CN108958329B - Drip irrigation water and fertilizer integrated intelligent decision-making method - Google Patents

Abstract

The invention discloses a drip irrigation water and fertilizer integrated intelligent decision-making method in the technical field of drip irrigation intelligent control. The intelligent decision-making method includes drip irrigation, fertilization decision-making method and database construction. The drip irrigation and decision-making comprehensively consider the impact of rainfall on the amount of irrigation, and the fertilization decision-making is carried out at the same time; and by building a database, the discrete point data is converted into spatial data to achieve intelligence Data mining; the decision-making method of the present invention comprehensively considers future meteorological conditions and crop growth requirements, realizes accurate output of irrigation and fertilization amount, and can easily acquire, manage and mine data, and improve the intelligence of irrigation and fertilization decision-making.

Description

An intelligent decision-making method for drip irrigation, water and fertilizer integration

technical field

The invention belongs to the technical field of drip irrigation intelligent control, in particular to a drip irrigation water and fertilizer integrated intelligent decision-making method.

Background technique

The development of drip irrigation is an important way to achieve agricultural water saving, high yield and high efficiency. It is of great significance for ensuring the safe supply of agricultural products, driving farmers out of poverty and becoming rich, and promoting agricultural modernization. Intelligent drip irrigation is based on the crop growth environment and crop physiological index information collected by sensors, analyzes the water and fertilizer requirements of crops, gives management parameters such as drip irrigation water and fertilizer amount, irrigation and fertilization time, and drives electronically controlled valves, water pumps, etc. to execute Drip irrigation technology, which automatically performs irrigation and fertilization operations by equipment or provides scientific and reliable irrigation and fertilization decision support for drip irrigation managers, has the advantages of more precise control, high water and fertilizer utilization efficiency, and low labor intensity, and has become an important direction for the development of drip irrigation.

At present, the control and execution part of the intelligent drip irrigation system has been relatively mature. There are many intelligent drip irrigation systems composed of sensor acquisition equipment, drive execution equipment, and transmission equipment on the market. At present, there are mainly the following problems: 1. In the process of intelligent decision-making and control of drip irrigation, water and fertilizer integration, the above-mentioned system lacks the decision-making process of water and fertilizer integration. The patent realizes the decision-making and automatic control of drip irrigation, water and fertilizer integration, but the decision-making process does not involve the calculation of the amount of fertilization, and does not consider the future precipitation, which is prone to repeated irrigation, resulting in the problem of low rainwater utilization; 2. Decision-making control process It is necessary to input, collect, and generate a large number of data in different categories and formats. It is necessary to build an effective database structure to effectively process and store the above data, so as to reduce the difficulty of data management in the intelligent drip irrigation system and facilitate the accumulation and further development of system data. Data mining has not yet seen relevant reports. For example, the patent application number CN201611127292.0 proposes an irrigation decision-making system and method based on an agricultural system model, which mentions an irrigation decision-making model library, but has not built a complete water and fertilizer integrated decision-making system. database structure. At the same time, decision-making parameters are obtained through experiments or calculations in specific regions, so the calculation parameters involved in decision-making are closely related to geographical locations, and their distribution in geographical locations is mostly discrete points. The management of geographically related parameters is also of great significance to the intelligent decision-making of water and fertilizer integration.

Therefore, it is urgent to build a decision-making system that integrates water and fertilizer decision-making and fully considers precipitation conditions. At the same time, it is necessary to build a suitable database, which can facilitate data acquisition, management and mining, and improve the intelligence of irrigation and fertilization decision-making.

SUMMARY OF THE INVENTION

The object of the present invention is to provide a kind of intelligent decision-making method of intelligent drip irrigation water and fertilizer integration, and the specific technical scheme is as follows:

An intelligent decision-making method for drip irrigation, water and fertilizer integration includes drip irrigation, a fertilization decision-making method and database construction;

The database construction is specifically as follows: using the Internet to establish a nationwide drip irrigation and fertilization parameter database, and the user directly selects parameters according to geographic locations or inputs local parameter values to obtain corresponding drip irrigation and fertilization decisions;

The drip irrigation and fertilization decisions are as follows: when the soil moisture reaches the irrigation water limit, according to the parameters input by the user and the parameters in the comprehensive database, directly determine whether to irrigate through the irrigation amount calculation model, and calculate the corresponding irrigation amount; the fertilization decision is carried out at the same time, Use the fertilization amount calculation model to determine whether to fertilize and calculate the fertilization amount; and send the decision result to the execution equipment of the intelligent control system.

