Liquid fertilizer variable-rate fertilization control method capable of realizing automatic proportioning and algorithm self-adaptive switching
Through the liquid fertilizer variable fertilization control method with automatic proportion and algorithm adaptive switching, the problem of low control accuracy of variable fertilization in the existing technology is solved, the precise application and efficient proportion of liquid fertilizer are achieved, and the adaptability and accuracy of the fertilization system are improved. .
Patent Information
- Application Number
- CN202411122208.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-15
- Publication Date
- 2024-11-26
AI Technical Summary
Due to nonlinearity and hysteresis problems, the existing variable fertilization control system is difficult to maintain high accuracy in complex operating environments, and it is unable to effectively adapt to external disturbances, resulting in large fertilization errors.
The liquid fertilizer variable fertilization control method is adopted for automatic proportioning and algorithm adaptive switching. By calculating the liquid fertilizer application volume of the current and next plots, combined with the RBF-PID control and sliding mode control algorithm optimized by the particle swarm algorithm, switch according to the disturbance amount Control algorithms to achieve accurate proportioning and transportation of liquid fertilizers to ensure fertilization accuracy.
It effectively reduces the waste of liquid fertilizer, improves the fertilization accuracy and system adaptability, and ensures the efficiency and accuracy of fertilization in different operating environments.
Smart Images

Figure CN119024696A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of agricultural fertilization, and in particular to a liquid fertilizer variable fertilization control method with automatic proportioning and algorithm adaptive switching. Background Art
[0002] The valve-controlled hydraulic motor variable fertilizer control system is widely used in variable fertilizer spreaders. However, due to the nonlinear and hysteresis problems of the control system itself and the combined effects of external disturbances during the actual operation process, a suitable control algorithm is required to ensure the good operation effect of the system and ensure that the variable fertilization operation has good fertilization accuracy.
[0003] The variable fertilization control methods currently used mainly include conventional PID control, fuzzy PID control, and neural network optimized PID control. However, the operating scenarios of the control system under a single algorithm are limited and cannot adapt well to the actual complex operating environment. In addition, the adaptability of the algorithm in actual operation is not taken into account during the control process, resulting in low control accuracy and large fertilization error in the actual operation of the variable fertilizer spreader. Summary of the invention
[0004] In order to solve the problems existing in the prior art, the present invention proposes a liquid fertilizer variable fertilization control method with automatic proportioning and algorithm adaptive switching on the basis of the prior art, which can realize automatic proportioning of liquid fertilizer, avoid waste of fertilizer, and ensure that the variable fertilizer spreader still has excellent control performance when subjected to external disturbances during actual operation.
[0005] In order to achieve the technical effects of the present invention, the technical solutions adopted by the present invention are specifically as follows:
[0006] A variable fertilization method comprises the following steps:
[0007] S1. Calculate the volume of liquid fertilizer applied in the current and next variable fertilization unit plots: According to the fertilizer amount information F in the fertilizer prescription map carried by the liquid fertilizer variable fertilization control system, obtain the target fertilizer weight M of nitrogen, phosphorus and potassium in the current and next variable fertilization unit plots. N1 、M P1 、M K1 and M N2 、M P2 、M K2 , and calculate the application volume Q of nitrogen, phosphorus and potassium liquid fertilizers for the current plot and the next plot respectively N1 , Q P1 , Q K1 and Q N2 , Q P2 , Q K2; Among them, the fertilizer amount information F stored in the prescription map is represented by a data set, specifically: F = {(M N1 ,M P1 ,M K1 ),(M N2 ,M P2 ,M K2 ),...,(M Nm ,M Pm ,M Km )}, m is the total number of variable fertilization unit plots;
[0008] S2. Automatically proportion and deliver the liquid fertilizer required for the current plot: Since the liquid outlets of the nitrogen fertilizer tank, phosphate fertilizer tank, potash fertilizer tank, and clean water tank are all equipped with proportional valves, the mathematical relationship between the analog current signal of each proportional valve and the corresponding liquid volume discharged from the tank within time T1 is calculated as f(I A1 ,Q N )、f(I A2 ,Q P )、f(I A3 ,Q K )、f(I A4 ,Q w ) is controlled to discharge the nitrogen, phosphorus and potassium liquid fertilizers and the corresponding volume of clean water required for the current plot into the second fertilizer storage chamber, so that the fertilizers and clean water are fully mixed, and then the pipeline switch under the second fertilizer storage chamber is opened to transport all the prepared liquid fertilizers to the first fertilizer storage chamber;
[0009] S3, automatically proportioning and delivering the liquid fertilizer required for the next plot: after the liquid fertilizer of the initial plot is delivered to the first fertilizer storage chamber, according to the mathematical relationship f(I A1 ,Q N )、f(I A2 ,Q P )、f(I A3 ,Q K ), the nitrogen, phosphorus and potassium liquid fertilizer stock solutions and the corresponding amount of clean water required for the next plot are mixed and discharged into the second fertilizer storage chamber, and when the liquid fertilizer variable fertilizer spreader is detected to enter the next plot, it is discharged into the first fertilizer storage chamber to apply the liquid fertilizer to the corresponding plot;
[0010] S4. Real-time calculation of the target fertilization flow value and fertilization flow error of the control system: According to the actual operating speed of the liquid fertilizer variable fertilizer applicator and the volume of liquid fertilizer to be applied to the plot in the first fertilizer storage chamber, the target fertilization flow value of the current system is calculated. The calculation formula is as follows:
[0011]
[0012] Among them, q(t) is the target fertilizer flow rate of the current system, in L / min; v(t) is the actual operating speed of the fertilizer spreader, in m / s; Q N1 , Q P1 , Q K1 They are the target fertilizer amounts of nitrogen, phosphorus and potassium for the current variable fertilization plot, in L; Q w1 is the amount of clean water required to be discharged into the current variable fertilization plot, in L; D is the side length of the variable fertilization unit plot, in m;
[0013] The calculation formula of fertilization flow error is as follows:
[0014] e(t)=q(t)-y(t)
