy, the living standard of the people improves day by day, winter booth vegetable market day by day Expanding, especially northern area uses Plastic canopy culture vegetable in cold winter, more embodies economic worth.Greenhouse temperature control System, i.e. according to the needs of crop growthing development, by sensor technology, microcomputer and singlechip technology and artificial intelligence Energy technology realizes greenhouse temperature and is automatically adjusted, and makes crops in the anti-season being not suitable for growth promoter, it is thus achieved that than outdoor growth More excellent environmental condition.Agricultural greenhouse is directly in open-air atmosphere, and usual area is relatively big, and conventional threshold values control system is in reply During this complex process object of greenhouse temperature, there is certain limitation, may result in temperature control not in time, control performance The best.

Current greenhouse temperature control system many employings wired mode carries out data transmission.Wired network system has movement Property is poor, dumb, expansibility is poor, networking and the shortcoming such as maintenance is not convenient.Along with the development of present radio network technique, wireless The stability of network and real-time have had the biggest improvement, and its performance is enough to ensure that control system is run with security and stability.

Summary of the invention

The technical problem to be solved in the present invention is to provide a kind of temperature control in time, accurately, and based on wireless data transmission Greenhouse temperature monitoring system.

In order to solve above-mentioned technical problem, the technical scheme is that a kind of long-range booth temperature based on WI-FI of offer Degree monitoring system, it is characterised in that: being made up of administrative unit, wireless transmission unit and control unit, administrative unit is by wireless Transmission unit access control unit, monitors greenhouse temperature real-time status；Control unit is calculated by control based on single neuron PID Greenhouse temperature is regulated by method in real time.

Preferably, described administrative unit is user side PC, on the one hand for arranging the control desired value of greenhouse temperature, another Aspect is for monitoring the real-time temperature values of greenhouse temperature.

Preferably, described wireless transmission unit is made up of multiple WI-FI modules, receives joint including MST admin site, AP Point, RST relay, RAP relay reception node, CST control website, and all information between MST, AP, RST, RAP, CST pass Defeated mode is for being wirelessly transferred, and described administrative unit and MST wired connection, CST connects described control unit.

Preferably, described AP is for realizing the proper communication between described RST and described MST.

Preferably, described RAP and described RST collectively constitutes one group of repeater, is used for extending network transmission range.

Preferably, described control unit includes:

Greenhouse temperature object；

For the real-time temperature sensor detecting greenhouse temperature object data and passing to MMF monitoring server；

On the one hand receive greenhouse temperature object real time data, and be sent to administrative unit by wireless transmission unit；Another Aspect receives the control instruction of administrative unit, and controls greenhouse temperature pair according to described control algolithm based on single neuron PID The MMF monitoring server of elephant；

For receiving the instruction of MMF, and by heating element heater and/or cooling element, greenhouse temperature object is adjusted in real time The temperature of joint controls executor.

Preferably, described heating element heater uses resistance wire.

Preferably, described cooling element includes drip irrigation emitter and controls the electromagnetic valve of drip irrigation emitter opening and closing.

Preferably, described control unit also includes:

Be located at temperature sensor rear end, for filter out in greenhouse-environment interference signal LPF active low-pass filter Device；

Be located between LPF and MMF for realizing the ADI analog-digital converter that analog quantity and digital quantity convert；

Be located at MMF and temperature control between executor for the DAI digital-to-analogue conversion realizing digital quantity and analog quantity converts Device.

Preferably, described resistance wire, electromagnetic valve are connected with described DAI, by regulation resistance wire both end voltage or electricity The opening and closing of magnet valve door realizes the control to greenhouse temperature object.

Preferably, described MMF is provided with software filter, and software filter is by die filter and moving average filter Composition.

Preferably, described greenhouse temperature object typically fluctuates in the range of crop tolerable temperature, now by based on single god Control algolithm control heating element heater through unit PID, it is achieved the function of regulation temperature；And if only if, and greenhouse temperature object exceedes crop During tolerance maximum temperature, carry out spray cooling by controlling cooling element.

