CN113994829B - LED light supplement lamp operation regulation and control method considering time shifting and cost factors - Google Patents
LED light supplement lamp operation regulation and control method considering time shifting and cost factors Download PDFInfo
- Publication number
- CN113994829B CN113994829B CN202111376082.6A CN202111376082A CN113994829B CN 113994829 B CN113994829 B CN 113994829B CN 202111376082 A CN202111376082 A CN 202111376082A CN 113994829 B CN113994829 B CN 113994829B
- Authority
- CN
- China
- Prior art keywords
- led
- illumination
- light
- cost
- neural network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01G—HORTICULTURE; CULTIVATION OF VEGETABLES, FLOWERS, RICE, FRUIT, VINES, HOPS OR SEAWEED; FORESTRY; WATERING
- A01G7/00—Botany in general
- A01G7/04—Electric or magnetic or acoustic treatment of plants for promoting growth
- A01G7/045—Electric or magnetic or acoustic treatment of plants for promoting growth with electric lighting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P60/00—Technologies relating to agriculture, livestock or agroalimentary industries
- Y02P60/14—Measures for saving energy, e.g. in green houses
Abstract
The invention relates to an operation regulation and control method of an LED light supplement lamp considering time shifting and cost factors, which comprises the following steps: s1: arranging a plurality of acquisition terminals in a plant greenhouse to acquire illumination acquisition data of multiple days to form a large amount of historical illumination sample data; then constructing a BP neural network prediction model, and obtaining target neural network parameters through training and correction of data for multiple days; s2: obtaining an illumination curve in the greenhouse of the next day according to the obtained BP neural network prediction model; then according to the illumination demand of different plants and different growth stages, obtaining the LED light supplement amount of the next day, and reversely deducing the total LED light supplement electric load of the next day; s3: constructing an optimized operation control model of the LED light supplement lamp; s4: and solving the optimal LED light supplement time and light supplement power of the next day by adopting a particle swarm algorithm with a penalty function. The LED supplementary lighting power can be directly and conveniently regulated, the effect of lowest comprehensive cost on the farmer side is achieved, and the comprehensive benefit of the farmer is effectively improved.
Description
Technical Field
The invention relates to the field of supplementary lighting control of electric power LED lamps, in particular to an operation regulation and control method of an LED supplementary lighting lamp considering time shifting and cost factors.
Background
Compared with a traditional incandescent lamp or an energy-saving lamp, the LED lamp has the remarkable advantages of long service life, small volume, low energy consumption, fixed wavelength and a cold light source, is an economic and environment-friendly light supplement lamp, is popularized and applied in large scale in planting modes such as plant factories and plant greenhouses at present, and gradually replaces traditional agricultural light supplement equipment.
At present, the silicon controlled dimming technology such as linear simulation dimming and PWM dimming technology is adopted to realize accurate dimming on light quality, light intensity and light cycle, so that the LED lamp has the technical conditions of real-time regulation and control. However, on one hand, the conventional LED light supplement lamp has the characteristic that the light supplement cost and the energy consumption cost account for high proportion of the total operation cost, and the energy consumption cost needs to be reduced as much as possible under the condition of meeting the light supplement requirement of crops; on the other hand, crops of different types and different growth stages have different optimal adaptation points to the light quantity, when the supplementary lighting is insufficient, the crops grow slowly, and when the lighting is excessive, the crops are inhibited from generating, so that the scientific daily supplementary lighting quantity needs to be determined; in addition, agricultural power consumption has "peak valley" price difference, and large-scale LED light filling lamp should be as far as possible in "valley" period power consumption, so not only be favorable to electric power system stable, but also can reduce peasant household cost, consequently, need the light filling period of rational planning LED lamp.
At present, there are some methods in the aspect of an LED fill-in light control method as follows: the method includes the steps that a terminal temperature sensor, a light sensor and a humidity sensor are built, data are uploaded to a mobile terminal in real time, then a computer realizes decision calculation to determine a light supplement scheme, an instruction is issued to a light supplement controller, and the red and blue light proportion of an LED is controlled to meet light supplement requirements of different plants; secondly, adjusting the color and the intensity of the tricolor LED lamp according to the environmental information and the fluorescence information on the basis; constructing an intelligent light supplementing terminal, an integrated controller and a mobile computer, wherein the communication is realized by adopting wireless ad hoc network communication, the WIFI of the intelligent light supplementing terminal supports AP and STATION modes, and the LED light supplementing power can be controlled in real time through the mobile terminal or the field integrated controller; based on the characteristic of illumination demand of plant growth, the photosynthetic efficiency is fed back by the plant fluorescent sensor at different temperatures, and the maximum photosynthetic efficiency of the plant can be guaranteed constantly by adopting PWM (pulse-width modulation) digital dimming to control the power of the LED light supplementing lamp in real time, accurately and in an energy-saving manner.
Although the light supplement system considers the requirement of plant growth on illumination and realizes the function of LED real-time intelligent light supplement, the energy consumption cost and the service life cost of farmers are not considered according to the time-of-use electricity price factor, and a reasonable light supplement control strategy is lacked, so that the LED light supplement optimization regulation and control aspect can be further researched.
Disclosure of Invention
The invention aims to provide a method for regulating and controlling the operation of an LED fill-in light, which directly and conveniently considers time-shifting and cost factors.