[0015] Among them, e(t) is the flow error of the current system, in L / min; q(t) is the target fertilization flow of the current system, in L / min; y(t) is the actual flow output value measured by the flow sensor of the current system, in L / min;
[0016] S5. Select the initial control algorithm of the system and calculate the disturbance amount: the control system is equipped with two different control algorithms, namely the RBF-PID control algorithm optimized by the particle swarm algorithm and the sliding mode control algorithm SMC. At the initial moment, the system uses the RBF-PID control algorithm optimized by the particle swarm algorithm to control the proportional valve under the first fertilizer storage chamber. At the same time, the disturbance amount Δq of the current control system is calculated;
[0017] S6. Switch the control algorithm used by the system according to the disturbance amount: judge according to the calculated disturbance amount Δq. When the disturbance is When , the system continues to use the RBF-PID control algorithm optimized by the particle swarm algorithm to control the proportional valve under the first fertilizer storage chamber;
[0018] S7, update of target fertilizer application amount information of the control system: the fertilizer spreader determines the current location of the working plot according to the GPS module, and updates the target fertilizer application weight M of the control system when it detects that the working plot has changed. N 、M P 、M K , repeat steps S1 to S5 until the variable fertilization operation is completed; when the operation is completed, the liquid in the tank is stopped from being discharged by controlling the proportional valves under each liquid fertilizer tank and the clean water tank;
[0019] Furthermore, the calculation formula for the volume of nitrogen, phosphorus and potassium liquid fertilizer applied in step S1 is as follows:
[0020]
[0021] Among them, Q N , Q P , Q K They are the target fertilizer amounts of nitrogen, phosphorus and potassium for variable fertilization plots, in L; M N 、M P 、M K are the target fertilizer weights of nitrogen, phosphorus and potassium in the variable fertilization plot, in kg; ω N ,ω P ,ω K They are the nitrogen content of liquid nitrogen fertilizer, the phosphorus content of liquid phosphorus fertilizer, and the potassium content of liquid potassium fertilizer, in %; ρ N , P , K They are the original solution densities of liquid nitrogen fertilizer, liquid phosphate fertilizer and liquid potassium fertilizer, in kg / L;
[0022] Furthermore, the mathematical relationship of the proportional valve in step S2 is obtained through a calibration test, and the specific steps are as follows:
[0023] S2-1. According to the analog current value signal range of the proportional valve, 10 analog current values are taken at equal intervals, and the above 10 current signal values are input into the proportional valve in order from small to large. At the same time, the volume of liquid nitrogen fertilizer, liquid phosphate fertilizer, liquid potash fertilizer and clean water discharged by the proportional valve within time T1 is measured with a measuring cylinder. Each measurement point is repeated three times to take the average value, and the corresponding discharged liquid volume value under each current signal value is recorded respectively;
[0024] S2-2. Considering the hysteresis loop of the proportional valve, the above 10 current signal values are input into the proportional valve in order from large to small. At the same time, the volume of liquid discharged by the proportional valve within time T1 is measured with a measuring cylinder. Each measurement point is repeated three times to take the average value, and the volume of liquid discharged under each current signal value is recorded respectively.
[0025] S2-3, respectively, the current signal value and liquid volume value obtained in the above steps are used as the horizontal and vertical coordinates, and a linear curve is used to fit the data points, and the mathematical relationship between the magnitude of the simulated current signal of each proportional valve and the volume of nitrogen fertilizer, phosphate fertilizer, potassium fertilizer and clean water discharged from the liquid fertilizer tank within time T1 is obtained. A1 ,Q N )、f(I A2 ,Q P )、f(I A3 ,Q K ) and f(I A4 ,Q w );
[0026] Furthermore, the time T1 in step S2 is required to be as follows:
[0027]
[0028] Wherein, T1 is the measurement time, in seconds; D is the side length of the variable fertilization unit plot, in meters; v max The maximum travel speed of the fertilizer spreader during operation, in m / s;
[0029] The volume Q of the clean water injected in step S2 w The calculation formula is:
[0030] Q w =C1Q N +C2Q P +C3Q K
[0031] Among them, Q w is the volume of clean water discharged, in L; C1, C2, and C3 are the dilution multiples of the original solutions of nitrogen fertilizer, phosphate fertilizer, and potassium fertilizer, respectively, and are all constants;
[0032] Furthermore, the calculation formula of the system disturbance Δq in step S4 is as follows:
[0033]
[0034] Among them, Δq is the system disturbance, in L / min; ΣT is the total sampling time, in seconds; T2 is the time taken by the fertilizer spreader to pass through the variable fertilization unit plot, in seconds; q1, q2, q n are the instantaneous flow values measured in the first, second, and nth sampling time, in L / min; t1, t2, t n They are the first, second and nth sampling times respectively, in seconds; and the requirements are:
[0035]
[0036] Among them, t s is the set sampling time, in seconds; D is the side length of the variable fertilization unit, in meters; v is the current speed of the fertilizer spreader, in meters per second;
[0037] Furthermore, the specific process of the sliding mode control algorithm SMC in S5 is as follows:
[0038] S5-1. Derive the equilibrium equation of the valve-controlled hydraulic motor variable fertilization system and remove the high-order differential terms, as follows:
[0039]
[0040] Among them, Q b D is the flow rate of liquid fertilizer output by the liquid fertilizer pump, in L / min; b is the displacement of the liquid fertilizer pump, in mL / r; η is the volumetric efficiency of the liquid fertilizer pump; ω is the speed of the hydraulic motor, in rad / s; K q is the flow gain of the valve-controlled hydraulic motor system, in m 3 ·s -1 ·V -1 ; xv is the displacement of the valve core of the proportional valve, in m; D m is the displacement of the hydraulic motor, in m 3 / rad;ω v is the natural frequency of the valve, in rad / s; ε v is the damping ratio of the valve; U is the control voltage input to the proportional valve, in V;
[0041] S5-2, real-time acquisition of the current target fertilization flow q(t) of the system and calculation of the real-time flow error e(t) of the system, and calculation of the sliding surface function of the current system through the balance equation of the above valve-controlled hydraulic motor variable fertilization system, as follows:
[0042]
[0043] Where, s is the sliding surface function; c is a constant greater than 0; t is time, in seconds; e(t) is the system flow error, in L / min;
[0044] S5-3. Determine the sliding mode reaching law:
[0045]
[0046] Among them, a and p are constants greater than 0;
[0047] S5-4. Calculate the control voltage value that the current system should input to the proportional valve according to the derived control voltage expression equation, and implement the control of the proportional valve by the sliding mode control algorithm SMC; the specific equation is as follows:
[0048]
[0049] S5-5, repeating steps S5-2 to S5-4 until the system switches to another control algorithm or the variable fertilization operation is completed;
[0050] Furthermore, the specific process of optimizing RBF-PID control by particle swarm algorithm in step S5 is as follows:
[0051] S5-A, build a simulation model of variable fertilization control system, set the total number of particles in the particle swarm algorithm to 50, the maximum number of iterations N = 100, and the position matrix and velocity matrix of each particle at the initial moment are randomly assigned; encode the network structure parameter values in the RBF neural network to be optimized as the initial particle position in the particle swarm algorithm, as follows:
[0052] P=[c1,b1,ω1,c2,b2,ω2,...,c6,b6,ω6]
[0053] Among them, P is the initial position matrix of the particle, c1, c2, ..., c6 are the hidden layer node center values of the RBF neural network, b1, b2, ..., b6 are the network hidden layer node width values, ω1, ω2, ..., ω6 are the network output layer weights;
[0054] Substitute the initial position matrix of each particle into the RBF-PID control model to perform unit step response simulation, and set the simulation time to 2s;
[0055] S5-B, after the simulation is completed, the following formula is the evaluation index of the particle swarm algorithm optimization effect, and the evaluation index value of each particle at the initial stage is calculated;
[0056]
[0057] Among them, T3 is the simulation duration, in seconds; t is the simulation time, in seconds; r is the unit step target value, r=1; y is the actual output value of the system;
[0058] S5-C, according to the evaluation index of each particle, update the position and speed of each particle, specifically as follows: and re-substitute the updated position matrix of each particle into the RBF-PID model to obtain the updated evaluation index value;
[0059]
[0060] Among them, P(t+1) is the position matrix of the particle at time t+1, P(t) is the position matrix of the particle at time t, w(t) is the inertia weight of the particle, V(t+1) is the velocity matrix of the particle at time t+1, V(t) is the velocity matrix of the particle at time t, P b (t) is the individual optimal position matrix found by the same particle from the beginning of the simulation to time t, G b (t) is the global optimal position matrix found by all particles from the start of simulation to time t, k1 and k2 are constants, r1 and r2 are random numbers between 0 and 1;
[0061] The inertia weight changes accordingly with the number of iterations, as follows:
[0062]
[0063] Among them, w(t) is the inertia weight at time t, n is the current iteration number of the particle swarm algorithm, N is the maximum iteration number, and w max is the maximum weight set, w min is the minimum weight set;
[0064] S5-D, resubstituting the updated position matrix of each particle into the RBF-PID model to obtain the updated evaluation index value, and updating the global optimal position matrix and the optimal position matrix of each particle in the particle swarm algorithm, and jumping to step S5-C to continue the iterative process;
[0065] S5-E, repeat steps S5-C to S5-D until the current number of iterations n is equal to the maximum number of iterations N; return the optimal position matrix G of the group obtained at this time b (t), substitute the corresponding parameters in the matrix back into the RBF neural network to complete the particle swarm optimization process;
[0066] Furthermore, the specific process of the RBF-PID control algorithm after particle swarm optimization in step S5 is as follows:
[0067] S5-a, initialize the network parameters, and use the optimal position matrix G obtained by the particle swarm algorithm b (t) Decomposition, and substitute the optimized hidden layer node center value c, hidden layer node width value b, and network output layer weight w into the RBF neural network;
[0068] S5-b, collect the control voltage of the fertilization control system at the current moment and the actual flow output value of the system, and form the input vector x with the control voltage difference Δu, the actual flow output value y, and the flow deviation value e before and after, x = [Δu(t), y(t), e(t)], use the vector x as the RBF neural network input signal, and calculate the network output as follows:
[0069]
[0070] Among them, y m is the flow identification output of the network, in L / min; x = [Δu(t), y(t), e(t)], Δu is the control voltage difference between the two times, in V; y is the actual flow output value measured by the flow sensor of the current system, in L / min; e is the flow deviation value, in L / min; c j is the center value of the hidden layer node after particle swarm optimization, b j is the hidden layer node width value after particle swarm algorithm optimization;
[0071] S5-c, update PID parameters according to network output results
[0072]
[0073] Among them, K P , K I , K D They are the proportional coefficient, differential coefficient and integral coefficient of PID control respectively; η P , η I , η D are the learning rates of the proportional coefficient, differential coefficient and integral coefficient, respectively, all of which are constants; Δu is the control voltage difference between the previous and the next two times, in V; e(t) is the liquid fertilizer flow error at the current moment, in L / min; e(t-1) is the liquid fertilizer flow error at the previous moment, in L / min; e(t-2) is the liquid fertilizer flow error at the previous two moments, in L / min;
[0074] S5-d, calculate the control voltage of the system at the current moment according to the following formula, and output it through the control system;
[0075]
[0076] Among them, U is the control voltage that the system should output at the current moment, in V;
[0077] S5-e, update the network parameter value according to the following formula:
[0078]
[0079] Among them, c j0 , b j0 ,ω j0 are the hidden layer node center value, hidden layer node width value, and output layer weight value of the RBF neural network at the initial moment; η is the network learning rate, 0<η<1; y is the actual output flow of the system, in L / min; y m is the identified flow rate of the neural network, in L / min;
[0080] S5-f, repeating steps S5-b to S5-e until the system switches to another control algorithm or the variable fertilization operation is completed;
[0081] Furthermore, the liquid fertilizer variable fertilization control system includes the following key components:
[0082] There are three liquid fertilizer tanks, namely liquid nitrogen fertilizer tank, liquid phosphate fertilizer tank and liquid potassium fertilizer tank. There is a liquid fertilizer outlet at the bottom of the liquid fertilizer tank to ensure the smooth discharge of liquid fertilizer;
[0083] A clean water tank stores a sufficient amount of clean water, and a liquid outlet is provided at the bottom of the tank to ensure smooth discharge of the clean water;
[0084] Proportional valves: There are 4 proportional valves in total, which are installed at the outlets of each liquid fertilizer tank and clean water tank to control the volume of discharged liquid. The feedback signal type of the proportional valve used is an analog current signal.