The crop irrigation amount calculation in the drip irrigation decision-making takes the soil water status as the decision-making index. When the control system detects that the current soil moisture θ (soil volume moisture content, %) is lower than the irrigation water limit set in the irrigation system θ _min , start irrigation; the irrigation amount comprehensively calculates the weather forecast information and crop water consumption during the decision-making period. The decision-making period is selected according to the drought tolerance of the crop, and the basic range is 5 to 10 days; the weather forecast information during the decision-making period is preferably through the Internet. Obtain, including temperature forecast, precipitation forecast and weather phenomenon forecast; comprehensive geographic location parameters and weather forecast information calculation of crop water consumption;

The irrigation water calculation model in the drip irrigation decision-making is:

The effective rainfall during the decision-making period is calculated according to formula 1

P _e =αP Formula 1

Among them, P _e represents the effective rainfall, α is the rainfall infiltration coefficient, and P is the future rainfall obtained from the weather forecast information; when P is less than 5mm, α takes 0; when P is 5-50mm, α takes 1.0-0.8 , in this interval, α decreases with the increase of P; when P is greater than 50mm, α takes 0.7-0.8; ∑P _e represents the sum of multiple precipitations during the decision-making period.

(1) The cumulative effective rainfall during the irrigation decision-making period is ∑P _e = 0, that is, there is no rainfall during the decision-making period, and the irrigation amount is directly calculated and irrigated immediately. The calculation formula of the irrigation amount W ₁ is shown in Equation 2:

In the formula, W ₁ is the amount of irrigation water, in m ³ ; θ _max is the upper limit of irrigation, that is, the target soil volume moisture content for irrigation, %, dimensionless; z is the planned wet layer depth of the crop, in mm; p is the drip irrigation soil wetting ratio , %, dimensionless; M is the area of the decided irrigation area, the unit is m ² ;

(2) During the irrigation decision-making period, if the accumulated effective rainfall ∑P _e ≤ the maximum irrigation amount W ₁ ′, the irrigation is carried out immediately. The calculation strategy of the irrigation amount is to supplement the water that cannot be satisfied by the rainfall and the crop water consumption before the first rainfall date; The formula for calculating the amount of irrigation water W ₂ is shown in Equation 3:

∑P_e为预报期内单位灌溉面积下的有效降水之和，单位为

In the formula, W ₂ is the amount of irrigation water, in m ³ ; θ _max is the upper limit of irrigation, that is, the target soil volume moisture content for irrigation, %, dimensionless; z is the planned wet layer depth of the crop, in mm; p is the drip irrigation soil wetting ratio , %, dimensionless; M is the area of the decided irrigation area, the unit is m ² ; ∑ET _c is the sum of crop water consumption per unit irrigation area from now to before the first effective rainfall occurs, the unit is

∑P _e is the sum of the effective precipitation per unit irrigation area during the forecast period, the unit is

The maximum irrigation amount W ₁ ′ in the unit irrigation area is obtained by converting W ₁ ′=W ₁ ×1000/M, and the unit is mm.

The formula for calculating the daily water consumption of crops is shown in Equation 4:

ET _C = K _C ·ET ₀ formula 4

In the formula: ET _c is the daily water consumption of crops, the unit is mm; K _c is the crop coefficient, dimensionless, determined by the crop growth period, assuming that the seedling stage is 20 days, and the seedling stage K _c is 0.8, then it belongs to 20 days after sowing. In the seedling stage, K _c was taken as 0.8;

The reference crop water requirement is calculated according to the following formula 5

ET ₀ =aT _max +bR _S +c Equation 5

Among them, ET ₀ is the daily water requirement of the reference crops predicted by this model. According to the FAO56 document, ET ₀ refers to the evaporation on the short grassland with sufficient soil moisture, complete ground cover, normal growth, and neat height, which is called the reference Crop water demand, this value is only related to meteorological elements, so it can be directly calculated by meteorological parameters, also known as potential crop evapotranspiration, the unit is mm; T _max is the daily maximum temperature (°C); R _s is the actual solar radiation, MJ/m ² ·d; a, b, c are temperature coefficient, radiation coefficient, constant coefficient, dimensionless, and are obtained by multiple regression analysis according to the temperature, radiation and ET ₀ in a certain area for many years; in this calculation formula , the T _max data comes directly from the highest temperature forecast in the weather forecast. Since the weather forecast issued by the meteorological department for the public does not include the actual solar radiation R _s , it usually only contains the description of the weather phenomenon, such as sunny, sunny to overcast, etc. , in order to obtain the value of the actual solar radiation, the description of the weather phenomenon in the weather forecast is analyzed by the following methods. The specific formulas and methods are as follows:

R _S = βR _S0 Equation 6

R _S0 =(a _S +b _S )R _a Formula 7

为地理纬度(弧度，rad)；δ为太阳磁偏角(弧度，rad)；J为日序数；In the formula: R _s is the actual solar radiation obtained by conversion (MJ/m ² ·d); R _s0 is the clear sky radiation (MJ/m ² ·d), this value is only related to the latitude and date in a year ordinal, β conversion coefficient, the database of the present invention stores the corresponding conversion coefficient β of common weather phenomena such as "showers", "overcast", "cloudy to sunny", "cloudy", "showers to sunny", etc. in the weather forecast. The coefficients are shown in Table 1; R _a is the total solar radiation in the clear sky (MJ/m ² ·d); a _s is a regression constant, which represents the part of the radiation that reaches the earth's surface on a dark day, that is, when n=0; a _S + b _S represents In the radiation part reaching the earth's surface on a clear and cloudless day, a _s and b _s are obtained from local meteorological data. In the area where there is no long-series relationship between solar radiation and sunshine return in our country, a _S + b _S is usually taken as 0.75; G _SC is the sun The constant is 0.0820MJ/(m ² ·d); d _r is the reciprocal of the relative distance between the sun and the earth; ω _S is the solar hour angle (radian, rad), calculated according to the geographic latitude and the solar magnetic declination;

is the geographic latitude (radian, rad); δ is the solar magnetic declination (radian, rad); J is the daily serial number;

Table 1 Solar radiation conversion coefficient table for common weather phenomena

(3) During the irrigation decision-making period, the cumulative effective rainfall ∑P _e > the maximum irrigation amount W ₁ ′, in which the maximum irrigation amount W ₁ ′ is obtained by converting W ₁ ′=W ₁ ×1000/M, and the unit is mm; The use of precipitation to save irrigation water is divided into two categories according to whether the crops suffer from drought before the precipitation:

a. The moisture content of the crop root zone θ _i ≤ the soil moisture content of the crop withering point θ _dry , that is, before the effective rainfall occurs, the crops will be affected by drought and should be irrigated immediately to supplement the water consumed by the crops from the decision time to the occurrence of rainfall, To ensure that crops are not affected by severe drought, the calculation formula of irrigation amount W ₃ is shown in Equation 13:

In the formula, W ₃ is the amount of irrigation water, in m ³ ; p is the soil wetness ratio of drip irrigation, %, dimensionless; M is the area of the irrigation area to be decided, in m ² ; ∑ET _c is the time from now to the first occurrence The sum of crop water consumption per unit irrigated area before effective rainfall, the unit is mm;

b. Crop root zone moisture content θ _i > crop withered point soil moisture content θ _dry , no irrigation;

Soil volume moisture content θ _dry (%) is an indicator of crop drought. When it is lower than this value, crop growth will be irreversibly affected. This value is usually not allowed in crop growth management. This value can be controlled by the control system manager. Setting, preferably the soil moisture content value at the crop withering point is used as θ _dry ;

Among them, the water content θ _i in the root zone of the crop is calculated as shown in Equation 14:

In the formula, i is the number of days from the decision time to the precipitation, i=1, 2, 3...; z is the planned wet layer depth in the current growth period of the crop, in mm; θ _i is the planned wet layer volumetric water content on the ith day rate (%); θ is the current volumetric water content (%); ∑ET _c is the sum of the water consumption per unit irrigation area from now to before the first effective rainfall, the unit is mm.

The crop fertilization calculation is based on the calculation of nitrogen, phosphorus and potassium macroelements required by the crop with irrigation water as the carrier. The nutrient balance method is used to calculate the fertilization amount, and the nutrient application amount under the target yield is estimated from the difference between the crop's target yield and the soil fertilizer supply.

The calculation model of each fertilization amount in the fertilization decision is shown in Equation 15:

In the formula, F is the amount of fertilizer applied each time, the unit is kg; F _c is the total amount of nutrients required to calculate the target yield per mu at the decision point, the unit is kg; S is the available nutrient amount per mu of soil at the calculation decision point, the unit is kg; U is the seasonal utilization rate of fertilizer, usually taken as: nitrogen 30% _-45 %; phosphorus 5%-25%; potassium 40%-50%; The proportion of distribution in each growth period), %; f _i is the distribution ratio of top dressing during the growth period, that is, the proportion of the fertilizer required by the crop in this growth stage in the entire growth period, %; t _f is the experience fertilization frequency, that is, the growth period. The experience value of fertilization times is different for different crops and different growth periods;

Among them, the total amount of nitrogen, phosphorus or potassium F _c required to calculate the target yield per mu at the decision point is calculated according to formula 16:

In the formula, Y is the target yield of the calculation decision point, the unit is kg, and the value is manually input; F ₁₀₀ is the amount of nitrogen, phosphorus or potassium required for the economic yield of 100kg of the crop. For some crops, please refer to the following table 2:

Table 2 The amount of nitrogen, phosphorus and potassium required for 100kg economic output (kg)

Calculate the amount of available nutrients S per acre of soil at the decision point, that is, the amount of nutrients that the soil can provide for the crops to be absorbed and utilized in the current season, calculated according to formula 17:

S=Ne Equation 17

In the formula, N is the soil measurement value of nitrogen, phosphorus and potassium at the calculation decision point, that is, the soil background value, mg/kg; e is the soil effective nutrient correction coefficient, which is used to indicate the fraction of available nutrients in the soil that can be used by crops, Obtained through field experiments, dimensionless, %, if there is no experimental value, the default value is the empirical value: nitrogen 60%; phosphorus 30%, potassium 40%;

In each integrated irrigation with water and fertilizer, record the cumulative amount of fertilization ∑F and judge whether ∑F is greater than F· _{t f} . until the next growth period of the crop.

The database uses spatial interpolation method and data rasterization method to convert parameter point data used in drip irrigation and fertilization calculation models into spatial data. Reasonable and effective calculation parameters; after the database is formed, data cleaning can be performed on the parameters input by the user to remove abnormal values in the data.

Spatial interpolation is a method of transforming discrete point measurement data into a continuous data surface for comparison with the distribution patterns of other spatial phenomena. The data for the location points in the same area can be extrapolated from the known sample data points. In the decision-making process of irrigation and fertilization of the present invention, there are a large number of parameters related to spatial geographic location. These parameters are usually obtained through experiments or long-term data calculation in a certain place. These points are usually discrete, and the parameters will change with the geographic location. For example, the parameters a, b, and c in Equation 5 are calculated by analyzing the historical meteorological data of meteorological stations for many years. It is very difficult to obtain the parameter values of the adjacent locations of the meteorological station by actual calculation. Because the climate change has the characteristics of gentleness on the geographical scale, the adjacent point data can be estimated by the spatial interpolation method in a certain area;

The spatial interpolation method is preferably an inverse distance weighted interpolation method. This method obtains the unit value by averaging each unit value in the adjacent area. The weighting is inversely proportional to the distance. The closer the point is to the center of the prediction unit, the greater the weight. The calculation formula is shown in Equation 18:

In formula 18, Z(x ₀ ) is the predicted value at x ₀ ; n is the number of samples around the prediction point to be used in the prediction calculation process; λ _i is the weight of each sample point used in the prediction calculation process, the The value decreases as the distance between the sample point and the predicted point increases; Z( _xi ) is the measured value obtained at x _i ;

The calculation formula of the weight λ _i is shown in Equation 19:

In the formula, p is the exponential value, which significantly affects the result of the interpolation, and usually defaults to 2; d _i0 is the distance between the predicted point x ₀ and the known sample point x _i .

The data rasterization method preferably uses ArcGIS software, and the rasterized spatial data distribution map is stored in the database; after the database structure is formed, a certain amount of input and output data can be accumulated through a certain period of work. Using data mining methods such as support vector machines and random forests, machine learning training can be carried out, and new irrigation and fertilization models can be obtained, stored in the database, and called according to the needs of drip irrigation and fertilization decisions.

In the decision-making method, fertilization is carried out for each irrigation, and the cumulative amount of fertilization ∑F is recorded in each integrated irrigation of water and fertilizer. If after a certain irrigation and fertilization, the cumulative amount of fertilization ∑F≥F·t _f No fertilization is applied until the next growth period of the crops.

An intelligent decision-making system for realizing the intelligent decision-making method includes an input module, a crop module, a model parameter module and an output module;

Among them, the input module includes soil moisture data, irrigation area, moisture ratio, geographic location, meteorological data and sowing time;

The crop module includes growth period parameters and crop fertilization parameters. Growth period parameters include crop coefficient, upper limit of irrigation water, upper limit of irrigation water, crop drought index, planned wet layer depth and management recommendations. Crop fertilization parameters include the number of experienced fertilization and the distribution ratio of top dressing during the growth period. and top dressing ratio;

The model parameter module includes the calculation of the amount of irrigation and the amount of fertilizer; the output module includes the amount of fertilizer, the amount of irrigation and management advice.

The decision-making system also has a data cleaning function and a data mining function.