In order to solve the above problems, the method for regulating and controlling the operation of the LED fill light considering time-shifting and cost factors includes the following steps:
s1: arranging a plurality of acquisition terminals in a plant greenhouse to acquire illumination acquisition data of multiple days to form a large amount of historical illumination sample data; then constructing a BP neural network prediction model, predicting a next day illumination curve according to illumination data of the previous day, comparing the next day illumination curve with actual next day illumination data, correcting neural network parameters, and obtaining target neural network parameters through training and correction of multi-day data;
s2: inputting the illumination data of the previous day according to the obtained BP neural network prediction model to obtain an illumination curve in the greenhouse of the next day; then, according to the illumination demand of different plants and different growth stages, the difference value is made between the illumination demand of the plants and the predicted illumination cumulative quantity to obtain the LED light supplement quantity of the next day; then, reversely deducing the total LED supplementary lighting electric load of the next day through an LED supplementary lighting lamp power theoretical calculation model;
s3: constructing an optimized operation control model of the LED light supplement lamp according to the LED light supplement electric load of the next day by taking time-sharing electric energy cost, LED energy consumption cost and LED service life cost as targets;
s4: and solving the optimal LED light supplement time and light supplement power of the next day by adopting a particle swarm algorithm with a penalty function.
The BP neural network prediction model in step S1 is obtained by the following method:
s11: arranging a plurality of illumination sensors at crops in the greenhouse for collecting natural illumination intensity; the system comprises a light supplementing controller, an intelligent terminal, LEDs and a wireless communication module, wherein the light supplementing controller is connected with the light supplementing controller through the wireless communication module, is connected with the intelligent terminal and is used for transmitting sunshine data and receiving control instructions such as the LEDs;
s12: and (3) normalization calculation: the BP neural network prediction model takes the hourly illumination data of the previous day in the history as an input vector, the 24h illumination of the next day as an output quantity, the output quantity is compared with the actual illumination intensity in the history, and model parameters are fed back and corrected; wherein, the 24h illumination input quantity needs to be normalized according to the following formula:
in the formula (I), the compound is shown in the specification,
meaning that the value range is [0,1 ]]Per unit value of;
representing the amount of light at time t, in W/m 2 ；
Is the average annual illumination quantity and has the unit of W/m 2 ；
S13: setting parameters of the neural network:
setting the number of network input nodes tonNumber of hidden nodeslNumber of nodes of output layermThe connection weight of the input layer and the hidden layer isw ij The connection weight of the hidden layer and the output layer isw jk The threshold value of each neuron of the hidden layer isa j (j=1,2,…,l)，jThe number of nodes of the hidden layer; the threshold of each neuron of the output layer isb k (k=1,2,…,m)，kThe number of nodes of the output layer;
s14, hidden layer output calculation: according to the light input amount of 24 hours
Connection weight between input layer and hidden layerw ij And hidden layer thresholda j Computing hidden layer outputR j ；
In the formula (I), the compound is shown in the specification,f(x) Is a hidden layer excitation function;
s15: output layer output calculation: output according to hidden layerR j Connection weight of hidden layer and output layerw jk Output layer thresholdb k Calculating the predicted value of the BP neural networkM k ；
S16: calculating errors and updating network parameters:
outputting the results of the layerM k And the actual data vectorY k Set as the error between
Updating according to the error
Connection weight of each layer
And
and updating the network node threshold
And
the formula is as follows:
wherein λ is learning rate and its value range is [0,1 ]]；
And
respectively representing the node threshold values updated by the network training in the current round;
s17: when error occurs
Less than target error
If so, stopping the training, otherwise, continuing to perform the processes of S13-S16 until reaching the error
The condition is satisfied; therefore, data training in one day in history is realized, and a final BP neural network prediction model is obtained through historical data training in multiple days.
In the step S2, the illumination brightness is required according to the plantsDLIReverse-thrust LED fill-in lamp powerP LED The method comprises the following steps:
s21: determining the required daylight illumination according to the table 1 for different types and different stages of plantsDLI；
TABLE 1 LED light parameters for different colors
S22: according to predicted indoor illuminance per hourS(t) And the cumulative amount of light required by the plantDLIThe daily supplement quantity of the LED lamp is determined according to the following formula:
in the formula (I), the compound is shown in the specification,DLIthe daily supplement quantity of the LED lamp is in mol/(m) 2 Day(s))；DLI max AndDLI min cumulative daily illumination intensity for different types of plantsDLIUpper and lower limit values of (d);
s23: according to the daily supplementary lighting quantity of the LED lamp, the electric quantity of the LED lamp is reversely pushed according to the following formulaP LED ：
In the formula (I), the compound is shown in the specification,P LED the total electric energy of the LED fill light is in kWh;I e the average illumination of the LED lamp is in lx;Ais the total area in the greenhouse, and has the unit of m 2 ；Φ 0 Is luminous flux per lx unit area, and has unit lm/lx;C 1 andC 2 taking 1 for correction coefficients;η e the unit is Lm/W for the lighting effect of the light supplementing lamp;k e is an illuminance conversion factor in unit ofμmol/(m 2 S · klx) associated with the fill-in lamp device.
In the step S3, the optimization target in the model for controlling the optimal operation of the LED fill light is the lowest comprehensive cost of time-sharing electric energy cost, LED energy consumption cost and LED life cost, the constraint conditions are the inequality constraint of the maximum rated power of the LED and the equality constraint of the total fill light power, and the optimization variable is the fill light power of the LED light every hour the next day.
The time-sharing electric energy cost is calculated according to the following formula:
in the formula (I), the compound is shown in the specification,
the cost of supplementing the light and the electric quantity for the LED lamp,
is composed oftThe LED light supplement electric quantity of the hour is in kWh;
the unit is yuan/kWh for the agricultural time-of-use electricity price cost.
