Design and Experiment of a Targeted Variable Fertilization Control System for Deep Application of Liquid Fertilizer

Abstract

Given the problems of targeted variable deep application of liquid fertilizer in the field, such as low precision, inaccurate fertilization amount, and poor fertilization effect, a targeted variable fertilization control system of liquid fertilizer based on a fuzzy PID algorithm was designed in this study to realize the combination of precise variable fertilization technology and targeted deep-fertilization technology. Specifically, the fertilization equipment and adaptive fuzzy PID control strategy of targeted variable fertilization were designed first. Then, the mathematical model of the targeted variable fertilization control system of liquid fertilizer was established following the requirements of intertillage and fertilization of corn crops. Afterward, the response time and overshoot of the control system were simulated through the Simulink tool of MATLAB software, in which the fuzzy PID control and traditional PID control were compared. Then, the control effect of the targeted variable fertilization control system was verified through field experiments. The test results demonstrated that in the process of simulation analysis, the response time of the variable fertilization control system based on fuzzy PID control was shortened by nearly 5 s on average compared to the system based on traditional PID control, and the error was controlled within 10%. In the field test, the target rate of targeted variable fertilization equipment for liquid fertilizer reached more than 80%, and the control accuracy of the liquid fertilizer application amount also remained above 90%. Finally, the tracking experiment to check the fertilization effect proved that the targeted variable deep-fertilization method of liquid fertilizer could further improve the yield of maize crops under the premise of reducing the fertilization cost. The study provides a feasible solution for the method of precise variable fertilization combined with targeted fertilization.

1. Introduction

Corn is a crucial food crop in the world, widely distributed in the United States, China, Brazil, and other countries [1]. China is a big country in corn planting and production. In 2022, China’s corn planting area reached 43,066 thousand hectares, ranking at the forefront of food crops [2]. Nowadays, agricultural production costs and production demands are increasing day by day. Planting crops requires long-term large-scale application of chemical fertilizers on farmland, and the soil environment is severely polluted [3,4,5]. Moreover, farmland is prone to issues such as soil aggregate structure damage and soil acidification and compaction. Therefore, it is urgent to realize the sustainable development of precision agriculture. Scientific and reasonable fertilization plays an imperative role in protecting the farmland environment, increasing crop yield, saving water resources, and improving soil quality [6]. Precise variable fertilization technology is one of the main methods to achieve precision agriculture and rationally distribute fertilizers [7].

Liquid fertilizer is widely applied due to its advantages of convenient production, flexible proportioning, and high absorption efficiency of crops [8]. The deep application of liquid fertilizer in agricultural production can promote the liquid fertilizer in reaching the roots of crops quickly, effectively enhancing the utilization rate of liquid fertilizer, and avoiding the waste of liquid fertilizer volatilization [9]. The existing equipment for deep application of liquid fertilizer is primarily in the mode of strip deep fertilization and hole deep fertilization [10]. However, the fertilizer consumption of strip deep-fertilization equipment is large and thus easily induces the wasting of fertilizer. There are problems such as inaccurate fertilizer injection position in a hole deep-fertilization machine, which may easily cause the liquid fertilizer to fail in being absorbed and utilized by the crops. Therefore, it is necessary to effectively improve the accuracy and utilization rate of liquid fertilizer application while guaranteeing the reduction in liquid fertilizer application, so as to realize scientific fertilization [11,12].

Under the above background, in this study, target detection technology is integrated with variable fertilization control, liquid fertilizer deep application, and other technologies. Furthermore, the existing deep-fertilization equipment of liquid fertilizer is improved to achieve the targeted and precise deep application of liquid fertilizer. Moreover, the current application status of liquid fertilizer and precise variable fertilization technology are combined. Finally, a targeted variable fertilization control system of liquid fertilizer based on the PID control algorithm is designed. This system can realize the fertilization operation of the variable fertilization system in cooperation with the target identification system, so as to enhance the high efficiency and precise operation level of the deep-fertilization device for liquid fertilizer and effectively reinforce the fertilizer utilization rate.

Many enterprises and researchers around the world have conducted research on the intelligent control of targeted deep-fertilization machines for liquid fertilizer. The AT3000 deep-fertilization machine for liquid fertilizer developed by Blu Jet is equipped with an intelligent fertilization control system based on GPS/GIS [13]. When the machine is working, the system can adjust the amount of fertilizer in real time according to the prescription map information. The opener opens a ditch near the root system of the plant, and the fertilizer injection mechanism injects the regulated liquid fertilizer into the ditch to realize the variable deep application of liquid fertilizer. The speed of the machine is unstable during driving, making the actual fertilization amount uneven and resulting in a substandard fertilization effect. The 8400-type disc-type deep-fertilization machine for liquid fertilizer developed by FAST is equipped with a fertilization control system based on crop detection [14]. The sensor in the intelligent fertilization system detects the plants and sends the control data to the controller, and the fertilizer injection needle injects the corresponding amount of fertilizer into the ditch to realize the precise and deep application of liquid fertilizer. The deep-fertilization machine for liquid fertilizer developed by Nutri Max is mainly adopted for the intertillage and fertilization of upland crops such as corn [15]. The machine is equipped with CASE’s AFS intelligent fertilization system to realize real-time regulation of liquid fertilizer injection. The disc opener creates a fertilizer ditch near the root of the crop plant, the user can adjust the amount of fertilizer applied at any time in the intelligent fertilization system, and the fertilizer injection needle injects the liquid fertilizer regulated by the intelligent fertilization system into the ditch, so as to achieve the precise and deep application of liquid fertilizer. Wang et al. designed a deep-fertilization machine using photoelectric sensors, solenoid valves, hydraulic pumps, and other devices, equipped with a fertilization control system based on target recognition [16]. The electromagnetic valve is controlled to open when the photoelectric sensor recognizes the crop, realizing the precise targeted deep application of liquid fertilizer. Compared with variable fertilization of liquid fertilizer, targeted deep fertilization cannot perform variable fertilization for individual crops or following soil fertilization information. Bai et al. established a corn variable fertilization machine with a variable rate fertilization control system for liquid fertilizer based on the beetle antenna search algorithm [17]. The implement can change the real-time fertilizer application rate upon the implement speed. The variable fertilization control system adopts the beetle antenna search algorithm to adaptively optimize the control parameters and thus effectively improve the transient response performance of the system. Simultaneously, extensive bench experiments were performed, revealing that the algorithm can effectively reinforce the control precision of liquid fertilizer variable fertilization. Zhao et al. [18] designed and optimized the internal structure of a variable fertilization device based on PWM pulse width modulation, realized the variable fertilization process of liquid fertilizer, and designed a fertilizer control system based on a fuzzy control algorithm. However, the study lacked experimental process and data, and did not consider the working state and specific fertilization effect of this variable fertilization control system in actual field work. To sum up, existing deep-fertilization equipment cannot achieve targeted fertilization and variable fertilization at the same time, and innovative research on targeted variable deep fertilization is still required.