[0085] The second fertilizer storage chamber is used to mix the corresponding amount of liquid fertilizer and clean water, and also serves to store the liquid fertilizer required for the next fertilization plot in advance;
[0086] The first fertilizer storage room is used to store the liquid fertilizer required for the current plot of land and is directly connected to the liquid fertilizer pump through a pipeline;
[0087] A valve-controlled hydraulic motor system is used to provide power to the liquid fertilizer pump, wherein the proportional valve input signal is a voltage signal, and the hydraulic motor output shaft is connected to the liquid fertilizer pump through a rigid coupling;
[0088] A liquid fertilizer pump, used for discharging liquid fertilizer in the first fertilizer storage chamber;
[0089] Flow sensor, used to detect the real-time output flow of the system, installed at the outlet of the liquid fertilizer pump;
[0090] Compared with the prior art, the present invention has the following beneficial effects:
[0091] The present invention adopts variable fertilization technology with automatic liquid fertilizer proportioning, and performs precise proportioning of liquid fertilizer according to a fertilization prescription map during fertilization operation, and applies liquid fertilizer immediately after proportioning, which can ensure the efficiency of fertilization operation and effectively reduce the waste of liquid fertilizer;
[0092] The present invention adopts a sliding mode control algorithm SMC for control when the system disturbance is large, and utilizes the characteristic that the sliding mode control is insensitive to the external disturbance and uncertain parameters of the system, thereby solving the problem of low fertilization accuracy of the system under large disturbance and improving the adaptability of the system; when the system disturbance is small, a particle swarm optimized RBF-PID control algorithm is adopted for control, thereby improving the nonlinear control performance of the system and ensuring the accuracy of variable application of liquid fertilizer. BRIEF DESCRIPTION OF THE DRAWINGS
[0093] Figure 1 It is the operation flow chart of the present invention;
[0094] Figure 2 It is a schematic diagram of the operation process of the liquid fertilizer variable-rate fertilizer applicator of the present invention;
[0095] Figure 3 It is a structural schematic diagram of the liquid fertilizer variable fertilization control system in the present invention. DETAILED DESCRIPTION
[0096] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0097] According to one aspect of an embodiment of the present invention, a variable fertilization method for automatic proportioning of liquid fertilizer is provided, which can automatically proportion the liquid fertilizer according to the application amount of liquid nitrogen fertilizer, liquid phosphate fertilizer, and liquid potassium fertilizer in the fertilization prescription diagram in combination with the dilution multiple of the liquid fertilizer, so that the liquid fertilizer can be used immediately after proportioning, thereby avoiding waste of liquid fertilizer. On the other hand, a variable fertilization method for switching the control algorithm of the system according to the size of the system disturbance is provided, and the current operation effect is judged by real-time calculation of the size of the system disturbance, so as to solve the problem of low fertilization accuracy of the liquid fertilizer variable fertilization control system under large disturbance.
[0098] like Figure 1 As shown, the present invention first obtains the target fertilizer amount information of the current plot according to the fertilizer prescription map, automatically proportions the liquid fertilizer according to the required amount of liquid fertilizer, and proportions the liquid fertilizer application amount of the next plot in advance to improve efficiency; different control algorithms are selected according to the size of the system disturbance to improve the fertilization accuracy. The specific steps are as follows:
[0099] S1. Calculate the volume of liquid fertilizer applied in the current and next variable fertilization unit plots: According to the fertilizer amount information F in the fertilizer prescription map carried by the liquid fertilizer variable fertilization control system, obtain the target fertilizer weight M of nitrogen, phosphorus and potassium in the current and next variable fertilization unit plots. N1 、M P1 、M K1 and M N2 、M P2 、M K2 , and calculate the application volume Q of nitrogen, phosphorus and potassium liquid fertilizers for the current plot and the next plot respectively N1 , Q P1 , Q K1 and Q N2 , Q P2 , Q K2 ; Among them, the fertilizer amount information F stored in the prescription map is represented by a data set, specifically: F = {(M N1 ,M P1 ,M K1 ),(M N2 ,M P2 ,M K2 ),...,(M Nm ,M Pm ,MKm )}, m is the total number of variable fertilization unit plots;
[0100] S2. Automatically proportion and deliver the liquid fertilizer required for the current plot: Since the liquid outlets of the nitrogen fertilizer tank, phosphate fertilizer tank, potash fertilizer tank, and clean water tank are all equipped with proportional valves, the mathematical relationship between the analog current signal of each proportional valve and the corresponding liquid volume discharged from the tank within time T1 is calculated as f(I A1 ,Q N )、f(I A2 ,Q P )、f(I A3 ,Q K )、f(I A4 ,Q w ) is controlled to discharge the nitrogen, phosphorus and potassium liquid fertilizers and the corresponding volume of clean water required for the current plot into the second fertilizer storage chamber, so that the fertilizers and clean water are fully mixed, and then the pipeline switch under the second fertilizer storage chamber is opened to transport all the prepared liquid fertilizers to the first fertilizer storage chamber;
[0101] S3, automatically proportioning and delivering the liquid fertilizer required for the next plot: after the liquid fertilizer of the initial plot is delivered to the first fertilizer storage chamber, according to the mathematical relationship f(I A1 ,Q N )、f(I A2 ,Q P )、f(I A3 ,Q K ), the nitrogen, phosphorus and potassium liquid fertilizer stock solutions and the corresponding amount of clean water required for the next plot are mixed and discharged into the second fertilizer storage chamber, and when the liquid fertilizer variable fertilizer spreader is detected to enter the next plot, it is discharged into the first fertilizer storage chamber to apply the liquid fertilizer to the corresponding plot;
[0102] S4. Real-time calculation of the target fertilization flow value and fertilization flow error of the control system: According to the actual operating speed of the liquid fertilizer variable fertilizer applicator and the volume of liquid fertilizer to be applied to the plot in the first fertilizer storage chamber, the target fertilization flow value of the current system is calculated. The calculation formula is as follows:
[0103]
[0104] Among them, q(t) is the target fertilizer flow rate of the current system, in L / min; v(t) is the actual operating speed of the fertilizer spreader, in m / s; Q N1 , Q P1 , Q K1 They are the target fertilizer amounts of nitrogen, phosphorus and potassium for the current variable fertilization plot, in L; Q w1 is the amount of clean water required to be discharged into the current variable fertilization plot, in L; D is the side length of the variable fertilization unit plot, in m;
[0105] The calculation formula of fertilization flow error is as follows:
[0106] e(t)=q(t)-y(t)
[0107] Among them, e(t) is the flow error of the current system, in L / min; q(t) is the target fertilization flow of the current system, in L / min; y(t) is the actual flow output value measured by the flow sensor of the current system, in L / min;