The LED energy consumption cost is obtained by fitting according to experimental data LED illumination intensity-LED energy consumption power to obtain an energy consumption model of LED illumination intensity as follows:
in the formula (I), the compound is shown in the specification,
the unit is the unit of the energy consumption power of the LED lamp;
is an indoor LED lamptThe total illumination at the moment is in lx;
、
parameters are fitted for energy consumption.
The life cost of the LED is calculated according to the following formula:
in the formula (I), the compound is shown in the specification,
the service life cost of the LED lamp on the same day is unit of Yuan;
is a firstnThe starting time of each LED lamp is hour;
is as followsnThe purchase cost of each LED lamp is yuan; n is the total number of the turned-on LED lamps, and the unit is one;
the service life of the LED lamp leaves factory is in unit of hour.
Compared with the prior art, the invention has the following advantages:
1. according to the invention, a BP neural network model is constructed through a large number of historical sunlight illumination curves, and the illumination curve of the next day is predicted. The method comprises the steps of collecting data with the light accumulation amount per hour being 1 moment in a plant greenhouse, forming a daily light curve by data with the light accumulation amount per hour being 24 moments, and then homogenizing the data to obtain input variables. And simultaneously setting various functions and parameters of the BP neural network, setting a target error, outputting 24h illumination quantity with a variable of the next day, and obtaining a neural network model through multiple times of training. The model can predict the illumination condition of the next day, and provides a basis for determining a subsequent LED light supplement regulation and control scheme.
2. The optimal daily cumulative illumination quantity of the planted crops is determined according to the requirements of different types of plants on the illumination quantity and the light color, and the difference value is made between the cumulative illumination quantity and the predicted illumination quantity of the next day to obtain the light supplement brightness of the LED of the next day. Particularly, the electric quantity of the LED light supplement lamp is reversely deduced by the daily light supplement quantity of the LED through the theoretical calculation model provided by the inventionP LED The light demand of crops is converted into the electric load demand, and the LED light supplement power can be quantitatively and accurately controlled, so that the light supplement amount can be accurately controlled.
3. The invention constructs an LED supplementary lighting optimization control model, comprehensively considers each cost of an LED supplementary lighting system for an optimization target, and particularly provides a calculation method for each cost. a) The electric energy cost of the time-of-use electricity price of farmers is fully considered, and the running cost is reduced as much as possible by controlling the LED light supplementing time period; b) the loss cost of the LED lamp is considered, the relation between the illumination brightness and the electric energy loss of the LED lamp in the PWM dimming mode is fitted to obtain a power loss model, and therefore the electric energy loss of the LED lamp is determined according to the illumination brightness correlation of the LED lamp; c) the optimization target also considers the service life cost of the LED, and the daily operation service life cost is calculated according to the purchase cost, the delivery service life limit and the sunlight illumination duration of the LED lamp.
4. The LED light supplementing optimization control model is solved by adopting an improved particle swarm algorithm. In particular, for the constraint conditions including equality constraint conditions: the total daily supplementary lighting power of the plant is constant, a penalty function H (x) is introduced, when the particles are optimized, if the daily supplementary lighting power is lower than the total demand of the plant, a large penalty function is added, and the optimizing direction of the particles is limited; meanwhile, the lowest cost is found by optimizing the model, and the supplementary lighting power is forced not to be too high, so that the superposition effect of the two is to enable the particle illumination power to approach the required power of the plants as much as possible.
5. The LED light supplementing power control system can directly and conveniently regulate and control the LED light supplementing power, achieves the effect of lowest comprehensive cost of a peasant household side, effectively improves the comprehensive benefit of the peasant household, and meets the requirements of 'peak clipping and valley filling' of a power grid side.
Drawings
The following describes embodiments of the present invention in further detail with reference to the accompanying drawings.
FIG. 1 is a flow chart of the present invention.
FIG. 2 is a BP neural network model training process according to the present invention.
Fig. 3 is a comparison of a predicted illumination curve and an actual illumination curve in an embodiment of the present invention.
Fig. 4 is a convergence process of particle swarm optimization solution in the embodiment of the present invention.
Fig. 5 is a scheme for adjusting the LED fill light of the next day in the embodiment of the present invention.
Detailed Description
An operation regulation and control method of an LED fill light considering time shifting and cost factors comprises the following steps:
s1: arranging a plurality of acquisition terminals in a plant greenhouse to acquire illumination acquisition data of multiple days to form a large amount of historical illumination sample data; and then constructing a BP neural network prediction model, predicting a next day illumination curve according to the illumination data of the previous day, comparing the next day illumination curve with the actual next day illumination data, correcting the neural network parameters, and obtaining the target neural network parameters through training and correction of multi-day data.
Wherein: as shown in fig. 2, the BP neural network prediction model is obtained as follows:
s11: arranging a plurality of illumination sensors at crops in the greenhouse for collecting natural illumination intensity; then be connected with the light filling controller through wireless communication module, the light filling controller is connected with intelligent terminal simultaneously for the transmission data in sunshine, receive control command such as LED simultaneously, the light filling controller is connected with each LED lamp, and LED adopts constant current drive for the light filling intensity of control.