The fuzzy PID control algorithm is widely used in other fields of agriculture. Wang et al. [19] designed an adaptive fuzzy PID control algorithm through theoretical analysis and identification of the push-out automatic seedling-picking mechanism to realize the position control of the automatic seedling-picking and transplanting mechanism. They also simulated the control system in the Simulink toolbox in MATLAB software. Concerning the time-domain response and control performance, the designed controller based on the fuzzy PID control algorithm has a better dynamic response and realizes the effective control of the position of the automatic seedling and transplanting mechanism compared with the traditional PID controller. Cao et al. [20] proposed a new self-checking seedling device suitable for automatic transplanting machines and conducted systematic analysis and research on the seedling positioning detection control system. Furthermore, they developed an automatic detection system based on a fuzzy PID control algorithm to achieve more efficient and accurate positioning of seedlings. Yin et al. [21] constructed a compound fuzzy PID controller to adjust the real-time data of PID parameters through a new fractional fuzzy proportional–integral–derivative (PID) controller and a kinematic model. It can be applied in the automatic steering control of a rice transplanter. They also designed a control system for the automatic steering of a rice transplanter based on a fuzzy PID control algorithm. The experiment verified the automatic steering effect of the automatic steering transplanter under different steering angles. The above studies suggest that the fuzzy PID control algorithm has better control performance and is broadly employed in various fields of agricultural engineering.

This study aimed to realize the practical application of the targeted variable fertilization technology in the field. Specifically, this paper designs a directional variable depth fertilization control system for liquid fertilizer based on a fuzzy PID control algorithm. In Section 2, firstly, the targeted variable deep-application machine of liquid fertilizer is designed, the mathematical model of the targeted variable fertilization control system of liquid fertilizer is established, and the virtual simulation is carried out. In Section 3, field tests are carried out on the control system to verify the feasibility, efficiency, and accuracy of deep application of targeted variables of liquid fertilizer. The results are analyzed statistically. In Section 4, the fertilization effect of the targeted variable deep-application machine of liquid fertilizer is tested, and through long-term experimental tracking, it is verified that this system could effectively reduce the consumption of liquid fertilizer. Section 5 gives some conclusions. This study provides enlightenment for the research on directional deep fertilization of liquid fertilizer, drives the promotion of the technology of deep fertilization of liquid fertilizer, and helps to reduce the amount of local fertilizer application and protect land resources [12,17,22].

2. Materials and Methods

2.1. Machine Structure and Working Principle

The targeted deep-fertilization device of liquid fertilizer meets the agronomic requirements of deep application of liquid fertilizer during the inter-cultivation period of corn. The fertilization row spacing of the fertilization device was designed to be adjustable to meet the fertilization needs of corn crops in various regions. During the intertillage period, the corn root system is developed, and the fertilization ditch is kept at 50 mm from the plant, so as to assure that the roots of the corn are not damaged, and the seedlings are prevented from burning. The fertilization device should have the advantages of simple structure, overall portability, high degree of automation, time saving, and high efficiency.

The liquid fertilizer targeted fertilization device was designed according to the above design requirements, as illustrated in Figure 1. The device consists of a frame assembly, a three-point suspension frame, a reducer, a fertilizer supply system, a ditching deep-application mechanism, a ground wheel assembly, and a targeted variable fertilization system. The deep-fertilization mechanism comprises a high-efficiency soil-returning liquid fertilizer opener, fertilizer spraying needle, mounting frame, and connecting rod. The opener is connected to the frame through the installation frame and the connecting rod, and the fertilizer spraying needle is embedded in the tail of the opener. The fertilizer supply system is composed of a liquid fertilizer pump, a liquid fertilizer tank, a filter, an overflow valve, and a fertilizer delivery pipe. The liquid fertilizer pump is the power source of the device, the liquid fertilizer tank is the liquid fertilizer storage, and the overflow valve is the pipeline pressure regulating component. The targeted variable fertilization system consists of a control box, a photoelectric sensor, an electromagnetic flowmeter, an electromagnetic proportional regulating valve, and a solenoid valve. The photoelectric sensor is fixed on the rear side of the connecting rod through the sensor mounting bracket, serving as a plant detection element. The electromagnetic valve group composed of an electromagnetic proportional regulating valve and electromagnetic valve is fixed on the rear side of the connecting rod through nylon cable ties, serving as the actuator for spraying fertilizer.

In this paper, a targeted variable deep-fertilization operation scheme of liquid fertilizer is proposed to improve the utilization rate of liquid fertilizer and the accuracy of deep-application location of liquid fertilizer. It uses a targeted variable fertilization control system and is equipped with a ditch opener with excellent soil-return performance to complete a series of processes such as precise plant detection, liquid fertilizer variable spraying, targeted fertilizer spraying, and soil self-falling and fertilization. The working principle of the targeted variable deep fertilization of liquid fertilizer designed in this paper is exhibited in Figure 2.

2.2. Working Principle of the Targeted Variable Fertilization System

Upon the demand for targeted variable fertilization of liquid fertilizer, the working principle of the targeted variable fertilization control system is to control the liquid fertilizer flow through the closed-loop control of the electromagnetic proportional regulating valve while controlling whether the solenoid valve is opened or not through the real-time detection of the corn plant by the photoelectric sensor. The speed of the vehicle, the amount of fertilization required by the soil, and the opening of the valve core of the electromagnetic proportional regulating valve are positively correlated. The amount of fertilizer sprayed out by the fertilizing needle is detected by the flow meter and fed back to the variable fertilization controller [23].