The beneficial effects of the present invention are:

(1) The intelligent decision-making method of drip irrigation, water and fertilizer integration of the present invention, according to the integrated requirements of intelligent drip irrigation, water and fertilizer, and comprehensively considering future meteorological conditions and crop growth requirements, a decision-making method for irrigation and fertilization under intelligent drip irrigation conditions is constructed, and can output Accurate irrigation and fertilization amount, output the decision results to the drip irrigation intelligent control system, improve the intelligence of the decision-making system;

(2) Based on the database of the drip irrigation, water and fertilizer integrated intelligent decision-making method, the present invention effectively classifies the data in the intelligent drip irrigation decision-making process, solves the problems of many data types and complex classification in the intelligent drip irrigation system, and proposes corresponding data analysis With the mining scheme, the decision model can be further optimized and the potential value of the data can be mined;

(3) The database of the present invention based on the drip irrigation, water and fertilizer integrated intelligent decision-making method has a data interface, which can not only be applied in the control system, but also can output the decision-making result as long as the input data required for decision-making is input. Drip irrigation and fertilization related equipment or software;

(4) According to the characteristics of spatial parameters in the intelligent drip irrigation water and fertilizer integration model, a decision parameter management method based on the difference method and data rasterization method is proposed. According to the geographic coordinates of the intelligent drip irrigation project, the corresponding The numerical value of the parameters solves the problem of parameter value selection in irrigation and fertilization decision-making and spatial distribution, and improves the accuracy and efficiency of parameter acquisition.

Description of drawings

Fig. 1 is the irrigation decision-making calculation flow chart in the decision-making system of the present invention;

FIG. 2 is an example diagram of the gridization result of temperature coefficient a in Formula 5 of the present invention, the left diagram is a distribution diagram of discrete data points before gridization, and the right diagram is a spatial data distribution diagram after gridization;

3 is a schematic diagram of the drip irrigation water and fertilizer integrated intelligent decision-making system provided by the present invention;

Detailed ways

The present invention provides an intelligent decision-making method for drip irrigation, water and fertilizer integration. The present invention will be further described below with reference to the accompanying drawings and embodiments.

Fig. 1 is the irrigation decision-making calculation flow chart in the decision-making system of the present invention, and the control system collects soil moisture parameters, and the meteorological information calculates the effective rainfall during the comprehensive decision-making period:

If the accumulated effective rainfall ∑P _e = 0 during the irrigation decision-making period, the condition (1) is satisfied, the irrigation amount W ₁ is calculated according to the formula 2, and the irrigation is performed immediately;

If the condition (1) is not satisfied, and the accumulated effective rainfall ∑P _e ≤ the maximum irrigation amount W ₁ ′ during the irrigation decision-making period, then the condition (2) is satisfied, the irrigation amount W ₂ is calculated according to formula 3, and the irrigation is performed immediately;

If the condition (2) is not met, that is, the cumulative effective rainfall ∑P _e > the maximum irrigation amount W ₁ ′ during the irrigation decision-making period, if the crop is drought, the irrigation amount W ₃ is calculated according to formula 13 and irrigated immediately; otherwise, no irrigation is required.

Figure 2 is an example diagram of the gridization result of the temperature coefficient a in Formula 5 of the present invention. Figure 2-a on the left is the distribution diagram of discrete data points before gridding, and Figure 2-b on the right is the spatial data distribution diagram after gridding. ; By converting point data into spatial data, to achieve the goal of quickly retrieving the required parameters through geographic coordinates. The parameters related to the geographic location in the present invention, including temperature coefficient, radiation coefficient, constant coefficient, nutrient background value, etc., all undergo a rasterization process to convert point data into spatial data, and users can quickly retrieve the required data through geographic coordinates. The parameters can also be directly retrieved from the database through the geographic location parameters for model calculation.

3 is a schematic diagram of the drip irrigation water and fertilizer integrated intelligent decision-making system provided by the present invention, which specifically includes information collection equipment and software interface, input module, crop module, model parameter module, output module and drip irrigation automatic control system execution equipment module; wherein drip irrigation water and fertilizer The integrated intelligent decision-making includes input module, crop module, model parameter module and output module.

The input module includes soil moisture data θ, irrigation area M, wetting ratio p, geographic location, meteorological data and sowing time. The specific value of soil moisture data θ is provided by the soil moisture monitor in real time; the moisture ratio p needs to be set by the user who uses the decision-making system, and is determined by the designer according to the soil texture of the crop and the project location when the drip irrigation project is designed. A fixed parameter; the role of the geographic location parameter is: when making decisions on irrigation and fertilization, the database can be used to retrieve parameters related to geographic location, and it is also used by the system to obtain the corresponding meteorological data from the Internet; the system directly determines the current crop according to the sowing time. Fertility period, and then determine the parameters of the fertile period part.