Preferably, the frequency-domain model of described greenhouse temperature object is:

$\left\{\begin{array}{c}Y\left(s\right)=G\left(s\right)U\left(s\right)\\ G\left(s\right)=\frac{1}{s^2+2s+1}{e}^{-10s}\end{array}\right.$

Wherein, Y (s) is the output signal of canopy temperature sensor；U (s) is control based on single neuron PID algorithm The control signal of device output；G (s) is the transmission function of greenhouse temperature object, e^-10sFor the purely retarded ring during greenhouse temperature Joint；

The frequency-domain model of greenhouse temperature object is converted to discrete time-domain model:

Y (k)=1.8097y (k-1)-0.8187y (k-2)+0.00468u (k-1)+0.00438u (k-2)

Wherein, k is sample sequence number；Y (k) is current time greenhouse temperature sampled value；Y (k-1), y (k-2) are respectively and work as Greenhouse temperature sampled value in front 1,2 sampling periods in front moment；U (k-1), u (k-2) are respectively first 1,2 of current time The output valve of sampling period internal controller.

Preferably, the expression formula of described control algolithm based on single neuron PID is:

$u\left(k\right)=K_{p}e\left(k\right)+K_t{T}_s\underset{m=0}{\overset{k}{\Σ}}e\left(m\right)+K_d\frac{e\left(k\right)-e(k-1)}{T_s}$

Wherein, u (k) is controller output；E (k) is current time error；E (k-1) is 1 sampling before current time Error in cycle；K is sample sequence number；K_pFor proportionality constant；K_iFor integral constant；K_dFor derivative constant；T_sFor sampling week Phase；

Controller model based on single neuron PID is as follows:

Neuron inputs: e (k),e(k)-e(k-1)

Neuron exports: u (k)

Error: e (k)=r (k)-y (k)

Performance indications:

Gradient descent method is used to carry out weighed value adjusting:

K_p(k+1)=K_p(k)+ΔK_p；K_i(k+1)=K_i(k)+ΔK_i；K_d(k+1)=K_d(k)+ΔK_d

Wherein, Δ e (k)=e (k)-e (k-1)；U (k) is controller output；R (k) is desired temperature；Y (k) is current Moment object temperature value；μ is learning rate.

Preferably, the step of described control algolithm based on single neuron PID is as follows:

Step 1, set each initial value, the given sampling period；

Step 2, calculating error and learning rate；

Step 3, recursive calculation weights；

Step 4, computing controller export.

Step 5, repetition step 1～step 4, until error is in tolerance interval or beyond frequency of training restriction.

Utilize the above-mentioned neural principle calculated that PID controller is carried out parameter tuning based on mononeuric PID controller, Regulatory PID control system weak point when in the face of complex object can be made up, improve and control effect, optimize systematic function. Adjust controller output according to control algolithm based on single neuron PID, control heating element heater, make greenhouse temperature in less overshoot On the premise of be rapidly achieved predetermined temperature stationary value；And when greenhouse temperature is too high exceed crop tolerance maximum temperature time then Turn off heating element heater, open drip irrigation electrically operated valve, spray cooling.

Due to the fact that and take above technical scheme, it has the advantages that

1, the present invention uses WI-FI network as transmission medium, layouts simple, and the duration of arranging net is short, after network is destroyed easily Recover.

2, the present invention uses multilayer distributed wireless network, can expand or shrink monitoring range, convenient increase and decrease neatly The greenhouse temperature monitor node quantity needed, improves greenhouse temperature control efficiency simultaneously.

3, neural computing principle is combined by the present invention with regulatory PID control algorithm.Play neutral net height intelligence The advantage of energy property, makes up regulatory PID control system weak point when in the face of complex object, it is achieved during PID controls Parameter self-tuning function, improves control efficiency, optimizes systematic function.