The process of the photoelectric sensor detecting the corn plants is detailed as follows. When the plant does not block the photoelectric sensor in the working direction, the speed measurement module collects the working speed of the machine tool and transmits it to the variable fertilization controller. The spool of the proportional valve performs a corresponding relative displacement according to the vehicle speed and the fertilization information in the storage unit, and the valve opens to the corresponding opening. At this time, the electromagnetic valve does not work, the liquid fertilizer is not sprayed, and the liquid fertilizer in the fertilizer delivery pipeline flows back to the liquid fertilizer tank by the overflow valve. When the plant in the working direction blocks the photoelectric sensor, the solenoid valve is opened, and the control unit transmits the electric signal to the solenoid valve and the electromagnetic proportional regulating valve. Immediately, the fertilizer spraying needle sprays liquid fertilizer, and the electromagnetic flowmeter counts and measures the fertilization amount of liquid fertilizer in real time. When the liquid fertilizer flow rate is close to the pre-applied fertilizer amount, the solenoid valve is closed to realize the targeted deep application of liquid fertilizer. When the vehicle speed changes, the valve core opening of the electromagnetic proportional valve is adjusted in real time to achieve targeted variable fertilization of liquid fertilizer.

The targeted variable fertilization controller analyzes and compares the current flow of liquid fertilizer, the amount of fertilization required in the current field, and the real-time vehicle speed. The spool opening of the electromagnetic proportional valve is adjusted in real time through the fuzzy PID control strategy. Finally, a closed-loop negative feedback regulation system is formed, so as to fulfill targeted variable fertilization with liquid fertilizer (Figure 3).

2.3. Design of Variable Fertilization Control System

2.3.1. Selection of the Control System Material

Following the requirements of the liquid fertilizer variable fertilization control system, the STM32F103C8T6 minimum system board was selected as the bearing hardware of the variable fertilization control system to enable the variable fertilization control system to efficiently and stably control the targeted fertilizer spraying system for fertilizer spraying operations, as exhibited in Figure 4a. The solenoid valve group is the key control component of the variable fertilization control system. It consists of a solenoid valve and a solenoid proportional regulating valve, and the HD-4WRE-2X solenoid proportional valve (China Henan Nasiwei Valve Co., Ltd., Zhengzhou, China, Figure 4b) was selected. The KVE21PS12N2N651A normally closed solenoid valve (Shanghai Kamoer Fluid Technology Co., Ltd., Shanghai, China) was selected, as displayed in Figure 4c. In this study, the WTGPSBD speed measuring module was used to control solenoid valve opening and closing, as shown in Figure 4d.

2.3.2. Design of Variable Fertilization Control System Based on the Fuzzy PID Control Algorithm

A Fuzzy PID control model of the liquid fertilizer fertilization control system was established to optimize the control process of the liquid fertilizer fertilization system and reduce the response time of the control system as much as possible under the premise of ensuring the control accuracy. The fuzzy PID control algorithm can control the value of the actual fertilization amount near the value of the ideal fertilization amount and control the error within the ideal range as much as possible [19,24]. Fuzzy control has a good control effect and control ability, but fuzzy control is essentially hierarchical control. In the case of a limited number of levels, it is impossible to guarantee that the value of the fertilization amount can be stabilized around the ideal value, and it is easy to oscillate and deviate from the ideal value. Therefore, the PID algorithm can be combined with the fuzzy control algorithm, allowing it to not only adapt to complex systems but also improve the accuracy of the control system while eliminating errors [21,25].

The input signal of the control model of the liquid fertilizer variable fertilization control system in this paper is the real-time vehicle speed collected by the speed measurement module and the amount of fertilization required by the soil. The amount of fertilizer required for the soil can be obtained from the fertilization prescription map. After these signals are converted by the system, the controller transmits the calculated high- and low-level signals to the electromagnetic proportional control valve. The electromagnetic proportional regulating valve controls the opening of the valve. Then, the photoelectric sensor identifies the crops as a fertilization start switch to achieve the purpose of variable fertilization. Finally, the flow rate is fed back to the controller through the flowmeter in the control system, and the solenoid valve of the liquid fertilizer flow rate is closed to stop fertilization.

Considering that the continuous signals involved in the PID control algorithm cannot be represented when using an STM microcontroller for control, it is necessary to discretize these continuous signals. In other words, the PID control algorithm should be discretized. The control system in this paper combines the read fertilization data and other information to control the actual fertilization amount to the ideal fertilization amount through the PID algorithm while minimizing the error and system response time as much as possible [26].

Upon the system requirements, closed-loop control was adopted in this paper to control the electromagnetic proportional control valve. The system consists of a PID regulating controller and an electromagnetic proportional regulating valve. The control quantity is formed by linear combination of e(t) according to proportion, integral, and differential, so as to control the electromagnetic proportional regulating valve. Among them, the control system constitutes a bias equation [27]:

$$e\left(t\right)=u_0\left(t\right)-u\left(t\right)$$

(1)

$$e_a(t)=K_p\left(e(t)+\frac{1}{T_i}{\displaystyle {\int }_o^{t}e(t)dt+T_d}\frac{de(t)}{dt}\right)$$

(2)

where e(t) denotes deviation; u₀(t) represents the set fertilization value; u(t) is the current fertilization value; e_a(t) indicates the corrected deviation value; K_p refers to the constant of proportionality; T_i signifies the integral time constant; T_d stands for the differential time constant.

The commonly used incremental PID algorithm was employed to prevent the system from generating too much calculation and errors in the calculation process [28]. The opening of the electric proportional regulating valve in the fertilization actuator is controlled by ∆u(k), where K_p is the constant of proportionality; K_i denotes the integral coefficient, K_i = T/T_i; K_d refers to the differential coefficient, K_d = T_d/T.

$$u(k)=K_{p}e(k)+K_i{\displaystyle \sum _{i=1}^{k}e(i)+K_d\left[e(k)-e(k-1)\right]}$$

(3)

$$∆u(k)=u(k)-u(k-1)=K_p\left[e\left(k\right)-2e\left(k-1\right)+e\left(k-2\right)\right]$$

(4)

where u(i) represents the system fertilization value at the current i-th moment, and e(i) indicates the input error of the controller at the current moment [19,22]. A parameter adaptive fuzzy PID control strategy was designed following the control system designed in this paper. Among them, the control parameters such as proportion, integral, and differential can be adjusted through the combination of the incremental PID algorithm and fuzzy control strategy.

The fuzzy PID controller was based on the fuzzy reasoning mechanism and the PID regulator of the tuning parameters to realize the control function (Figure 5). The input amount is the error e and the error change rate e_c between the actual fertilization amount and the pre-applied fertilizer amount. The output is the variation ∆K_p, ∆K_i, and ∆K_d of the three parameters K_p, K_i, and K_d in the conventional PID control algorithm. The relationship between the three parameters of the conventional PID controller and the error about the amount of fertilization and its error rate of change was explored. Then, the three parameters were adjusted using fuzzy reasoning [29].