[0108] S5. Select the initial control algorithm of the system and calculate the disturbance amount: the control system is equipped with two different control algorithms, namely the RBF-PID control algorithm optimized by the particle swarm algorithm and the sliding mode control algorithm SMC. At the initial moment, the system uses the RBF-PID control algorithm optimized by the particle swarm algorithm to control the proportional valve under the first fertilizer storage chamber. At the same time, the disturbance amount Δq of the current control system is calculated;
[0109] S6. Switch the control algorithm used by the system according to the disturbance amount: judge according to the calculated disturbance amount Δq. When the disturbance is When , the system continues to use the RBF-PID control algorithm optimized by the particle swarm algorithm to control the proportional valve under the first fertilizer storage chamber;
[0110] S7, update of target fertilizer application amount information of the control system: the fertilizer spreader determines the current location of the working plot according to the GPS module, and updates the target fertilizer application weight M of the control system when it detects that the working plot has changed. N 、M P 、M K , repeat steps S1 to S5 until the variable fertilization operation is completed; when the operation is completed, the liquid in the tank is stopped from being discharged by controlling the proportional valves under each liquid fertilizer tank and the clean water tank;
[0111] Furthermore, the calculation formula for the volume of nitrogen, phosphorus and potassium liquid fertilizer applied in step S1 is as follows:
[0112]
[0113] Among them, Q N , Q P , Q K They are the target fertilizer amounts of nitrogen, phosphorus and potassium for variable fertilization plots, in L; M N 、M P 、M Kare the target fertilizer weights of nitrogen, phosphorus and potassium in the variable fertilization plot, in kg; ω N ,ω P ,ω K They are the nitrogen content of liquid nitrogen fertilizer, the phosphorus content of liquid phosphorus fertilizer, and the potassium content of liquid potassium fertilizer, in %; ρ N , P , K They are the original solution densities of liquid nitrogen fertilizer, liquid phosphate fertilizer and liquid potassium fertilizer, in kg / L;
[0114] Furthermore, the mathematical relationship of the proportional valve in step S2 is obtained through a calibration test, and the specific steps are as follows:
[0115] S2-1. According to the analog current value signal range of the proportional valve, 10 analog current values are taken at equal intervals, and the above 10 current signal values are input into the proportional valve in order from small to large. At the same time, the volume of liquid nitrogen fertilizer, liquid phosphate fertilizer, liquid potash fertilizer and clean water discharged by the proportional valve within time T1 is measured with a measuring cylinder. Each measurement point is repeated three times to take the average value, and the corresponding discharged liquid volume value under each current signal value is recorded respectively;
[0116] S2-2. Considering the hysteresis loop of the proportional valve, the above 10 current signal values are input into the proportional valve in order from large to small. At the same time, the volume of liquid discharged by the proportional valve within time T1 is measured with a measuring cylinder. Each measurement point is repeated three times to take the average value, and the volume of liquid discharged under each current signal value is recorded respectively.
[0117] S2-3, respectively, the current signal value and liquid volume value obtained in the above steps are used as the horizontal and vertical coordinates, and a linear curve is used to fit the data points, and the mathematical relationship between the magnitude of the simulated current signal of each proportional valve and the volume of nitrogen fertilizer, phosphate fertilizer, potassium fertilizer and clean water discharged from the liquid fertilizer tank within time T1 is obtained. A1 ,Q N )、f(I A2 ,Q P )、f(I A3 ,Q K ) and f(I A4 ,Q w );
[0118] like Figure 2 As shown, the time T1 in step S2 is required as follows:
[0119]
[0120] Where, T1 is the measurement time, in seconds; D is the side length of the variable fertilization unit, in meters; v max The maximum travel speed of the fertilizer spreader during operation, in m / s;
[0121] The volume Q of the clean water injected in step S2 w The calculation formula is:
[0122] Q w =C1Q N +C2Q P +C3Q K
[0123] Among them, Q w is the volume of clean water discharged, in L; C1, C2, and C3 are the dilution multiples of the original solutions of nitrogen fertilizer, phosphate fertilizer, and potassium fertilizer, respectively, and are all constants;
[0124] Furthermore, the calculation formula of the system disturbance Δq in step S4 is as follows:
[0125]
[0126] Among them, Δq is the system disturbance, in L / min; ΣT is the total sampling time, in seconds; T2 is the time taken by the fertilizer spreader to pass through the variable fertilization unit plot, in seconds; q1, q2, q n are the instantaneous flow values measured in the first, second, and nth sampling time, in L / min; t1, t2, t n They are the first, second and nth sampling times respectively, in seconds; and the requirements are:
[0127]
[0128] Among them, t s is the set sampling time, in seconds; D is the side length of the variable fertilization unit, in meters; v is the current speed of the fertilizer spreader, in meters per second;
[0129] Furthermore, the specific process of the sliding mode control algorithm SMC in S5 is as follows:
[0130] S5-1. Derive the equilibrium equation of the valve-controlled hydraulic motor variable fertilization system and remove the high-order differential terms, as follows:
[0131]
[0132] Among them, Q b D is the flow rate of liquid fertilizer output by the liquid fertilizer pump, in L / min; b is the displacement of the liquid fertilizer pump, in mL / r; η is the volumetric efficiency of the liquid fertilizer pump; ω is the speed of the hydraulic motor, in rad / s; K q is the flow gain of the valve-controlled hydraulic motor system, in m 3 ·s -1 ·V-1 ;x v D is the displacement of the valve core of the proportional valve, in m; m is the displacement of the hydraulic motor, in m 3 / rad;ω v is the natural frequency of the valve, in rad / s; ε v is the damping ratio of the valve; U is the control voltage input to the proportional valve, in V;
[0133] S5-2, real-time acquisition of the current target fertilization flow q(t) of the system and calculation of the real-time flow error e(t) of the system, and calculation of the sliding surface function of the current system through the balance equation of the above valve-controlled hydraulic motor variable fertilization system, as follows:
[0134]
[0135] Where, s is the sliding surface function; c is a constant greater than 0; t is time, in seconds; e(t) is the system flow error, in L / min;
[0136] S5-3. Determine the sliding mode reaching law:
[0137]
[0138] Among them, a and p are constants greater than 0;
[0139] S5-4. Calculate the control voltage value that the current system should input to the proportional valve according to the derived control voltage expression equation, and implement the control of the proportional valve by the sliding mode control algorithm SMC; the specific equation is as follows:
[0140]
[0141] S5-5, repeating steps S5-2 to S5-4 until the system switches to another control algorithm or the variable fertilization operation is completed;
[0142] Furthermore, the specific process of optimizing RBF-PID control by particle swarm algorithm in step S5 is as follows:
[0143] S5-A, build a simulation model of variable fertilization control system, set the total number of particles in the particle swarm algorithm to 50, the maximum number of iterations N = 100, and the position matrix and velocity matrix of each particle at the initial moment are randomly assigned; encode the network structure parameter values in the RBF neural network to be optimized as the initial particle position in the particle swarm algorithm, as follows:
[0144] P=[c1,b1,ω1,c2,b2,ω2,…,c6,b6,ω6]
[0145] Among them, P is the initial position matrix of the particle, c1, c2, ..., c6 are the hidden layer node center values of the RBF neural network, b1, b2, ..., b6 are the network hidden layer node width values, ω1, ω2, ..., ω6 are the network output layer weights;
[0146] Substitute the initial position matrix of each particle into the RBF-PID control model to perform unit step response simulation, and set the simulation time to 2s;
[0147] S5-B, after the simulation is completed, the following formula is the evaluation index of the particle swarm algorithm optimization effect, and the evaluation index value of each particle at the initial stage is calculated;
[0148]
[0149] Among them, T3 is the simulation duration, in seconds; t is the simulation time, in seconds; r is the unit step target value, r=1; y is the actual output value of the system;
[0150] S5-C, according to the evaluation index of each particle, update the position and speed of each particle, specifically as follows: and re-substitute the updated position matrix of each particle into the RBF-PID model to obtain the updated evaluation index value;
[0151]
[0152] Among them, P(t+1) is the position matrix of the particle at time t+1, P(t) is the position matrix of the particle at time t, w(t) is the inertia weight of the particle, V(t+1) is the velocity matrix of the particle at time t+1, V(t) is the velocity matrix of the particle at time t, P b (t) is the individual optimal position matrix found by the same particle from the beginning of the simulation to time t, G b (t) is the global optimal position matrix found by all particles from the start of simulation to time t, k1 and k2 are constants, r1 and r2 are random numbers between 0 and 1;
[0153] The inertia weight changes accordingly with the number of iterations, as follows:
[0154]
[0155] Among them, w(t) is the inertia weight at time t, n is the current iteration number of the particle swarm algorithm, N is the maximum iteration number, and w max is the maximum weight set, w min is the minimum weight set;
[0156] S5-D, resubstituting the updated position matrix of each particle into the RBF-PID model to obtain the updated evaluation index value, and updating the global optimal position matrix and the optimal position matrix of each particle in the particle swarm algorithm, and jumping to step S5-C to continue the iterative process;
[0157] S5-E, repeat steps S5-C to S5-D until the current number of iterations n is equal to the maximum number of iterations N; return the optimal position matrix G of the group obtained at this time b (t), substitute the corresponding parameters in the matrix back into the RBF neural network to complete the particle swarm optimization process;
[0158] Furthermore, the specific process of the RBF-PID control algorithm after particle swarm optimization in step S5 is as follows:
[0159] S5-a, initialize the network parameters, and use the optimal position matrix G obtained by the particle swarm algorithm b (t) Decomposition, and substitute the optimized hidden layer node center value c, hidden layer node width value b, and network output layer weight w into the RBF neural network;
[0160] S5-b, collect the control voltage of the fertilization control system at the current moment and the actual flow output value of the system, and form the input vector x with the control voltage difference Δu, the actual flow output value y, and the flow deviation value e before and after, x = [Δu(t), y(t), e(t)], use the vector x as the RBF neural network input signal, and calculate the network output as follows:
[0161]
[0162] Among them, y m is the flow identification output of the network, in L / min; x = [Δu(t), y(t), e(t)], Δu is the control voltage difference between the two times, in V; y is the actual flow output value measured by the flow sensor of the current system, in L / min; e is the flow deviation value, in L / min; c j is the center value of the hidden layer node after particle swarm optimization, b j is the hidden layer node width value after particle swarm algorithm optimization;
[0163] S5-c, update PID parameters according to network output results
[0164]
[0165] Among them, K P , K I , K D They are the proportional coefficient, differential coefficient and integral coefficient of PID control respectively; ηP , η I , η D are the learning rates of the proportional coefficient, differential coefficient and integral coefficient, respectively, all of which are constants; Δu is the control voltage difference between the previous and the next two times, in V; e(t) is the liquid fertilizer flow error at the current moment, in L / min; e(t-1) is the liquid fertilizer flow error at the previous moment, in L / min; e(t-2) is the liquid fertilizer flow error at the previous two moments, in L / min;
[0166] S5-d, calculate the control voltage of the system at the current moment according to the following formula, and output it through the control system;
[0167]
[0168] Among them, U is the control voltage that the system should output at the current moment, in V;
[0169] S5-e, update the network parameter value according to the following formula:
[0170]
[0171] Among them, c j0 , b j0 ,ω j0 are the hidden layer node center value, hidden layer node width value, and output layer weight value of the RBF neural network at the initial moment; η is the network learning rate, 0<η<1; y is the actual output flow of the system, in L / min; y m is the identified flow rate of the neural network, in L / min;
[0172] S5-f, repeating steps S5-b to S5-e until the system switches to another control algorithm or the variable fertilization operation is completed;
[0173] like Figure 3 As shown, the liquid fertilizer variable fertilization control system includes the following key components:
[0174] There are three liquid fertilizer tanks, namely liquid nitrogen fertilizer tank, liquid phosphate fertilizer tank and liquid potassium fertilizer tank. There is a liquid fertilizer outlet at the bottom of the liquid fertilizer tank to ensure the smooth discharge of liquid fertilizer;
[0175] A clean water tank stores a sufficient amount of clean water, and a liquid outlet is provided at the bottom of the tank body to ensure smooth discharge of the clean water;
[0176] Proportional valves, there are 4 proportional valves in total, which are installed at the outlets of each liquid fertilizer tank and clean water tank respectively to control the volume of discharged liquid. The feedback signal type of the proportional valve used is an analog current signal;
[0177] The second fertilizer storage chamber is used to mix the corresponding amount of liquid fertilizer and clean water, and also serves to store the liquid fertilizer required for the next fertilization plot in advance;
[0178] The first fertilizer storage room is used to store the liquid fertilizer required for the current plot of land and is directly connected to the liquid fertilizer pump through a pipeline;
[0179] A valve-controlled hydraulic motor system is used to provide power to the liquid fertilizer pump, wherein the proportional valve input signal is a voltage signal, and the hydraulic motor output shaft is connected to the liquid fertilizer pump through a rigid coupling;
[0180] A liquid fertilizer pump, used for discharging liquid fertilizer in the first fertilizer storage chamber;
[0181] The flow sensor is used to detect the real-time output flow of the system and is installed at the outlet of the liquid fertilizer pump.