Accompanying drawing explanation

Based on WI-FI the long-range greenhouse temperature monitoring system block diagram that Fig. 1 provides for the present embodiment；

Fig. 2 is greenhouse temperature monitoring system structural representation.

Fig. 3 is that distributed wireless controls network node schematic diagram.

Fig. 4 is greenhouse temperature monitoring system control principle block diagram.

Fig. 5 is based on mononeuric pid control algorithm theory diagram.

Fig. 6 is greenhouse temperature monitoring system tracking characteristics figure.

Fig. 7 is greenhouse temperature monitoring system noiseproof feature figure.

Detailed description of the invention

Below in conjunction with specific embodiment, the present invention is expanded on further.Should be understood that these embodiments are merely to illustrate the present invention Rather than restriction the scope of the present invention.In addition, it is to be understood that after having read the content that the present invention lectures, people in the art The present invention can be made various changes or modifications by member, and these equivalent form of values fall within the application appended claims equally and limited Scope.

Based on WI-FI the long-range greenhouse temperature monitoring system block diagram that Fig. 1 provides for the present embodiment, described based on WI- The long-range greenhouse temperature monitoring system of FI uses Hierarchical network topological structure, carries out wireless data transmission by WI-FI, it is achieved Real-time monitoring and control to greenhouse temperature, uses greenhouse temperature controller based on single neuron PID algorithm simultaneously, optimizes control Dynamic and the steady-state characteristic of system processed.

Based on WI-FI the long-range greenhouse temperature monitoring system that the present embodiment provides is by hardware system and is implanted to MMF mould Software two parts of block are constituted.Hardware system is made up of administrative unit, wireless transmission unit and control unit.Administrative unit is for using Family end PC；Wireless transmission unit is used for connection management unit and control unit, is made up of several WI-FI modules；Control unit By monitoring server (MMF), analog-digital converter (ADI), digital to analog converter (DAI), active low-pass filter (LPF), temperature control Executor processed (heating element heater and cooling element) and greenhouse temperature object composition.Software section is controlled plan by software filtering, selection Slightly form with control algolithm based on single neuron PID.

This working-flow is: on the one hand, the wireless network access MMF (prison that user side PC consists of WI-FI module Control server) module, monitor greenhouse temperature real-time status；On the other hand, MMF module is according to control based on single neuron PID Algorithm adjusts controller output, controls heating element heater, makes greenhouse temperature be rapidly achieved on the premise of less overshoot and preset Temperature stabilization value；And when greenhouse temperature is too high exceed crop tolerance maximum temperature time then turn off heating element heater, open and drip Fill electrically operated valve, spray cooling.

User side PC, on the one hand for arranging the control desired value of described greenhouse temperature object, on the other hand is used for monitoring The real-time temperature values of described greenhouse temperature object.

Wireless transmission unit is by MST (admin site), AP (receiving node), CST (control website), and (relay reception saves RAP Point), RST (relay) forms.All information transmission sides between MST, AP, CST, RAP, RST in wireless transmission unit Formula is for being wirelessly transferred.

Described MST is for connecting described administrative unit and described wireless transmission unit.The company of user side PC Yu MST The mode of connecing is wired connection.

Described AP is for realizing the proper communication between described RST and described MST.

Described RAP and described RST collectively constitutes one group of repeater, is used for extending network transmission range.

Described CST is for connecting described wireless transmission unit and described control unit.

Described MMF mono-aspect is for receiving the data of described ADI, on the other hand according to described based on mononeuron The control algolithm of PID controls described greenhouse temperature object.The mode of the MMF described in user side pc access is for input phase by webpage The IP address answered accesses.

Described ADI and described DAI is for realizing the mutual conversion of digital quantity and analog quantity.

Described LPF is for filtering out the interference signal in greenhouse-environment.