S12: and (3) normalization calculation: the BP neural network prediction model takes the hourly illumination data of the previous day in the history as an input vector, the 24h illumination of the next day as an output quantity, the output quantity is compared with the actual illumination intensity in the history, and model parameters are fed back and corrected; wherein, the 24h illumination input quantity needs to be normalized according to the following formula:
in the formula (I), the compound is shown in the specification,
meaning that the value range is [0,1 ]]Per unit value of;
representing the amount of light at time t, in W/m 2 ；
Is the average annual illumination quantity and has the unit of W/m 2 ；
S13: setting parameters of the neural network:
setting the number of network input nodes tonNumber of hidden nodeslNumber of nodes of output layermThe connection weight of the input layer and the hidden layer isw ij The connection weight of the hidden layer and the output layer isw jk The threshold value of each neuron of the hidden layer isa j (j=1,2,…,l)，jNumber of nodes of hidden layer; the threshold of each neuron of the output layer isb k (k=1,2,…,m)，kNumber of output layer nodes;
s14, hidden layer output calculation: according to the light input amount of 24 hours
Connection weight between input layer and hidden layerw ij And hidden layer thresholda j Computing hidden layer outputR j ；
In the formula (I), the compound is shown in the specification,f(x) Is a hidden layer excitation function;
s15: output layer output calculation: output according to hidden layerR j Connection weight of hidden layer and output layerw jk Output layer thresholdb k Calculating the predicted value of the BP neural networkM k ；
S16: calculating errors and updating network parameters:
outputting the results of the layerM k And the actual data vectorY k Set as the error between
Updating according to the error
Connection weight of each layer
And
and updating the network node threshold
And
the formula is as follows:
wherein λ is learning rate and its value range is [0,1 ]]；
And
respectively representing the node threshold values updated by the network training in the current round;
s17: when error occurs
Less than target error
If so, stopping the training, otherwise, continuing to perform the processes of S13-S16 until reaching the error
The condition is satisfied; therefore, data training in one day in history is realized, and a final BP neural network prediction model is obtained through historical data training in multiple days.
S2: inputting the illumination data of the previous day according to the obtained BP neural network prediction model to obtain an illumination curve in the greenhouse of the next day; then, according to the illumination demand of different plants and different growth stages, the difference value is made between the illumination demand of the plants and the predicted illumination cumulative quantity to obtain the LED light supplement quantity of the next day; and then, the total LED supplementary lighting electric load of the next day is reversely deduced through an LED supplementary lighting lamp power theoretical calculation model.
According to the illumination brightness of the plantDLIReverse-thrust LED fill-in lamp powerP LED The method comprises the following steps:
s21: determining the required daylight illumination according to Table 1 for different types and stages of plantsDLI；
TABLE 1 LED light parameters for different colors
For very low light crops, the daily cumulative light exposureDLIShould be less than 5 mol/(m) 2 Day); for low-light crops, the daily cumulative illumination is 5-10 mol/(m) 2 Day); cumulative daily light exposure for medium-light cropsDLIIs 10 to 20 mol/(m) 2 Day); for high light crops, fatigueIntegrated light quantityDLIIs 20 to 30 mol/(m) 2 Day); cumulative daily light exposure for very high light cropsDLIIs 30 to 60 mol/(m) 2 Day). If the amount of light isDLIIf the amount of light is too large, the growth is inhibited.
S22: according to predicted indoor illuminance per hourS(t) And the cumulative amount of light required by the plantDLIThe daily supplement quantity of the LED lamp is determined according to the following formula:
in the formula (I), the compound is shown in the specification,DLIthe daily supplement quantity of the LED lamp is in mol/(m) 2 Day);DLI max andDLI min cumulative daily illumination intensity for different types of plantsDLIUpper and lower limit values of (d);
s23: according to the daily supplementary lighting quantity of the LED lamp, the electric quantity of the LED lamp is reversely pushed according to the following formulaP LED ：
In the formula (I), the compound is shown in the specification,P LED the total electric energy of the LED fill light is in kWh;I e the average illumination of the LED lamp is in lx;Ais the total area in the greenhouse, and has the unit of m 2 ；Φ 0 Is luminous flux per lx unit area, and has unit lm/lx;C 1 andC 2 taking 1 for correction coefficients;η e the unit is Lm/W for the lighting effect of the light supplementing lamp;k e is an illuminance conversion factor in unit ofμmol/(m 2 S · klx) associated with the fill-in lamp device.
S3: and constructing an optimized operation control model of the LED light supplement lamp according to the LED light supplement electric load of the next day by taking time-sharing electric energy cost, LED energy consumption cost and LED service life cost as targets.
The specific process is as follows:
the optimization target in the LED light supplement lamp optimization operation control model is the lowest comprehensive cost of time-sharing electric energy cost, LED energy consumption cost and LED service life cost, the constraint conditions are LED maximum rated power inequality constraint and total light supplement power equality constraint, and the optimization variable is LED light supplement power of each hour on the next day.
The time-sharing electric energy cost is calculated according to the following formula:
in the formula (I), the compound is shown in the specification,
the cost of supplementing the light and the electric quantity for the LED lamp,
is composed oftThe LED light supplement electric quantity of the hour is in kWh;
the unit is yuan/kWh for the agricultural time-of-use electricity price cost.
The LED energy consumption cost is obtained by fitting according to experimental data LED illumination intensity-LED energy consumption power to obtain an energy consumption model of LED illumination intensity as follows:
in the formula (I), the compound is shown in the specification,
the unit is the unit of the energy consumption power of the LED lamp;
is an indoor LED lamptThe total illumination at the moment is in lx;
、
parameters are fitted for energy consumption.
The life cost of the LED is calculated according to the following formula:
in the formula (I), the compound is shown in the specification,
the service life cost of the LED lamp on the same day is unit of Yuan;
is as followsnThe time length of the LED lamp is set as hour;
is as followsnThe purchase cost of each LED lamp is yuan; n is the total number of the turned-on LED lamps, and the unit is one;
the service life of the LED lamp leaves factory is in unit of hour.
S4: and solving the optimal LED light supplement time and light supplement power of the next day by adopting a particle swarm algorithm with a penalty function.