The core control strategy of the fuzzy PID controller is the conditional statement “IF A and B THEN C and D and E” [30]. The basic principles of fuzzy control rules are described as follows. When the error between the ideal value of the liquid fertilization amount and the actual value of the liquid fertilization amount is large, the output of the fuzzer should be as large as possible, and the focus of the output is to eliminate the error. When the error between the ideal value of the liquid fertilization amount and the actual value of the liquid fertilization amount is small, the output of the fuzzer is as small as possible, and the focus of the output is to stabilize the fertilization amount near the ideal value and avoid excessive overshoot. The most imperative thing in fuzzy PID control is to create the correct fuzzy rule table and obtain the fuzzy control table for setting the three parameters of K_p, K_i, and K_d. Meanwhile, the values of the three parameters K_p, K_i, and K_d in the PID algorithm are adjusted in real time [31,32]:

$$K_p=K_{p0}+∆K_p$$

(5)

$$K_i=K_{i0}+∆K_i$$

(6)

$$K_d=K_{d0}+∆K_d$$

(7)

where K_p0, K_i0, and K_d0 denote the last adjusted parameter value in the PID control algorithm; ∆K_p, ∆K_i, and ∆K_d indicate the compensation value obtained by the fuzzy control algorithm; K_p, K_i, and K_d represent the final value obtained through tuning.

According to the characteristics of the control object and the analysis of some experimental results of the control system, the fuzzy sets of the input quantities e and e_c of the control system are divided into seven levels: {negative large, negative medium, negative small, zero, positive small, positive medium, positive large}, denoted as {NB, NM, NS, ZO, PS, PM, PB}. The triangular membership function is selected, and the domain of discourse is set as (−5, 5). Fuzzy control rules, as the main core of fuzzy controllers, are designed in two main ways. One is to continuously summarize the knowledge and operational experience of professional technical personnel, and the other is to continuously test the control system, and then statistically analyze and summarize the input and output data of the system. This article summarizes the operational experience of some technical personnel and the fuzzy rules of relevant scholars, and the fuzzy rule table of the output quantities ∆K_p, ∆K_i, and ∆K_d of the fuzzy PID controller is obtained as follows [26,27,28,29]. The fuzzy control table is shown in

Table 1. Fuzzy inference rule.

e/ec	∆Kp/∆Ki/∆Kd
	NB	NM	NS	ZO	PS	PM	PB	NB
NB	PB/NB/PS	PB/NB/NS	PM/NM/NB	PM/NM/NB	PS/NS/NB	ZO/ZO/NM	ZO/ZO/PS	PB/NB/NS
NM	PB/NB/PS	PB/NB/NS	PM/NM/NB	PS/NS/NM	PS/NS/NM	ZO/ZO/NS	NS/ZO/ZO	PB/NB/NS
NS	PM/NM/ZO	PM/NM/NS	PM/NS/NM	PS/NS/NM	ZO/ZO/NS	NS/PS/NS	NS/PS/ZO	PM/NM/NS
ZO	PM/NM/ZO	PM/NM/NS	PS/NS/NS	ZO/ZO/NS	NS/PS/NS	NM/PM/NS	NM/PM/ZO	PM/NM/NS
PS	PS/NM/ZO	PS/NS/ZO	ZO/ZO/ZO	NS/PS/ZO	NS/PS/ZO	NM/PM/ZO	NM/PB/ZO	PS/NS/ZO
PM	PS/ZO/PB	ZO/ZO/NS	NS/PS/PS	NM/PS/PS	NM/PM/PS	NM/PB/PS	NB/PB/PB	ZO/ZO/NS
PB	ZO/ZO/PB	ZO/ZO/PM	NM/PS/PM	NM/PM/PM	NM/PM/PS	NB/PB/PS	NB/PB/PS	ZO/ZO/PM

.

2.4. Design of Targeted Fertilizer Spraying System

The targeted fertilizer spraying system adopts photoelectric sensors to detect plants as the trigger signal for fertilizer spraying, and the speed measurement module is equipped with a single-chip microcomputer to collect speed information in real time. Combined with the pre-stored and set crop spraying range, the opening and closing timing of the solenoid valve group is calculated to realize the targeted spraying of liquid fertilizer [16,33].

The operating principle of the system is illustrated in Figure 6. The I position is the area when the sensor detects the plant for the first time. Due to the response time of the system, the solenoid valve does not close when the position is I, and the fertilizer injection needle does not spray fertilizer. When the device reaches the II position, the system drives the relay module to pull in, the solenoid valve opens, and the liquid fertilizer is sprayed instantaneously. The device moves on. At position III, the sensor does not detect the plant, the electromagnetic valve does not open, and the fertilizer spraying needle does not spray fertilizer. The operation basis of the targeted fertilizer spraying system is that the photoelectric sensor detects the stem of the corn plant as the trigger information for the fertilizer spraying command; it outputs a high level when the photoelectric sensor detects a plant, the single-chip microcomputer receives the signal to drive the relay to pull in, and the solenoid valve opens. This process requires the sensor to continuously detect the plant. The position of the corn plant should be accurately detected during the forward process of the liquid fertilizer targeted variable fertilizer applicator. The E18-D80NK photoelectric sensor (China Hufeng Machine Tool Electric Manufacturing Factory, Yueqing, China) was selected, as shown in Figure 7.

2.5. Model Establishment of Variable Fertilization Control System

2.5.1. Control System Transfer Function

A transfer function model of the variable fertilization control system was established to optimize the control process of the liquid fertilizer variable fertilization system and reduce the response time of the control system. The control model of the liquid fertilizer variable fertilization control system in this paper takes the real-time vehicle speed and target fertilization amount collected by the speed measurement module as input. After conversion, the controller sends the electric signal to the electromagnetic proportional regulating valve, and the proportional valve controls the opening of the valve. The output of the system is the flow rate of liquid fertilizer [34].

The flow is fed back to the controller through the flow meter in the control system block diagram, and the controller performs closed-loop negative feedback control. The key control object of the variable fertilization control system is the electromagnetic proportional control valve. The control of the fertilization amount is realized by controlling the opening of the spool magnet of the electromagnetic proportional control valve. Therefore, the control model of the electromagnetic proportional regulating valve should be established. Neglecting the friction in the regulating valve, the dynamic model of the electromagnetic proportional regulating valve can be expressed as [35]:

$$F_c=k_{i}i=B_g\pi D{N}_{e}i=K_{e}i$$

(8)

$$u_e=\left(Rc+r_c\right)i+L\frac{di}{dt}+K_e\frac{dx}{dt}$$

(9)

$$F_e=m\frac{d^{2}x}{d{t}^2}+B\frac{dx}{dt}+Kx$$

(10)

where F_e indicates the ferromagnetic force of the valve, N; k_i denotes the solenoid valve current, mA; B_g represents the air gap magnetic induction intensity, T; D is the mean diameter of the coil, mm; N_e signifies the number of coil windings, turns; i refers to the carrying current, A; u_e means the coil voltage, V; Rc stands for coil resistance, Ω; r_c is the internal resistance of the amplifier, Ω; L reflects the coil inductance, H; K_e is the coil-velocity-induced back electromotive force constant; m reveals the quality of the spool; B demonstrates the viscous damping coefficient of the solenoid valve; K is the elastic stiffness.