The function of the input module is to provide the user with a parameter input window. After the user inputs the parameters, the data is cleaned before being passed to the next step to remove abnormal values in the data, including invalid data exceeding the sensor range, negative values and other wrong data. At the same time, the same type of data format is unified to prevent intelligent decision-making errors caused by abnormal collected values. The principle of data cleaning is to compare whether the value conforms to the valid rules of the data. For example, the soil volumetric moisture content in a certain area is in the range of 5 to 40%. If the data obtained by the soil moisture sensor after filtering and correction is 50%, the data exceeds the valid rules. Invalid processing, and at the same time obtain data from irrigation plots whose data conforms to the rules for replacement.

The crop module includes growth period parameters and crop fertilization parameters. Growth period parameters include crop coefficient K _c , irrigation upper limit θ _max , irrigation water upper limit θ _min , crop drought index θ _dry , planned wet layer depth z and management recommendations, and crop fertilization parameters include Experience fertilization frequency t _f , top dressing ratio f _i and top dressing ratio f _z in growth period. Among them, the crop coefficient K _c , the upper limit of irrigation θ _max , the upper limit of irrigation water θ _min , the crop drought index θ _dry , and the planned wetting layer depth z are automatically converted by the decision-making system according to the sowing time parameters in the input module, and the specific values vary according to the growth period. The difference is also related to the type of crops; among them, "management advice" refers to the planting management advice that exists in the form of text and is applicable to different growth periods of crops. Crop fertilization parameters include experience fertilization times t _f , top dressing ratio f _i and top dressing ratio f _z , different crops have different growth stages, and are stored in the database by crop type and growth stage; the module parameters can be determined by the use decision system The user can input the settings by himself or choose to use the parameters corresponding to the corresponding crop growth period in the database. The specific parameter values in the database are manually collected and organized by personnel with relevant agricultural knowledge on the parameter data of different crops and different regions obtained from long-term field trials, and then entered into the database, or can be determined and entered into the database with reference to relevant literature.

The crop module is used to store the parameters related to the crop type, which can be directly retrieved and used by the decision-making system of the present invention during irrigation decision-making and fertilization decision-making, and stored in the database of the decision-making system of the present invention, which can be queried and used by users at any time. Experienced users You can also enter relevant experience values at any time, continuously improve the database, and achieve the purpose of sharing.

The model parameter module includes the calculation of irrigation amount and fertilization amount; the calculation part of irrigation amount stores the temperature coefficient a, radiation coefficient b, constant coefficient c and training irrigation model related to geographical location, temperature coefficient a, radiation coefficient b, constant coefficient c The specific values are calculated from the accumulated data of more than 700 meteorological stations across the country for many years, and the values of parameters a, b and c of more than 700 meteorological stations across the country have been obtained. The input geolocation parameters are directly called for decision calculation. The parameters required for the calculation of the fertilization amount are specifically the nutrient background value N, the effective nutrient correction coefficient e, the nutrient balance calculation parameters (except N and e, the various parameters used in the fertilization amount calculation model) and the training fertilization model. The background value N refers to the soil nutrient content status of the decided plot before sowing, which can be divided into soil nitrogen, phosphorus and potassium background values. The provided experience value is obtained; after the staff enters the above experimental value or experience value, the point data is formed, and the point data is converted into spatial data through data rasterization. The decision-making system directly calls the decision-making irrigation amount and fertilization amount according to the geographical location parameters input by the user. Irrigation and fertilization training model refers to the calculation model formed by data mining after long-term fertilization calculation for users to choose.

The output module includes fertilization amount F, irrigation amount W and management suggestions. The decision values of fertilization amount F and irrigation amount W are directly sent to the executive equipment of the intelligent control system such as water pumps, solenoid valves, fertilizer spreaders, etc. The management suggestions include to the drip irrigation system managers. Push agricultural machinery and agronomic management measures related to seedling management.

The decision-making method of the above-mentioned drip irrigation water and fertilizer integrated intelligent decision-making system is carried out according to the following steps:

(1) The user inputs soil moisture data, irrigation area, moisture ratio, geographic location, and sowing time at the user end, and the input module directly converts the geographic location parameters into meteorological data parameters based on the Internet, and removes abnormal values through data cleaning;

(2) When the soil moisture reaches the irrigation water limit, the irrigation decision is made: the decision-making system directly determines whether to irrigate through the irrigation amount calculation model according to the parameters input by the user and the parameters in the comprehensive database, and calculates the corresponding irrigation amount; at the same time, the fertilization decision is made. , and send the decision results to the execution equipment of the intelligent control system;

(3) The system automatically judges the growth period of crops, and pushes agricultural machinery and agronomic management measures related to seedling management to the drip irrigation system managers.

In the decision-making method, fertilization is performed for each irrigation, and the parameter data involved in the decision-making process are provided by the database in the present invention.