Greenhouse temperature monitoring system is by being positioned at the user side PC of Control Room, wireless network, temperature sensor, heating element heater (resistance wire), cooling element (drip irrigation emitter, electromagnetic valve), greenhouse temperature object composition, its structure is as shown in Figure 2.Temperature passes Sensor is connected with described active low-pass filter, and described resistance wire, Irrigation Electromagnetic Valve door are connected with described DAI.Booth temperature Degree control system realizes the control to greenhouse temperature object by the opening and closing of regulation resistance wire both end voltage or Irrigation Electromagnetic Valve door System.

Wireless network is distributed wireless networks, and this network structure is by being distributed in different location and by multiple functional modules The node interconnection of set forms.Wherein, each node comprises a controller module MMF, ADI module, a DAI Module, a heating element heater and a temperature sensor composition, its structure is as shown in Figure 3.ADI, DAI and MMF are integrated in one In device, and directly it is connected by wired mode with LPF.

Software filtering and being realized by code implant in described MMF based on mononeuric pid control algorithm.

Described software filtering uses moving average filter and limit filtration.

The control strategy of greenhouse temperature monitoring system includes selector and controller based on single neuron PID algorithm, its Structure is as shown in Figure 4.Described greenhouse temperature object typically fluctuates in the range of crop tolerable temperature, now by controlling heating unit Part realizes the function of regulation temperature.Drip irrigation is opened by selector during and if only if greenhouse temperature exceedes crop tolerance maximum temperature Electromagnetic valve carries out spray cooling.

Described controller based on single neuron PID algorithm realizes the function of regulation temperature by controlling heating element heater.Should Controller is made up of learning algorithm, neuron and conventional PID controller, and its structured flowchart is as shown in Figure 5.This controller is with nerve Computational theory is foundation, carries out pid parameter and adjust based on digital pid control law.Wherein, tri-parameters of P, I, D are The weights of input, systematic error and real system are output as training sample.By these neuron models are constantly trained Obtain optimal P, I, D parameter, described greenhouse temperature control system is carried out real-time optimum control.

The control algolithm of described Traditional PID unit is the positional PID control calculation in Digital PID Algorithm, control The expression formula of algorithm processed is:

$u\left(k\right)=K_{p}e\left(k\right)+K_t{T}_s\underset{m=0}{\overset{k}{\Σ}}e\left(m\right)+K_d\frac{e\left(k\right)-e(k-1)}{T_s}$

Wherein, u (k) is the output of PID controller；E (k) is current time error；E (k-1) is before current time 1 Error in sampling period；K is sample sequence number；K_pFor proportionality constant；K_iFor integral constant；K_dFor derivative constant；T_sFor sampling Cycle.

Described it is described as follows based on mononeuric PID controller model:

Neuron inputs: e (k),e(k)-e(k-1)

Neuron exports: u (k)

Error: e (k)=r (k)-y (k)

Performance indications:

Weighed value adjusting (gradient descent method):

K_p(k+1)=K_p(k)+ΔK_p；K_i(k+1)=K_i(k)+ΔK_i；K_d(k+1)=K_d(k)+ΔK_d

Wherein, Δ e (k)=e (k)-e (k-1)；U (k) is controller output；R (k) is desired temperature；Y (k) is current Moment object temperature value；μ is learning rate, and choosing of μ is related to convergence.

Described step based on Research on algorithm of single neuron adaptive PID is as follows:

1. set each initial value, given sampling period.

2. calculate error and learning rate.

3. recursive calculation weights.

4. computing controller output.

5. repeat the above steps is until error is in tolerance interval or beyond frequency of training restriction.

Utilize the above-mentioned neural principle calculated that PID controller is carried out parameter tuning based on mononeuric PID controller, Regulatory PID control system weak point when in the face of complex object can be made up, improve and control effect, optimize systematic function.