By performing the Laplace transform on the above formula, the transfer function of the electromagnetic proportional control valve can be expressed as [17,32]:

$$F_1(t)=\frac{1}{(m{t}^2+Bt+K)}$$

(11)

Figure 8 suggests that the input of the feedback channel of the variable fertilization system is the real-time flow rate read from the flow meter, and the feedback channel outputs a voltage signal to the controller. After the conversion of the controller, the speed of the input system is compared with the target fertilization amount and adjusted to realize the negative feedback control of the control system [17].

$$Q=\mu lv\cdot {10}^{-4}$$

(12)

where Q denotes the amount of liquid fertilizer output by the variable fertilization control system, L/min; v is the vehicle speed, m·s⁻¹; μ denotes the target fertilizer rate for each crop entered into the system; l represents the length of the targeted fertilization area, mm.

The flow feedback function in the control model is expressed as [17,20]:

$$K(t)=\frac{v(t)}{Q(t)}=\frac{450}{\mu l}{e}^{-\tau t}$$

(13)

where τ indicates the transmission delay time of the feedback signal; t is the complex variable of the transfer function after the Laplace transform; K represents the feedback link of the transfer function.

Since the electromagnetic proportional control valve selected in this study has a linear relationship between the opening and the flow rate under constant pressure conditions, the relationship between the flow rate and the opening can be expressed by the following formula. The following formula is also the transfer function of valve opening and flow [22]:

$$F_2(t)=\frac{Q(t)}{Z(t)}=K_s$$

(14)

where Z(t) denotes the Laplace transform function of the axis displacement of the electromagnetic proportional control valve.

Regarding the relationship between the driving module of the system and the input voltage and output voltage, the transfer function of the driving module is expressed as [24,33]:

$$F_3(t)=\frac{U_o}{U_i}=K_u\cdot e^{-\tau t}$$

(15)

where K_u indicates the amplification factor of the converter, and the value is 2 according to the system selection; U_o and U_i represent the output and input voltage signals of the drive module, V.

Since the voltage signal of the drive module in the variable fertilization system is delayed less than 0.05 s during transmission, the influence of the delay τ on the system is ignored. The control model F(t) of the electromagnetic flow regulating valve and the closed-loop feedback control transfer function F₀(t) of the variable fertilization control system are established upon the above comprehensive analysis of the system, research material selection, and agronomic design requirements [25].

$$F(t)=F_1(t)F_2(t)F_3(t)$$

(16)

$$F_0(t)=\frac{F(t)}{1+K(t)F(t)}=\frac{25}{90t^2+0.1t+30}$$

(17)

2.5.2. Control System Model Simulation

Simulink is a digital simulation tool under the MATLAB environment. It is an integrated environment for dynamic system modeling, simulation, and analysis. Compared with the method of modeling differential equations or differential equations under traditional simulation software, Simulink has the characteristics of intuition, convenience, and flexibility. The fuzzy control and PID control parts of the variable fertilization control system were simulated using MATLAB/Simulink to verify the accuracy and effectiveness of the variable fertilization control system based on fuzzy PID [36,37,38]. The simulation model of the PID control system and the fuzzy PID control system was simply built for comparison, and the block diagram of the simulation system was drawn, as presented in Figure 8 and Figure 9.

2.6. Design of Variable Fertilization Control Software

A targeted variable fertilization control software was designed, written by QT Creator software, involving the software interface design of the control system and the writing of the control process. According to the design requirements of the variable fertilization system, the system has two different interfaces: (1) inputting the amount of pre-fertilization and various data when the machine is not working; (2) displaying the operation and fertilization information of the machine when it is working [39,40,41,42].

When the implement is at rest, the targeted variable fertilization implement is assembled. At this time, the software interface is exhibited in Figure 10. Among them, the target variable fertilization system is not running, and the working status is displayed in red. After the targeted variable fertilization equipment is assembled, the fertilization information is imported into the software. Clicking the “Ready” button, the software displays the pre-fertilization amount at “Output per plant” following the soil fertilization information stored in the system. Clicking the start button to run the system, the software displays real-time vehicle speed at “START”, as well as the last fertilization amount and the cumulative fertilization amount since the system started in real time. Moreover, the system counts the number of fertilization times. If the movement of the machine tool is relatively stable and the system is working normally, the software working status will be displayed in green. The software also calculates and displays the current working mileage in real time following the information sent by the speed measurement module and the GPS signal for the user’s reference, as illustrated in Figure 11.

2.7. Experiment and Design

At the Xiangyang Farm of Northeast Agricultural University, the fertilizer spraying performance test, the position accuracy test, and the fertilizer amount accuracy test of the targeted variable deep-fertilization machine of liquid fertilizer were conducted to detect the actual situation of the liquid fertilizer targeted variable deep-fertilization machine working in the field.

The control system was tested for corn crops, the experimental conditions were good, the weather was clear, and the soil was suitable for the normal operation of the fertilization machine. The test was powered by a Diduo 704 tractor, using a 60-type three-cylinder plunger pump as the power source of the fertilizer spraying system, a decompression and overflow valve to balance the pressure of the fertilizer pipeline, and a T4-type cylindrical gear parallel reversing reducer. The test equipment mainly includes a steel ruler (range: 80 cm; accuracy: 1 mm), measuring tape (range: 10 m; accuracy: 1 mm), measuring cylinder (range: 50 mL; accuracy: 0.1 mL), calibration rod, and digital camera. The controller adopts an Intel processor, with 4G DDR3 memory and resistive touch screen.

The test valve group was composed of an electric proportional control valve and a solenoid valve, serving as the main control object of the experiment. The field experiments were performed with clean water to prevent environmental pollution.

The test area of the farm was 40 m long, the average corn plant spacing was 0.5 m, and the average row spacing was 0.65 m. Different tests need to be performed on different ridges to prevent the previous test results from affecting this test.

The test plan is described as follows.