[0182] Through the description of the above implementation modes, those skilled in the art can clearly understand the specific implementation process of the present invention. However, it should be pointed out that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention is described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the above embodiments, or replace some of the technical features therein by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A liquid fertilizer variable fertilization control method with automatic proportioning and algorithm adaptive switching, It is characterized in that The control method comprises the following steps: S1, calculating the liquid fertilizer application volume of the current and next variable fertilization unit plots; According to the fertilizer amount information F in the fertilizer prescription map carried by the liquid fertilizer variable fertilization control system, the target fertilizer weights of nitrogen, phosphorus and potassium for the current and next variable fertilization unit plots are obtained, and the application volumes of nitrogen, phosphorus and potassium liquid fertilizers for the current and next plots are calculated respectively; S2, automatically proportioning and delivering the liquid fertilizer required for the current plot; According to the analog current signal size and time T of each proportional valve 1 The inner box is controlled by a mathematical relationship of the corresponding liquid volume, and the nitrogen, phosphorus, and potassium liquid fertilizers and the corresponding volume of clean water required for the current plot are discharged into the second fertilizer storage chamber to fully mix the fertilizers and clean water. Then, the pipe switch under the second fertilizer storage chamber is opened to transport all the prepared liquid fertilizers to the first fertilizer storage chamber; S3, automatically proportioning and delivering the liquid fertilizer required for the next plot of land; After the liquid fertilizer of the initial plot is transported to the first fertilizer storage chamber, the nitrogen, phosphorus and potassium liquid fertilizer stock solutions and the corresponding amount of clean water required for the next plot are mixed and discharged into the second fertilizer storage chamber. When it is detected that the liquid fertilizer variable fertilizer spreader enters the next plot, it is discharged into the first fertilizer storage chamber to apply the liquid fertilizer to the corresponding plot; S4, real-time calculation of the target fertilization flow value and fertilization flow error of the control system; According to the actual operating speed of the liquid fertilizer variable fertilizer spreader and the volume of the liquid fertilizer to be applied to the land in the first fertilizer storage chamber, the target fertilization flow value of the current system is calculated; S5, select the initial control algorithm of the system and calculate the disturbance magnitude; The control system is equipped with two different control algorithms, namely the RBF-PID control algorithm optimized by the particle swarm algorithm and the sliding mode control algorithm SMC. At the initial moment, the system uses the RBF-PID control algorithm optimized by the particle swarm algorithm to control the proportional valve under the first fertilizer storage chamber, and at the same time, calculates the disturbance magnitude Δq of the current control system. S6. Switch the control algorithm used by the system according to the size of the disturbance; According to the calculated disturbance magnitude Δq, the When the disturbance is When , the system continues to use the RBF-PID control algorithm optimized by the particle swarm algorithm to control the proportional valve under the first fertilizer storage chamber; S7, controlling the update of target fertilizer application amount information of the control system; The fertilizer spreader determines the current location of the working field according to the GPS module. When it detects that the working field has changed, it updates the target fertilizer weight of the control system and repeats steps S1 to S6 until the variable fertilization operation is completed. When the operation is completed, the proportional valves under each liquid fertilizer tank and the clean water tank are controlled to stop the discharge of the liquid in the tank.
2. The liquid fertilizer variable fertilization control method according to claim 1, It is characterized in that The mathematical relationship of the proportional valve in step S2 is obtained through a calibration test, and the specific steps are as follows: S2-1. According to the analog current value signal range of the proportional valve, take 10 analog current values at equal intervals, input the above 10 current signal values to the proportional valve in order from small to large, and use a measuring cylinder to measure the time T respectively. 1 The volume of liquid nitrogen fertilizer, liquid phosphate fertilizer, liquid potash fertilizer and clean water discharged by the internal proportional valve is repeated three times at each measurement point to take the average value, and the corresponding discharged liquid volume value under each current signal value is recorded respectively; S2-2. Considering the hysteresis loop of the proportional valve, input the above 10 current signal values into the proportional valve in descending order, and measure the time T with a measuring cylinder. 1 The volume of liquid discharged by the internal proportional valve is measured three times at each measurement point and the average value is taken. The volume of liquid discharged at each current signal value is recorded respectively. S2-3, respectively use the obtained current signal value and liquid volume value as horizontal and vertical coordinates, and use a linear curve to fit the data points to obtain the relationship between the magnitude of the simulated current signal of each proportional valve and the time T 1 The mathematical relationship between the volume of nitrogen fertilizer, phosphate fertilizer, potassium fertilizer and clean water discharged from the internal liquid fertilizer tank.