In the present embodiment, the frequency-domain model of greenhouse temperature object is:

$\left\{\begin{array}{c}Y\left(s\right)=G\left(s\right)U\left(s\right)\\ G\left(s\right)=\frac{1}{s^2+2s+1}{e}^{-10s}\end{array}\right.$

Wherein, Y (s) is the output signal of canopy temperature sensor；U (s) is control based on single neuron PID algorithm The control signal of device output；G (s) is the transmission function of greenhouse temperature object, e^-10sFor the purely retarded ring during greenhouse temperature Joint.It is converted into discrete time-domain model:

Y (k)=1.8097y (k-1)-0.8187y (k-2)+0.00468u (k-1)+0.00438u (k-2)

Wherein, k is sample sequence number；Y (k) is current time greenhouse temperature sampled value；Y (k-1), y (k-2) are respectively and work as Greenhouse temperature sampled value in front 1,2 sampling periods in front moment；U (k-1), u (k-2) are respectively first 1,2 of current time The output valve of sampling period internal controller.

For this model, carry out temperature based on Research on algorithm of single neuron adaptive PID and control.

Wherein, the initial value of pid parameter is set to: K_p=0.4, K_i=0.35, K_d=0.1.It should be noted that in system During realization, concussion occurs, in described single neuron PID controller model in order to avoid PID gain coefficient is excessive Weighed value adjusting algorithm needs after terminating to be normalized the weights after adjusting.Method is as follows:

K (k+1)=| K_p(k+1)|+|K_i(k+1)|+|K_d(k+1)|

${\hat{K}}_p(k+1)=\frac{K_p(k+1)}{K(k+1)};{\hat{K}}_i(k+1)=\frac{K_i(k+1)}{K(k+1)};{\hat{K}}_d(k+1)=\frac{K_d(k+1)}{K(k+1)}$

Wherein,For the actual weights brought into and calculate next time.

In order to verify the system tracking characteristics to greenhouse temperature setting value, initial time target setting temperature value is 15 DEG C, After post case temperature stabilization, when 40s, target setting temperature value is 25 DEG C.The tracking characteristics of this control system is as shown in Figure 6, permissible Finding out that described control system can relatively quickly reach target set point, in system reaches steady-state process, overshoot is the least, without shake Swinging, dynamic property is preferable.In order to verify the capacity of resisting disturbance of system, after system reaches stable state, when 45s in roc temperature mistake A disturbed value is added in journey.The noiseproof feature of control system is as shown in Figure 7, it can be seen that this system can be in relatively short period of time Inside return to steady statue.

Claims (10)

1. a long-range greenhouse temperature monitoring system based on WI-FI, it is characterised in that: by administrative unit, wireless transmission unit Forming with control unit, administrative unit passes through wireless transmission unit access control unit, monitors greenhouse temperature real-time status；Control Greenhouse temperature is regulated in real time by unit by control algolithm based on single neuron PID.

A kind of long-range greenhouse temperature monitoring system based on WI-FI, it is characterised in that: described pipe Reason unit is user side PC.

A kind of long-range greenhouse temperature monitoring system based on WI-FI, it is characterised in that: described nothing Line transmission unit is made up of multiple WI-FI modules, including MST admin site, AP receiving node, RST relay, RAP relaying Receiving node, CST control website, and all information transmission modes between MST, AP, RST, RAP, CST are for being wirelessly transferred, described Administrative unit and MST wired connection, CST connects described control unit.

A kind of long-range greenhouse temperature monitoring system based on WI-FI, it is characterised in that described control Unit processed includes:

Greenhouse temperature object；

For the real-time temperature sensor detecting greenhouse temperature object data and passing to MMF monitoring server；

On the one hand receive greenhouse temperature object real time data, and be sent to administrative unit by wireless transmission unit；On the other hand Receive the control instruction of administrative unit, and control greenhouse temperature object according to described control algolithm based on single neuron PID MMF monitoring server；

For receiving the instruction of MMF, and by heating element heater and/or cooling element, greenhouse temperature object is regulated in real time Temperature controls executor.