In order to optimize the constraint condition containing equality, the invention introduces a penalty function, when the accumulated illumination brightness of particles in the optimizing process is less than the total fill-in powerP LED A sufficiently large constant is introduced in the objective function to limit the direction of the particle's optimization.
In the formula, F is the lowest comprehensive cost and the unit is element; h (x) is a penalty function, and when the total electric energy of the LED fill-in light is lower than that in the solving processP LED When H (x) is 100 yuan, the supplementary lighting power is higher thanP LED Then, H (x) is taken as 0-membered.
An embodiment is shown in fig. 1, a method for adjusting and controlling operation of an LED fill-in light considering time-shifting and cost factors, comprising the following steps:
s1: an LED intelligent acquisition and light supplement system is built, and illumination data in a plant greenhouse every hour is acquired; arrange a plurality of illuminance sensors in plant greenhouse for gather natural illumination intensity, then be connected with the light filling controller through wireless communication module, the light filling controller is connected with intelligent terminal simultaneously, is used for transmitting the data in sunshine, receives control instructions such as LED simultaneously, and the light filling controller is connected with each LED lamp, and LED adopts constant current drive, is used for the light filling intensity of control.
And then constructing a BP neural network prediction model and predicting the indoor illumination curve of the next day. The method comprises the following specific steps:
(a) acquiring illumination intensity data of 31 days in a greenhouse in a certain northwest region in a whole month, wherein the data acquisition period is 1h, namely the dimension of the illumination data is 31 multiplied by 24, and carrying out normalization processing according to a formula (1), so as to obtain the processed illumination data as shown in table 2:
table 2 normalized lighting data
(b) The input variable in the neural network model is the normalized historical 24h illumination amount, namely a 24 x 1 vector, and the output vector is the 24h illumination amount of the next day. The training data is the light quantity data of the previous 30 days, the check data is the light data of the 31 th day, and the target error is
The value was taken to be 0.001. Number of input nodes using BP neural networknValue of 24, number of output layer nodesmThe value is 24, the hidden layer is 2 layers, and the initial connection weight is [0,1 ]]Inner random number, hidden layer thresholda j Initial 1, output layer neuron thresholdb k The prediction result is shown in FIG. 3 as 1.
Fig. 3 shows that the illumination prediction curve of the next day of the method is basically consistent with the actual illumination curve, the average illumination intensity error is 0.35%, and the daily accumulated illumination error is 2.64%, so that the accuracy of the prediction result of the illumination intensity is more than 95%, and the requirement of the next step of LED light supplement operation regulation and control based on the predicted illumination is met.
S2: selecting high-light plants with a ratio of red light to blue light of 4:1 as an example, wherein the daily cumulative light is 20-30 mol/(m) 2 Day), i.e. DLI max =30 mol/(m 2 Day), DLI min =20 mol/(m 2 Day). Total solar illumination in Table 1 is 2797.3W/m 2 In the order of 1W/m 2 Is equal to 4.57μmol/(s·m 2 ) Namely, the amount of sunlight is estimated to be 18.408mol/m 2 . The indoor illumination intensity s (of the glass greenhouse in certain northwest region)t) As shown in table 3:
TABLE 3 indoor sunlight intensity predicted in certain northwest region
Calculating the sunlight output of the LED according to the formula (9) to be 25mol/m 2 -18.408mol/m 2 =6.592 mol/m 2 。
The average illuminance of the LED is reversely deduced by the fill-in light quantity of the LED, and then the average illuminance is calculated according to a formulaI e =6.592/(132×0.75+145×0.25)=4.874×10 7 lx。
In this embodiment, the blue LED lamp coefficient is 145, and the red LED lamp coefficient is 132; total area in greenhouseAIs 500m 2 (ii) a Luminous flux per lx per unit area phi 0 Is 2.8 lm/lx; correction factorC 1 AndC 2 are all 1; light effect of light-compensating lamp of LED lampη e Is 100 lm/W; conversion coefficient of illuminancek e Relating to the light supplement lamp equipment, the unit is mu mol/(m) 2 S · klx). And (3) reversely deducing the fill light power of the LED lamp according to the formulas (10) to (11):P LED =4.874×10 7 × 500 × 2.8 × 1 × 1/(100 × 3600) =189.54 kWh. Therefore, the light supplement electric energy of the LEDs on the next day is 189.54 kWh.
S3: and constructing an LED light supplement lamp model with the lowest electric energy cost, energy consumption cost and service life cost. In the embodiment, the LED lamp is dimmed in a PWM mode.
Wherein: the relation between the LED energy consumption power and the LED illumination quantity is shown in the formula (5). Obtained by fitting experimental parameters
、
Respectively taking the values of 3.1 and 0.1 to obtain the formula (9):
wherein the single power of the LED lamp is 18W, the cost is 19 yuan/count, and the total area is 500m 2 The maximum supplementary lighting power of the greenhouse is 25kW, the electricity price at the peak of agriculture in the northwest region is 0.671 yuan/kWh, the electricity price at the ordinary time is 0.449 yuan/kWh, and the electricity price at the valley is 0.227 yuan/kWh. The rated life of each light is 10000 hours.
S4: and solving the optimal LED light supplement time and light supplement power of the next day by adopting a particle swarm algorithm with a penalty function.