3. The liquid fertilizer variable fertilization control method according to claim 1, It is characterized in that The time T in step S2 1 The requirements are as follows: Among them, T 1 is the measurement time, in seconds; D is the side length of the variable fertilization unit, in meters; v max It is the maximum travel speed of the fertilizer spreader during operation, in m / s.
4. The liquid fertilizer variable fertilization control method according to claim 1, It is characterized in that The calculation formula of the system disturbance Δq in step S5 is as follows: Among them, Δq is the system disturbance, in L / min; ΣT is the total sampling time, in s; T 2 The time taken by the fertilizer spreader to pass through the variable fertilization unit plot, in seconds; q 1 ,q 2 ,q n are the instantaneous flow values measured in the first, second, and nth sampling time, in L / min; t 1 ,t 2 ,t n They are the first, second and nth sampling times respectively, in seconds; and the requirements are: Among them, t s is the set sampling time, in seconds; D is the side length of the variable fertilization unit, in meters; v is the current speed of the fertilizer spreader, in meters per second.
5. The liquid fertilizer variable fertilization control method according to claim 1, It is characterized in that The specific process of the sliding mode control algorithm SMC in step S5 is as follows: S5-1. Derive the equilibrium equation of the valve-controlled hydraulic motor variable fertilization system and remove the high-order differential terms; S5-2, real-time acquisition of the current target fertilization flow rate q(t) of the system and calculation of the real-time flow error e(t) of the system, and calculation of the sliding surface function of the current system through the balance equation of the valve-controlled hydraulic motor variable fertilization system; S5-3, determine the sliding mode reaching law; Among them, a and p are constants greater than 0; S5-4, calculate the control voltage value that the current system should input to the proportional valve according to the derived control voltage expression equation, and realize the control of the proportional valve by the sliding mode control algorithm SMC; S5-5. Repeat steps S5-2 to S5-4 until the system switches to another control algorithm or the variable fertilization operation is completed.
6. The liquid fertilizer variable fertilization control method according to claim 1, It is characterized in that The specific process of RBF-PID control after particle swarm algorithm optimization in step S5 is as follows: S5-A, build a simulation model of variable fertilization control system, set the total number of particles in the particle swarm algorithm to 50, the maximum number of iterations N = 100, and the position matrix and velocity matrix of each particle at the initial moment are randomly assigned; encode the network structure parameter values in the RBF neural network to be optimized as the initial particle positions in the particle swarm algorithm; S5-B, after the simulation is completed, the following formula is the evaluation index of the particle swarm algorithm optimization effect, and the evaluation index value of each particle at the initial stage is calculated; Among them, T 3 is the simulation duration, in seconds; t is the simulation time, in seconds; r is the unit step target value, r=1; y is the actual output value of the system; S5-C, update the position and velocity of each particle according to the evaluation index of each particle; S5-D, resubstituting the updated position matrix of each particle into the RBF-PID model to obtain the updated evaluation index value, and updating the global optimal position matrix and the optimal position matrix of each particle in the particle swarm algorithm, and jumping to step S5-C to continue the iterative process; S5-E, repeat steps S5-C to S5-D until the current number of iterations n is equal to the maximum number of iterations N; return the optimal position matrix G of the group obtained at this time b (t), substitute the corresponding parameters in the matrix back into the RBF neural network to complete the particle swarm optimization process.
7. The liquid fertilizer variable fertilization control method according to claim 1, It is characterized in that The specific process of the RBF-PID control algorithm after particle swarm optimization in step S5 is as follows: S5-a, initialize the network parameters, and use the optimal position matrix G obtained by the particle swarm algorithm b (t) Decomposition, and substitute the optimized hidden layer node center value c, hidden layer node width value b, and network output layer weight w into the RBF neural network; S5-b, collect the control voltage of the fertilization control system at the current moment and the actual flow output value of the system, and form the input vector x with the control voltage difference Δu, the actual flow output value y and the flow deviation value e before and after, x = [Δu(t), y(t), e(t)], use the vector x as the RBF neural network input signal, and calculate the output of the network; S5-c, update PID parameters according to the network output results; S5-d, calculating the control voltage of the system at the current moment and outputting it through the control system; S5-e, update the network parameter value according to the following formula; Among them, c j0 , b j0 ,ω j0 are the hidden layer node center value, hidden layer node width value, and output layer weight value of the RBF neural network at the initial moment; η is the network learning rate, 0<η<1; y is the actual output flow of the system, in L / min; y m is the identified flow rate of the neural network, in L / min; S5-f, repeating steps S5-b to S5-e until the system switches to another control algorithm or the variable fertilization operation is completed.
8. The liquid fertilizer variable fertilization control method according to claim 1, It is characterized in that Includes the following key components: There are three liquid fertilizer tanks, namely liquid nitrogen fertilizer tank, liquid phosphate fertilizer tank and liquid potassium fertilizer tank. Liquid fertilizer outlets are set at the bottom of the liquid fertilizer tanks to ensure the smooth discharge of liquid fertilizer; A clean water tank stores a sufficient amount of clean water, and a liquid outlet is provided at the bottom of the tank to ensure smooth discharge of the clean water; Proportional valves: There are 4 proportional valves in total, which are installed at the outlets of each liquid fertilizer tank and clean water tank to control the volume of discharged liquid. The feedback signal type of the proportional valve used is an analog current signal. The second fertilizer storage chamber is used to mix the corresponding amount of liquid fertilizer and clean water, and also serves to store the liquid fertilizer required for the next fertilization plot in advance; The first fertilizer storage room is used to store the liquid fertilizer required for the current plot of land and is directly connected to the liquid fertilizer pump through a pipeline; A valve-controlled hydraulic motor system is used to provide power to the liquid fertilizer pump, wherein the proportional valve input signal is a voltage signal, and the hydraulic motor output shaft is connected to the liquid fertilizer pump through a rigid coupling; A liquid fertilizer pump, used for discharging liquid fertilizer in the first fertilizer storage chamber; The flow sensor is used to detect the real-time output flow of the system and is installed at the outlet of the liquid fertilizer pump.