• Experiment on spraying performance of liquid fertilizer targeted variable deep applicator

The liquid fertilizer targeted deep application device is intermittently spraying fertilizer, and it is difficult to accurately measure the amount of single spray fertilizer during field operations. Therefore, the fertilizer spraying performance test of the device was performed to explore the influence of liquid fertilizer pump pressure on the amount of fertilizer sprayed and determine the pressure of the liquid fertilizer pump under the amount of fertilizer sprayed according to agronomic requirements. Upon the analysis of the operation process of the targeted fertilizer spraying system, the opening time of the solenoid valve t4 = 100 ms was used as the quantitative factor, the pressure of the liquid fertilizer pump was adopted as the experimental factor, and the amount of fertilizer sprayed was employed as the experimental index. Four levels of liquid fertilizer pump pressure were selected as 0.2, 0.4, 0.6, and 0.8 MPa. Each level test was repeated 3 times, and the average value was taken as the test result.

2.

Experiment on the accuracy of fertilization position of liquid fertilizer targeted variable deep applicator

The test of the accuracy of the fertilization position of the liquid fertilizer targeting variable deep applicator was conducted with the operation speed as the test factor and the target rate as the test index. Four levels of operating speed were selected as 0.4, 0.6, 0.8, and 1.0 m·s⁻¹. Each level test was repeated 3 times, and the average value was taken as the test result. Before the test, the pressure of the liquid fertilizer pump and the operating depth were first adjusted to guarantee that the operating depth and the amount of fertilizer sprayed met the agronomic requirements. Moreover, 20 m was selected as the length of the test area, and 5 m before and after the test area were set as buffer zones. The statistical method was to divide the test area into 4 test plots along the operating direction of the device, with an area of 0.4 m (corn planting ridge width) × 2 m. The interval between each test plot was 0.5 m. The target rate is a vital index to quantify the targeted fertilizer spraying performance of the device and characterize the performance of the targeted fertilizer spraying device. The method for judging the performance of targeted fertilizer spraying is exhibited in Figure 12. With the corn plant as the center, a circle with a radius of 50 mm was utilized to spray fertilizer in this circular area. Beyond this area is wrong fertilization or missed fertilization. The actual situation of the operational performance test and the division of the test area are illustrated in Figure 13.

The target rate statistical method is detailed as follows. The number n₀ of corn plants in each test plot before operation was recorded, the number of corn plants fertilized to the target in each test plot after operation as n₁ was counted, and the target rate was the ratio of the number n₁ of targeted spraying plants to the total number n₀ of plants.

3.

Experiment on the accuracy of fertilization amount of liquid fertilizer targeted variable deep applicator

The liquid fertilizer targeted variable deep applicator changes the amount of fertilization at any time following the fertilization information when working in the field. Therefore, the accuracy test of the fertilization amount of the liquid fertilizer targeted variable deep-application machine was performed to explore the control accuracy of the liquid fertilizer targeted variable fertilization system. According to the agronomic requirements of corn planting, four groups of pre-fertilization amount of 10, 20, 30, and 40 mL were selected; the length of the fertilization block was l = 40 m; the machine speed was the test factor; the actual average amount of fertilizer sprayed per plant was the test index. Before the test, the number of corn plants in the four test plots was counted as 53, 59, 63, and 60, respectively.

In the test area, the driving speed of the liquid fertilizer targeted variable deep applicator was kept at four levels of 0.4, 0.6, 0.8, and 1.0 m·s⁻¹ to conduct the variable speed experiment. During the test, the machine tool speed was first accelerated to 0.4 m·s⁻¹ and maintained, and then accelerated to 0.6 m·s⁻¹ and maintained after driving for 10 m. After the machine travelled at a speed of 1 m·s⁻¹, it travelled to the end position of the test block and turned off the fertilization system. In the test, the fertilizer measuring container was hung under the fertilizer spraying needle, and the test personnel replaced the fertilizer measuring container every time the speed was changed. After the test, measuring instruments such as measuring cylinders were adopted to measure the amount of liquid fertilizer in each fertilizer testing container. The average amount of total liquid fertilizer was the test result of the actual amount of fertilizer applied to each plant.

The main parameters of the liquid fertilizer targeted variable deep-application machine are listed in

Table 2. Main technical parameters of liquid fertilizer targeted deep-application device.

No.	Project	Features and Parameters
1	Mounting method	Three-point suspension
2	Dimensions/mm	2000 × 1000 × 1400
3	Number of job lines	1
4	Adapt to line spacing/mm	2000
5	Working width/mm	2000
6	Working speed/(m·s−1)	0.4~1.0

.

3. Results and Discussion

3.1. Simulation Results and Analysis of Fertilization Control System

Using the Simulink simulation tool in MATLAB, the dynamic simulation of the targeted variable fertilization control based on the fuzzy PID control algorithm was performed to analyze the performance of the control system. At t = 0, the value of the input signal was set to 1, and the simulation was run to analyze the output waveform of the system signal, as shown in Figure 14. The error of the control system was stable within 10%, and the instantaneous response time of the system was reduced as much as possible. The system tracking curve for a given step input signal is exhibited in the figure. In steady-state error analysis, the time within the error band of the corresponding step to the final value of ±5% is usually calculated [43]. The simulation curve demonstrated that the overshoot of the PID strategy control system was 20%, and the adjustment time was about 3.2 s; the overshoot of the fuzzy PID strategy control system was 9%, and the adjustment time was about 6.7 s. After adopting the fuzzy PID control strategy, the adjustment time of the system was significantly shortened, and the overshoot was reduced.

3.2. Experimental Results and Analysis of Spraying Performance of Liquid Fertilizer Targeted Variable Deep Applicator

The curve of the test results of the liquid fertilizer pump on the amount of fertilizer sprayed suggests that when other operating conditions are constant, the amount of fertilizer sprayed increases with the increase in the pressure of the liquid fertilizer pump, as shown in Figure 15.

3.3. Experiment on the Target Rate of the Machine of Liquid Fertilizer Targeted Variable Deep Applicator

The working depth of the deep-ditching device of the liquid fertilizer targeted variable fertilization machine was 80 mm. Under the condition of setting the fertilizer spraying amount of 30 mL each time, a single factor experiment was performed to analyze the influence of the operating speed of the fertilization device on the target rate. During the test, the four levels of operating speed were selected as 0.4, 0.6, 0.8, and 1.0 m·s⁻¹. Three replicates were performed at each level. Statistical analysis was conducted on the obtained results. The curve of target alignment rate and device operating speed is demonstrated in Figure 16.