A kind of long-range greenhouse temperature monitoring system based on WI-FI, it is characterised in that: described control Unit processed also includes:

Be located at temperature sensor rear end, for filter out in greenhouse-environment interference signal LPF active low-pass filter；

Be located between LPF and MMF for realizing the ADI analog-digital converter that analog quantity and digital quantity convert；

Be located at MMF and temperature control between executor for the DAI digital to analog converter realizing digital quantity and analog quantity converts.

6. a kind of based on WI-FI the long-range greenhouse temperature monitoring system as described in claim 4 or 5, it is characterised in that: described MMF is provided with software filter, and software filter is made up of die filter and moving average filter.

A kind of long-range greenhouse temperature monitoring system based on WI-FI, it is characterised in that: described greatly Canopy temperature object typically fluctuates in the range of crop tolerable temperature, controls now by control algolithm based on single neuron PID Heating element heater, it is achieved the function of regulation temperature；During and if only if greenhouse temperature object exceedes crop tolerance maximum temperature, by control Cooling element processed carries out spray cooling.

A kind of long-range greenhouse temperature monitoring system based on WI-FI, it is characterised in that: described greatly The frequency-domain model of canopy temperature object is:

$\left\{\begin{array}{c}Y\left(s\right)=G\left(s\right)U\left(s\right)\\ G\left(s\right)=\frac{1}{s^2+2s+1}{e}^{-10s}\end{array}\right.$

Wherein, Y (s) is the output signal of canopy temperature sensor；U (s) is that controller based on single neuron PID algorithm is defeated The control signal gone out；G (s) is the transmission function of greenhouse temperature object, e^-10sFor the pure lag system during greenhouse temperature；

The frequency-domain model of greenhouse temperature object is converted to discrete time-domain model:

Y (k)=1.8097y (k-1)-0.8187y (k-2)+0.00468u (k-1)+0.00438u (k-2)

Wherein, k is sample sequence number；Y (k) is current time greenhouse temperature sampled value；When y (k-1), y (k-2) are respectively current Greenhouse temperature sampled value in front 1,2 sampling periods carved；U (k-1), u (k-2) are respectively front 1,2 samplings of current time The output valve of cycle internal controller.

A kind of long-range greenhouse temperature monitoring system based on WI-FI, it is characterised in that: described base Expression formula in the control algolithm of single neuron PID is:

$u\left(k\right)=K_{p}e\left(k\right)+K_i{T}_s\underset{m=0}{\overset{k}{\mathrm{\Σ}}}e\left(m\right)+K_d\frac{e\left(k\right)-e(k-1)}{T_s}$

Wherein, u (k) is controller output；E (k) is current time error；E (k-1) is 1 sampling period before current time Interior error；K is sample sequence number；K_pFor proportionality constant；K_iFor integral constant；K_dFor derivative constant；T_sFor the sampling period；

Controller model based on single neuron PID is as follows:

Neuron inputs: e (k),e(k)-e(k-1)

Neuron exports: u (k)

Error: e (k)=r (k)-y (k)

Performance indications:

Gradient descent method is used to carry out weighed value adjusting:

K_p(k+1)=K_p(k)+ΔK_p；K_i(k+1)=K_i(k)+ΔK_i；K_d(k+1)=K_d(k)+ΔK_d

Wherein, Δ e (k)=e (k)-e (k-1)；U (k) is controller output；R (k) is desired temperature；Y (k) is current time Object temperature value；μ is learning rate.

A kind of long-range greenhouse temperature monitoring system based on WI-FI, it is characterised in that: described base As follows in the step of the control algolithm of single neuron PID:

Step 1, set each initial value, the given sampling period；

Step 2, calculating error and learning rate；

Step 3, recursive calculation weights；

Step 4, computing controller export.

Step 5, repetition step 1～step 4, until error is in tolerance interval or beyond frequency of training restriction.