Example 1

Taking the decision-making system and database construction of an intelligent drip irrigation project in Tongzhou District, Beijing as an example, the project is planted with summer corn, each plot is equipped with sensors that can collect soil moisture, and the irrigation decision-making period is set to 10 days.

In the data collection step, 0% appears in the collected real-time soil moisture data. During the data cleaning process, it is found that this value is not within the normal data range, so this value is removed from the collected data, and the collected soil moisture and meteorological data , irrigated area, humidity ratio, summer corn sowing time, geographic location and other input data are stored into the database. When the soil moisture reaches the lower limit, the irrigation decision is made. The specific decision-making process is as follows:

In this example, the values of parameters a, b, and c located in Tongzhou District are searched in the spatial database through coordinates, and the specific value of parameter a is determined to be 0.454, the value of b is 0.1819, and the value of c is -0.5759.

The decision-making system automatically obtains 10 weather forecasts in the future from the website of the meteorological department according to the geographical location, and calculates the effective rainfall in the future. In this embodiment, the effective rainfall during the decision-making period is 0, and the irrigation amount W ₁ is:

The upper limit of irrigation θ _max is 30% (volume moisture content), soil moisture content θ is 19% (volume moisture content), z is 400 mm, p is 30%, M is 6670 m ² , and calculated by the above formula, the irrigation amount W ₁ =88.04m ³ , then the maximum irrigation amount in unit irrigation area W ₁ ′=13.2mm.

Enter the fertilization decision. Since the local unconditional output of test data, the local basic nutrient data is obtained. Taking nitrogen fertilizer decision-making as an example, the basic value of available nitrogen in Tongzhou District, Beijing is 160mg/kg, and the nutrient correction coefficient e is 60%, then The amount of nutrients S available per acre of soil at the decision point is:

S=0.15×160mg/kg×60%=14.4kg/mu

The user sets the target yield Y as 600kg, the nitrogen F ₁₀₀ required for the economic yield of 100kg of the crop is 3.4kg, and the total amount of nutrients Fc required to calculate the target yield per mu at the decision point is:

The seasonal utilization rate U of fertilizer is 35%; the data in the database is retrieved, and the top-dressing ratio f _z is 60%; the top-dressing fertilizer distribution ratio f _i in the growth period required for the seedling stage is 10%; the empirical value of the number of fertilization in the growth period t _f is 2, then the fertilization amount F is:

The output result is that 0.5kg of nitrogen fertilizer needs to be applied, which is carried out with irrigation, and the decision result is sent to the execution equipment of the intelligent control system; it is judged that the current summer corn is in the seedling stage, so the agricultural machinery and agronomic management related to the seedling stage management are pushed to the drip irrigation system manager. measure.

Claims (9)

1. A drip irrigation water and fertilizer integrated intelligent decision method is characterized in that the intelligent decision method comprises a drip irrigation and fertilizer application decision method and database construction;

the database construction specifically comprises the following steps: establishing a nationwide drip irrigation and fertilization parameter database by utilizing the internet, and directly selecting parameters or inputting local parameter values by a user according to the geographical position to obtain corresponding drip irrigation and fertilization decisions;

the drip irrigation and fertilization decision is as follows: when the soil moisture reaches the underwater irrigation limit value, directly determining whether irrigation is needed or not through an irrigation quantity calculation model according to parameters input by a user and parameters in a comprehensive database, and calculating corresponding irrigation quantity; the fertilization decision is carried out simultaneously, whether fertilization is carried out or not is determined by using a fertilization amount calculation model, and the fertilization amount is calculated; sending the decision result to an executing device of the intelligent control system;

according to the accumulated effective rainfall sigma P in the irrigation decision period_eDetermining the precipitation condition:

∑P_ethe sum of effective precipitation in unit irrigation area in the forecast period is in mm;

P_ealpha is a rainfall infiltration coefficient, and P is the future rainfall obtained from weather forecast information; when P is less than 5mm, alpha is 0; when P is 5-50mm, taking 1.0-0.8; when the rainfall is more than 50mm, taking 0.7-0.8;

(1) cumulative effective rainfall sigma P in irrigation decision period_eCalculating the irrigation quantity as 0, and immediately irrigating;

(2) cumulative effective rainfall sigma P in irrigation decision period_eNot more than the maximum irrigation quantity W₁', calculating the irrigation quantity and immediately irrigating;

(3) cumulative effective rainfall sigma P in irrigation decision period_eMaximum irrigation quantity W₁', there are two categories of conditions depending on whether the crop is drought before precipitation:

a. water content theta in root zone of crop_iSoil moisture content not greater than crop withering point theta_dryImmediately irrigating;

b. water content theta in root zone of crop_iSoil water content theta higher than crop withering point_dryNo irrigation;

W₁' is the maximum irrigation per unit irrigation area; theta is the water content percentage of the soil volume detection, and is dimensionless.