As revealed in the target alignment rate at different operating speeds, the average target alignment rate of the device is 84.03%. The change curve of the influence of device operating speed on the target rate in Figure 16 implies that when the operating speed is 0.4–0.8 m·s⁻¹, the target rate increases slowly as the speed increases; when the working speed is between 0.8 and 1.0 m·s⁻¹, the target alignment rate decreases rapidly as the speed increases.

To sum up, the program needs to continuously detect the operating speed and make calculations and judgments when the operating speed is low. However, there is also a certain time delay between the continuous calculation and judgment processes. This part of the time delay accumulates continuously, resulting in prolonged response time and misjudgment by the system. Then, the device sprays fertilizer in advance. As the operation speed increases, the delay phenomenon is alleviated, and the target accuracy increases slowly. Furthermore, the photoelectric sensor cannot continuously detect the plants when the speed is too high. Therefore, after the fertilizer spraying needle has entered the fertilization range, the solenoid valve opens with a delay or even does not open, the fertilizer spraying lag occurs, and the target rate decreases rapidly.

3.4. Experiment on the Accuracy of Fertilization Amount of Liquid Fertilizer Targeted Variable Deep Applicator

From Table 3 and Figure 17, it can be observed that within the range of fertilization amount of no more than 40 mL per plant, the error of fertilization does not exceed 10%. The actual amount of fertilization is slightly lower than the amount of pre-established fertilizer when the amount of pre-established fertilizer does not exceed 20 mL. When the pre-established fertilizer amount exceeds 20 mL, the actual fertilization amount is slightly higher than the pre-established fertilizer amount and the error rate is not higher than 10%. After analysis, the main reasons for the error are summarized as follows:

(1)

When designing the fuzzy PID control strategy, there was a deviation in the parameter setting of the drive components such as the drive motor, and the transmission relationship between some transmission components was ignored. Therefore, the experimental effect deviated from the effect produced by the actual motor drive.

(2)

The action of the electromagnetic proportional control valve lagged behind and the error occurred when the solenoid valve received the control voltage, and it took a certain amount of time for the valve to open. Thus, the coordination between the electromagnetic proportional regulating valve and the solenoid valve was poor. The valve response time can be reduced by replacing the electric proportional regulating valve with one of higher sensitivity and higher performance.

(3)

When the amount of pre-fertilized fertilizer did not exceed 20 mL, the speed of the vehicle was low when starting, and the response of the speed measurement module was not sensitive. As a result, the system response time was longer, and the opening of the electromagnetic proportional control valve was smaller. Under the condition of limited fertilization time, some liquid fertilizers were sprayed late or missed, leading to the actual amount of fertilizer being lower than the pre-applied fertilizer amount.

Table 3. Fertilizing amount under the pressure of the different hydraulic pump.

Test Number	Preset the Amount of Fertilizer Applied to Each Plant/mL	The Number of Plants in the Test Ridge	Moving Speed/(m·s−1)	Measured Average Fertilization Amount/mL	Error/mL	Error Rate/%
1	10	53	0.4	9.68	−0.32	−3.2
			0.6	9.82	−0.18	−1.8
			0.8	9.93	−0.07	−0.7
			1.0	10.21	0.21	2.1
2	20	59	0.4	19.15	−0.85	−4.25
			0.6	20.31	0.31	1.55
			0.8	20.6	2.8	3
			1.0	21.30	1.3	6.5
3	30	63	0.4	28.90	−1.1	−3.67
			0.6	29.13	−0.87	−2.9
			0.8	30.60	0.6	2
			1.0	31.9	1.9	6.33
4	40	60	0.4	36.39	−1.61	−4.03
			0.6	38.95	−1.05	−2.625
			0.8	40.30	0.3	0.75
			1.0	41.10	1.1	2.75

Table 3. Fertilizing amount under the pressure of the different hydraulic pump.

Test Number	Preset the Amount of Fertilizer Applied to Each Plant/mL	The Number of Plants in the Test Ridge	Moving Speed/(m·s−1)	Measured Average Fertilization Amount/mL	Error/mL	Error Rate/%
1	10	53	0.4	9.68	−0.32	−3.2
			0.6	9.82	−0.18	−1.8
			0.8	9.93	−0.07	−0.7
			1.0	10.21	0.21	2.1
2	20	59	0.4	19.15	−0.85	−4.25
			0.6	20.31	0.31	1.55
			0.8	20.6	2.8	3
			1.0	21.30	1.3	6.5
3	30	63	0.4	28.90	−1.1	−3.67
			0.6	29.13	−0.87	−2.9
			0.8	30.60	0.6	2
			1.0	31.9	1.9	6.33
4	40	60	0.4	36.39	−1.61	−4.03
			0.6	38.95	−1.05	−2.625
			0.8	40.30	0.3	0.75
			1.0	41.10	1.1	2.75

Figure 17. Statistical chart of the test results of the accuracy of fertilizer application.

Figure 17. Statistical chart of the test results of the accuracy of fertilizer application.

To sum up, within the range of fertilization amount required by agronomy during the inter-cultivation period of corn, the error rate of the control system of the targeted variable fertilization control system based on the fuzzy PID control algorithm for the amount of fertilization did not exceed 8%, and the target rate of the targeted fertilization equipment equipped with the variable fertilization control system reached 84%. The experimental results revealed that the variable fertilization control system based on fuzzy PID control has good response speed and flow control stability and can meet the control requirements.

4. Tracking Test of Fertilization Effect

4.1. Conditions of the Test Area

In order to test the fertilization effect of the liquid fertilizer targeted variable deep applicator, a tracking experiment on the targeted variable control system was carried out. The test site is located at Xiangyang Farm (127.024 N, 45.529 E), Northeast Agricultural University, Harbin, Heilongjiang Province. It is located in the plain and belongs to the mid-temperate continental monsoon climate zone. The average annual sunshine number is 2475.4 h, the average annual temperature is 5.5 °C, and the average annual precipitation is 640 mm. The soil in the test plot is black soil, with soil bulk density of 1.3 g·cm⁻¹ and average organic matter content of 36.2 g·kg⁻¹ [45]. The area of the test plot was about 0.25 hm, and the test was divided into 16 test blocks, each of which was 15 m long and 6 m wide. A 1 m wide isolation band between cells was set, as shown in Figure 18. Moreover, the plots represented by F₁-F₄ in Figure 18 are shown in

Table 4. Fertilization scheme.