The population size of the particle swarm is 50, the upper iteration limit is 1000 times, and the penalty function is H (x) =100, the optimization variable is fill light power per hour of 24 hours, the objective function is the minimum of the three types of cost synthesis, and the solution result is as follows:
and solving the optimization model by adopting an improved particle swarm algorithm. The particle swarm optimization method adopts a penalty function, and when each particle is optimized, the total fill light power of the particle is lower than that of the particleP LED The penalty function H (x) in the objective function takes the valueIs 100 yuan, when the total supplementary lighting power is higher thanP LED When the penalty function H (x) takes a value of 0 element, the optimization direction of the algorithm is limited. The population scale of the particle swarm is 50, and the upper limit of the iteration times is 1000 times; and finally, when the iteration result is not changed for 100 times continuously, the algorithm is regarded as convergence, and an optimal light supplement lamp operation strategy is output, or the light supplement lamp operation strategy is output when the iteration number upper limit is reached. The strategy includes fill light power every hour the next day. The solution results are shown in fig. 4. It can be seen from the figure that the lowest light supplement energy cost in the same day obtained by 192 times of iterative convergence is 71.823 yuan, wherein the electricity cost is 51.804 yuan, the LED lifetime cost is 20.007 yuan, the loss cost is 0.012 yuan, and the LED operation regulation scheme at the cost is shown in fig. 5.
As can be seen from fig. 5, in consideration of energy consumption cost, the power of the LED fill light of the next day is mostly turned on at the night off-peak time, so that the optimal comprehensive cost of the user side can be further realized, and the user is guided to scientifically and reasonably regulate and control the power of the LED fill light. If an LED light supplement real-time control strategy is adopted, the illumination quantity of the plants is always maintained at a constant value on the basis of daylight illumination in the daytime, light supplement is still carried out in the peak electricity price period in the mode, the energy consumption cost is far higher than that of the optimal regulation and control scheme, and the light supplement period and the light supplement electric quantity cannot be predictively regulated and controlled according to illumination conditions and the crop growth demand.
Claims (5)
1. An operation regulation and control method of an LED fill light considering time shifting and cost factors comprises the following steps:
s1: arranging a plurality of acquisition terminals in a plant greenhouse to acquire illumination acquisition data of multiple days to form a large amount of historical illumination sample data; then constructing a BP neural network prediction model, predicting a next day illumination curve according to illumination data of the previous day, comparing the next day illumination curve with actual next day illumination data, correcting BP neural network parameters, and obtaining target BP neural network parameters through training and correction of multi-day data;
the BP neural network prediction model is obtained according to the following method:
s11: arranging a plurality of illumination sensors at crops in the greenhouse for collecting natural illumination intensity; the system comprises a wireless communication module, a light supplement controller, an intelligent terminal, LED lamps, a light supplement controller and a light supplement module, wherein the wireless communication module is used for transmitting sunlight data and receiving LED lamp control instructions;
s12: and (3) normalization calculation: the BP neural network prediction model takes the hourly illumination data of the previous day in the history as an input vector, takes the 24h illumination of the next day as an output quantity, compares the output quantity with the actual illumination intensity in the history, and feeds back and corrects the model parameters; wherein, the 24h illumination input quantity needs to be normalized according to the following formula:
in the formula (I), the compound is shown in the specification,
meaning that the value range is [0,1 ]]Per unit value of;
representing the amount of light at time t, in W/m 2 ；
Is the average annual illumination quantity and has the unit of W/m 2 ；
S13: setting BP neural network parameters:
setting the number of BP neural network input nodes asnNumber of hidden nodeslNumber of nodes of output layermThe connection weight of the input layer and the hidden layer isw ij The connection weight of the hidden layer and the output layer isw jk The threshold value of each neuron of the hidden layer isa j (j=1,2,…,l)，jNumber of nodes of hidden layer; the threshold of each neuron of the output layer isb k (k=1,2,…,m)，kNumber of output layer nodes;
s14, hiddenReservoir output calculation: according to the light input amount of 24 hours
Connection weight between input layer and hidden layerw ij And hidden layer thresholda j Computing hidden layer outputR j ；
In the formula (I), the compound is shown in the specification,f(x) Is a hidden layer excitation function;
s15: output layer output calculation: output according to hidden layerR j Connection weight of hidden layer and output layerw jk Output layer thresholdb k Calculating the predicted value of the BP neural networkM k ；
S16: calculating errors and updating BP neural network parameters:
outputting the results of the layerM k And the actual data vectorY k Set as the error between
Updating according to the error
Connection weight of each layer
And
and updating BP neural network node threshold
And
the formula is as follows:
wherein λ is learning rate and its value range is [0,1 ]]；
And
respectively representing the node threshold values updated by the round of BP neural network training;
s17: when error occurs
Less than target error
If so, stopping the training, otherwise, continuing to perform the processes of S13-S16 until reaching the error
The condition is satisfied; therefore, one-day historical data training is realized, and a final BP neural network prediction model is obtained through multiple-day historical data training;
s2: inputting the illumination data of the previous day according to the obtained BP neural network prediction model to obtain an illumination curve in the greenhouse of the next day; then, according to the illumination demand of different plants and different growth stages, the difference value is made between the illumination demand of the plants and the predicted illumination cumulative quantity to obtain the LED light supplement quantity of the next day; then, reversely deducing the total LED supplementary lighting electric load of the next day through an LED supplementary lighting lamp power theoretical calculation model;
wherein the illumination brightness is determined according to the plant requirementDLIReverse-thrust LED fill-in lamp powerP LED The method comprises the following steps:
s21: determining the sunlight illumination required by growth according to different types and stages of plantsDLI；
For very low light crops, the daily cumulative light exposureDLIShould be less than 5 mol/(m) 2 Day); for low-light crops, the daily cumulative illumination is 5-10 mol/(m) 2 Day); cumulative daily light exposure for medium-light cropsDLIIs 10 to 20 mol/(m) 2 Day); cumulative daily light exposure for high light cropsDLIIs 20 to 30 mol/(m) 2 Day); cumulative daily light exposure for very high light cropsDLIIs 30 to 60 mol/(m) 2 Day);
s22: according to the predicted indoor illuminance per hourS(t) And cumulative amount of light required by the plantDLIThe daily supplement quantity of the LED lamp is determined according to the following formula:
in the formula (I), the compound is shown in the specification,DLIbeing LED lampsThe daily supplement quantity is in mol/(m) 2 Day);DLI max andDLI min cumulative daily illumination intensity for different types of plantsDLIThe upper and lower limit values of (2);
s23: according to the daily supplementary lighting quantity of the LED lamp, the electric quantity of the LED lamp is reversely pushed according to the following formulaP LED ：
In the formula (I), the compound is shown in the specification,P LED the total electric energy of the LED fill light is in kWh;I e the average illumination of the LED lamp is in lx;Ais the total area in the greenhouse, and has the unit of m 2 ；Φ 0 Is luminous flux per lx unit area, and has unit lm/lx;C 1 andC 2 taking 1 for correction coefficients;η e =100 Lm/W, which is the lighting effect of the fill-in light;k e is an illuminance conversion factor in unit ofμmol/(m 2 S · klx) associated with a fill-in lamp device;
s3: constructing an optimized operation control model of the LED light supplement lamp according to the LED light supplement electric load of the next day by taking time-sharing electric energy cost, LED energy consumption cost and LED service life cost as targets;
s4: and solving the optimal LED light supplement time and light supplement power of the next day by adopting a particle swarm algorithm with a penalty function.