2. The intelligent decision-making method according to claim 1, wherein the irrigation quantity calculation model in the drip irrigation decision is as follows:

(1) cumulative effective rainfall sigma P in irrigation decision period_eWater filling quantity W is equal to 0₁The calculation formula is shown in formula 2:

(2) cumulative effective rainfall sigma P in irrigation decision period_eNot more than the maximum irrigation quantity W₁', amount of irrigation W₂The calculation formula is shown in formula 3:

(3) cumulative effective rainfall sigma P in irrigation decision period_eMaximum irrigation quantity W₁′：

a. Water content theta in root zone of crop_iSoil moisture content not greater than crop withering point theta_dryAmount of irrigation W₃The calculation formula is shown in formula 13:

in formulae 2,3 and 13, W₁、W₂、W₃For the amount of irrigation, unit m³；θ_maxThe upper limit of irrigation, namely the volume water content of the target soil to be irrigated,%, is dimensionless; theta is the water content percentage of the soil volume detection, and is dimensionless; z is the planned wetting layer depth of the crop in mm; p is the wetting ratio of the drip irrigation soil,%, and is dimensionless; m is the area of the irrigation area decided on, and the unit is M²；∑ET_cThe sum of the crop water consumption in unit irrigation area from the prior to the first occurrence of effective rainfall is expressed in mm;

W₁′＝W₁1000/M in mm.

3. The intelligent decision-making method according to claim 1, wherein the fertilization amount calculation model in the fertilization decision is as follows:

in the formula 15, F is the fertilizing amount per time and the unit is kg; f_cThe unit of the total amount of nutrients required by the target output per mu at the decision point is kg; s is the amount of the available nutrients in kg of soil per mu at the decision point; u is the season utilization rate of the fertilizer,%; f. of_zThe proportion of top dressing of crops in the whole growth period is percent; f. of_iThe proportion of the fertilizer is distributed for the additional fertilizer in the growth period; t is t_fThe number of fertilization times is empirical.

4. The intelligent decision-making method according to claim 1, wherein the database construction converts the parameter point data used by the drip irrigation and fertilization calculation model into spatial data by using a spatial interpolation method and a data rasterization method;

the spatial interpolation method is preferably an inverse distance weight interpolation method, and the calculation formula is shown as formula 18:

in formula 18, Z (x)₀) Is x₀The predicted value of (c); n is the number of sample points around the prediction point to be used in the prediction calculation process; lambda [ alpha ]_iFor predicting the weight of each sample point used in the calculation, the value decreases as the distance between the sample point and the predicted point increases; z (x)_i) Is at x_iObtaining the obtained measured value;

the data rasterization method preferably uses ArcGIS software.

5. An intelligent decision method according to claim 1, characterized in that after the database is formed, data mining is preferably performed by using a support vector machine and a random forest to obtain a new fertigation model.

6. The intelligent decision-making method according to claim 1, characterized in that after the database is formed, the data can be cleaned for the parameters input by the user to remove abnormal values in the data.

7. An intelligent decision method as claimed in claim 1, wherein the decision method performs fertilization every irrigation, records the cumulative fertilization amount sigma F every time the water and fertilizer are irrigated integrally, and if the cumulative fertilization amount sigma F is larger than or equal to F.t after a certain irrigation fertilization_fIrrigation after the growth period is not fertilized any more untilThe next growth period of the crop.

8. An intelligent decision making system based on the intelligent decision making method according to any one of claims 1 to 7, wherein the intelligent decision making system comprises an input module, a crop module, a model parameter module and an output module;

the input module comprises soil moisture data, irrigation area, wetting ratio, geographical position, meteorological data and sowing time;

the crop module comprises a growth period parameter and a crop fertilization parameter, the growth period parameter comprises a crop coefficient, an irrigation upper limit, an irrigation lower limit, a crop drought-enduring index, a planned wetting layer and a management suggestion, and the crop fertilization parameter comprises empirical fertilization times, a growth period topdressing distribution ratio and a topdressing ratio;

the model parameter module comprises irrigation quantity calculation and fertilization quantity calculation; the output module comprises fertilizing amount, irrigation amount and management suggestion.

9. The intelligent decision making system according to claim 8, wherein after the user inputs the geographical location or selects the corresponding parameters in the input module at the use end, the intelligent decision making system makes the decision of drip irrigation and fertilization, automatically judges the growth period of crops, and pushes agricultural and agronomic management measures related to seedling management to the management personnel of the drip irrigation system.

Images