Experimental Group Number	Fertilization Scheme	Liquid Fertilizer Consumption/L
F1	No fertilization	0
F2	Targeted deep fertilization	93.7
F3	Targeted variable deep fertilization	86.5
F4	Traditional deep fertilization	110.3

.

4.2. Test Materials and Conditions

The corn variety is Doctor Jin 825, which has good adaptability and stable yield. The fertilization effect tracking experiment was conducted from June to October. At this time, the average height of corn plants was 270 mm, the average row distance was 650 mm, and the average plant distance was 330 mm [46], as shown in Figure 19.

The corn plants in the experiment were in the topdressing period, and the base fertilizer was applied at the time of sowing. The test environment was relatively good, the weather was sunny, and the average temperature was 26~30 °C. The soil condition of the experimental field was suitable for the normal operation of the fertilization machine.

4.3. Experimental Design

According to the corn fertilization data of the previous year and the suggestions of local field managers, before the experiment, liquid nitrogen fertilizer with 45% nitrogen content was diluted at an 8% ratio, and the amount of liquid fertilizer used for each corn crop was about 20–30 mL [46]. Four fertilization treatment methods were designed, corresponding to four plots, respectively. The F₁ plot did not receive fertilizer. In the F₂ plot, the experiment was conducted on the premise that the variable fertilization control system was turned off. At this time, the targeted variable fertilization software did not access fertilization data, the variable fertilization system did not work, the targeted fertilization system worked normally, and the fertilization machine and tools targeted deep fertilization. In the F₃ plot, the targeted variable deep-application machine of liquid fertilizer developed in this study was used for fertilization. In F₄ plots, the traditional deep strip fertilization method was applied. During the experiment, it was ensured that in addition to different fertilization methods, other experimental factors such as fertilization pump pressure, liquid fertilizer nutrient concentration, and trencher size were the same. During the experiment, a liquid flowmeter was installed on the liquid fertilizer output pipe to read the amount of liquid fertilizer consumed by each plot. The fertilization scheme and liquid fertilizer consumption are shown in Table 4.

In Table 4, the experiment shows that the consumption of liquid fertilizer in experimental group F₃ is significantly lower than that in other groups. The experimental results showed that targeted variable deep application of liquid fertilizer could effectively reduce the application cost of liquid fertilizer. Under the condition of achieving the same fertilization effect, the consumption of liquid fertilizer is reduced to the minimum.

The growth of maize was monitored after the group fertilization experiment. In the monitoring process, the four groups of test blocks had unified weeding, application, and other work carried out to ensure that all experimental plants could develop healthily and normally.

After four months of growth of the corn crops, researchers harvested corn fruits, picked, peeled, threshed, and sampled the corn fruits, calculated the 1000-grain weight of corn fruits in each plot, and tested the fertilization effect of the four groups of fertilization experiments, as shown in Figure 20 and Figure 21.

4.4. Experimental Results and Analysis

After the completion of the threshing work of corn crops, 1000 corn grains produced in different experimental blocks of each fertilization method were randomly selected for drying and the 1000-grain weight of corn grains was calculated. Experiments were conducted in groups three times to obtain the actual yield of corn after different fertilization methods in each plot, as shown in Figure 22.

In Figure 22, the experimental data of the three groups all showed that the corn yield in F₃ of the experimental group was higher, and the fertilization effect of targeted variable deep fertilization of liquid fertilizer was better. Because the corn plants were not fertilized, the nutrients required for the growth of corn crops in the experimental group F₁ were fewer, and the incomplete development of corn led to the low yield of corn. F₂ of the experimental group adopted the targeted quantitative fertilization method without further controlling the application amount of fertilizer, which would make part of the liquid fertilizer not play a role, resulting in a low target ratio of the fertilizer machine and the phenomenon of liquid fertilizer waste. In the experimental group F₄, the traditional deep strip fertilization method was adopted, which had higher consumption of liquid fertilizer and higher cost than other fertilization methods. Excessive application of liquid fertilizer will destroy the soil environment, burning seedlings, and other phenomena. This is also the main reason for the low yield of this experimental group.

5. Conclusions

In this paper, the specific application of the targeted variable fertilization technology in the actual agricultural production process was studied, the liquid fertilizer targeted variable fertilization system was designed, the mathematical model of the targeted variable fertilization control system was established, and the targeted variable fertilization control software was constructed. When designing the target variable fertilization control system model, a virtual simulation comparison was performed between the traditional PID control and the fuzzy PID-based control, and the fertilizer spray performance, target rate, and flow control accuracy of the liquid fertilizer target variable deep applicator were tested. The conclusions can be drawn as follows.

(1) The fuzzy control strategy combined with the PID control algorithm was adopted to optimize the control system, and the control system model of liquid fertilizer variable fertilization was established. Additionally, the PID parameters were adaptively optimized to effectively improve the transient response performance of the system and curtail the system response time and overshoot.

(2) The feedback adjustment process of the targeted variable fertilization control system was analyzed, and the virtual simulation of the variable fertilization control system was performed using MATLAB software. The simulation results demonstrated that the response time of the traditional PID control to achieve steady-state control was 9.8 s, and the overshoot was 20%; the fuzzy PID control response time was 4.8 s, and the overshoot was 10%. Compared with the traditional PID control, the combination of fuzzy control strategy and PID control shortened the system response time by 5 s and lessened the overshoot by about 10%.

(3) With the purpose of verifying the feasibility and accuracy of the variable fertilization control system, a field experiment was conducted to test the performance, the accuracy of the fertilization position, and the control precision of the fertilization amount of the liquid fertilizer targeted variable deep applicator. The test results suggested that at the level of the liquid fertilizer pump pressure not less than 0.6 MPa, within 50 mm around the plant, the accuracy of the fertilization position of the liquid fertilizer targeted variable deep applicator designed in this paper reached more than 80%, which can achieve effective deep fertilization; the variable fertilization control system had a control accuracy of more than 90% for the fertilization amount, satisfying the design requirements of the control system.

(4) To verify the comprehensive fertilization effect of the targeted variable deep-application machine, a five-month fertilization effect tracking experiment was carried out, mainly through the yield of corn crops to reflect the overall fertilization effect of the fertilization machine. After applying fertilizer with different fertilization methods, the fertilization results showed that the targeted variable fertilization machine could significantly reduce the application amount of liquid fertilizer. After drying and threshing the harvested corn crops in each block, the 1000-grain weight of the harvested corn crops was randomly sampled for comprehensive comparison. The results showed that maize crops with targeted variable deep fertilization had higher and more prominent yields. The results of this study provided a feasible solution for targeted variable fertilization of maize crops.