2. The method for adjusting and controlling the operation of the LED fill-in light, considering the time-shifting and cost factors, as claimed in claim 1, wherein: in the step S3, the optimization target in the model for controlling the optimal operation of the LED fill light is the lowest comprehensive cost of time-sharing electric energy cost, LED energy consumption cost and LED life cost, the constraint conditions are the inequality constraint of the maximum rated power of the LED and the equality constraint of the total fill light power, and the optimization variable is the fill light power of the LED light every hour the next day.
3. The method as claimed in claim 2, wherein the method for adjusting and controlling the operation of the LED fill light considering time-shifting and cost factors comprises: the time-sharing electric energy cost is calculated according to the following formula:
in the formula (I), the compound is shown in the specification,
the cost of supplementing the light and the electric quantity for the LED lamp,
is composed oftThe LED light supplement electric quantity in hours is kWh;
the unit is yuan/kWh for the agricultural time-of-use electricity price cost.
4. The method for adjusting and controlling the operation of the LED fill-in light, considering the time-shifting and cost factors, as claimed in claim 2, wherein: the LED energy consumption cost is obtained by fitting according to experimental data LED illumination intensity-LED energy consumption power to obtain an energy consumption model of LED illumination intensity as follows:
in the formula (I), the compound is shown in the specification,
the unit is the unit of the energy consumption power of the LED lamp;
is an indoor LED lamptTotal illumination at a timeIn lx;
、
parameters are fitted for energy consumption.
5. The method as claimed in claim 2, wherein the method for adjusting and controlling the operation of the LED fill light considering time-shifting and cost factors comprises: the life cost of the LED is calculated according to the following formula:
in the formula (I), the compound is shown in the specification,
the service life cost of the LED lamp on the same day is unit of Yuan;
is as followsnThe starting time of each LED lamp is hour;
is as followsnThe purchase cost of each LED lamp is Yuan; n is the total number of the turned-on LED lamps, and the unit is one;
the service life of the LED lamp leaves factory is in unit of hour.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111376082.6A CN113994829B (en) | 2021-11-19 | 2021-11-19 | LED light supplement lamp operation regulation and control method considering time shifting and cost factors |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111376082.6A CN113994829B (en) | 2021-11-19 | 2021-11-19 | LED light supplement lamp operation regulation and control method considering time shifting and cost factors |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113994829A CN113994829A (en) | 2022-02-01 |
| CN113994829B true CN113994829B (en) | 2022-09-20 |
Family
ID=79929595
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111376082.6A Active CN113994829B (en) | 2021-11-19 | 2021-11-19 | LED light supplement lamp operation regulation and control method considering time shifting and cost factors |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113994829B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114845443B (en) * | 2022-07-04 | 2022-09-20 | 江苏南通二建集团讯腾云创智能科技有限公司 | Industrial data-based energy consumption optimization method for plant lighting equipment |
| CN116249246B (en) * | 2023-05-12 | 2023-09-05 | 南昌大学 | Lighting adjustment method, system, storage medium and equipment |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102138464A (en) * | 2010-12-14 | 2011-08-03 | 杭州汉徽光电科技有限公司 | Light supplement method for greenhouse plant cultivation |
| CN103237380A (en) * | 2013-03-15 | 2013-08-07 | 西北农林科技大学 | Method and system of intelligent light-environment controlling based on multi-factor coupling |
| CN203136275U (en) * | 2013-03-15 | 2013-08-14 | 西北农林科技大学 | Luminous environment intelligent regulation and control system based on multifactorial coupling |
| CN108983849A (en) * | 2018-07-12 | 2018-12-11 | 沈阳大学 | It is a kind of to utilize compound extreme learning machine ANN Control greenhouse method |
| CN109156196A (en) * | 2018-10-08 | 2019-01-08 | 中国农业大学 | A kind of plant light compensation control method and system for shining amount based on cumulative daylight |
| CN110334387A (en) * | 2019-05-09 | 2019-10-15 | 重庆大学 | A kind of indoor illumination predictor method based on BP neural network algorithm |
| CN110729764A (en) * | 2019-12-06 | 2020-01-24 | 国网江苏省电力有限公司南通供电分公司 | Optimal scheduling method for photovoltaic power generation system |
| CN210746303U (en) * | 2019-01-14 | 2020-06-16 | 内蒙古科技大学 | Warmhouse booth intelligence light filling monitored control system |
| CN111525547A (en) * | 2020-03-24 | 2020-08-11 | 云南电网有限责任公司临沧供电局 | Low-voltage intelligent management method based on optimal reactive compensation |
| CN112868435A (en) * | 2021-01-14 | 2021-06-01 | 同济大学 | NSGA-II-based blueberry greenhouse light and temperature coordination optimization method |
| CN113408206A (en) * | 2021-06-23 | 2021-09-17 | 陕西科技大学 | Indoor natural illuminance modeling method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111418381A (en) * | 2020-04-26 | 2020-07-17 | 南京格尼兹农业科技有限责任公司 | Dynamically-adjustable L ED plant light supplementing system and dynamic light adjusting method |
-
2021
- 2021-11-19 CN CN202111376082.6A patent/CN113994829B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102138464A (en) * | 2010-12-14 | 2011-08-03 | 杭州汉徽光电科技有限公司 | Light supplement method for greenhouse plant cultivation |
| CN103237380A (en) * | 2013-03-15 | 2013-08-07 | 西北农林科技大学 | Method and system of intelligent light-environment controlling based on multi-factor coupling |
| CN203136275U (en) * | 2013-03-15 | 2013-08-14 | 西北农林科技大学 | Luminous environment intelligent regulation and control system based on multifactorial coupling |
| CN108983849A (en) * | 2018-07-12 | 2018-12-11 | 沈阳大学 | It is a kind of to utilize compound extreme learning machine ANN Control greenhouse method |
| CN109156196A (en) * | 2018-10-08 | 2019-01-08 | 中国农业大学 | A kind of plant light compensation control method and system for shining amount based on cumulative daylight |
| CN210746303U (en) * | 2019-01-14 | 2020-06-16 | 内蒙古科技大学 | Warmhouse booth intelligence light filling monitored control system |
| CN110334387A (en) * | 2019-05-09 | 2019-10-15 | 重庆大学 | A kind of indoor illumination predictor method based on BP neural network algorithm |
| CN110729764A (en) * | 2019-12-06 | 2020-01-24 | 国网江苏省电力有限公司南通供电分公司 | Optimal scheduling method for photovoltaic power generation system |
| CN111525547A (en) * | 2020-03-24 | 2020-08-11 | 云南电网有限责任公司临沧供电局 | Low-voltage intelligent management method based on optimal reactive compensation |
| CN112868435A (en) * | 2021-01-14 | 2021-06-01 | 同济大学 | NSGA-II-based blueberry greenhouse light and temperature coordination optimization method |
| CN113408206A (en) * | 2021-06-23 | 2021-09-17 | 陕西科技大学 | Indoor natural illuminance modeling method |
Non-Patent Citations (5)
| Title |
|---|
| 基于可时移农业负荷的光伏智慧农业大棚微型能源网优化调度;王维洲等;《中国农业大学学报》;20180524(第06期);全文 * |
| 新组合模型在光电功率预测中的应用;王仕俊等;《数学的实践与认识》;20200323(第06期);全文 * |
| 日光温室内作物干物质积累预测新方法研究;张云鹤等;《林业机械与木工设备》;20050625(第06期);全文 * |
| 温室环境与叶类蔬菜生长态势模型研究;赵亚威等;《北方园艺》;20200915(第17期);全文 * |
| 面向设施农业的光伏-多形态储能联合优化调度控制;何欣等;《农业工程》;20200720(第07期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113994829A (en) | 2022-02-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113994829B (en) | LED light supplement lamp operation regulation and control method considering time shifting and cost factors | |
| CN107045367B (en) | A kind of greenhouse multiple-factor coordination energy-conserving and optimizing control method | |
| EP3277073B1 (en) | System and method of illuminating plants | |
| CA2914471C (en) | Method for controlling growth of a plant | |
| CN114208558B (en) | Light supplementing device and method based on thermoluminescent material | |
| EP3003009A1 (en) | Dynamic light recipe for horticulture | |
| CN111418381A (en) | Dynamically-adjustable L ED plant light supplementing system and dynamic light adjusting method | |
| CN111417238A (en) | developments illumination L ED plant lamp system | |
| TWI574608B (en) | Plant factory with light recipe verification platform | |
| CN109006002A (en) | A kind of intelligence plant LED dynamic light-supplementing system | |
| CN107894707A (en) | A kind of internet intelligent tracery wall sunlight analog control method and its system | |
| CN108717307B (en) | Multi-channel light supplement regulation and control method and system for light demand difference of whole plant of crop | |
| CN114868542B (en) | Environment processing method, device and system for strawberry planting | |
| NL2034401B1 (en) | Systems and methods for managing horticultural load shedding | |
| TWI601444B (en) | A plant-based upright LED light source drive system | |
| CN113507758A (en) | LED plant lighting system suitable for greenhouse | |
| Tomar | Sustainable photovoltaic based protective environment controlled farming technology as economy boosters for agro-sectors | |
| CN114609903A (en) | Energy-saving optimization control method and device for sunlight greenhouse | |
| CN115906913A (en) | Device power determination method and device for agricultural greenhouse system and electronic device | |
| WO2024023053A1 (en) | Method and system for illuminating plants with artificial light | |
| CN116528423A (en) | Full-spectrum intelligent plant growth lamp control method and system | |
| CN117015101A (en) | Self-adaptive dimming system and control method thereof | |
| Su et al. | The Optimization of Equipment in Greenhouse for Seedling Production in South Area of Guangxi Province |
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 |