Multivariable collaborative optimization control method and system for low-crushing threshing of soybeans

Abstract

The invention discloses a soybean low-crushing threshing multivariable collaborative optimization control method and system, and relates to the field of agricultural machinery automation control. The method comprises the following steps: acquiring grain harvesting and threshing related data at the current moment; based on grain harvesting and threshing related data, soybean low-crushing threshing multivariable cooperative control is carried out, and the rotating speed of a target roller and the opening degree of a concave grid are solved; and outputting a control instruction according to a solving result to realize online regulation and control of the rotating speed of the roller and the opening degree of the concave sieve. The intelligent level of the grain combine harvester can be effectively improved, high-quality and low-loss harvesting of grain harvesting is achieved, grain saving and loss reduction of grain harvesting are achieved, and agricultural economic benefits are effectively improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of agricultural machinery automation control, and more particularly to a soybean low-breakage threshing multivariable collaborative optimization control method and system. BACKGROUND

[0002] The existing grain combine harvester threshing system mostly adopts PID control or fixed parameter adjustment strategy, which has the following problems: first, multivariable coupling is difficult to handle, the key parameters such as cylinder speed, concave clearance and conveying speed influence each other, and traditional single variable control cannot coordinate and optimize; second, dynamic response is lagging, under the disturbance of feeding amount fluctuation and crop moisture content change, PID parameter setting is difficult, which easily leads to the increase of breakage rate; third, the constraint conditions are ignored, and the cylinder speed and concave clearance have physical safety limits, and traditional control is easy to cause mechanical damage or breakage aggravation due to over-limit. The multivariable collaborative optimization control method can realize multivariable collaborative control of the grain harvesting and threshing system through rolling optimization and feedback correction, but a complete solution has not been constructed for the low-breakage demand of soybean crops in the threshing system. SUMMARY

[0003] Therefore, the present application provides a soybean low-breakage threshing multivariable collaborative optimization control method and system, which is suitable for the parameter regulation and control of the harvester threshing process of soybean crops, and aims to reduce the grain breakage rate and improve the threshing rate.

[0004] In order to achieve the above purpose, the present application adopts the following technical scheme:

[0005] A soybean low-breakage threshing multivariable collaborative optimization control method, comprising the following steps:

[0006] Obtaining the grain harvesting and threshing related data at the current time;

[0007] Based on the grain harvesting and threshing related data, the soybean low-breakage threshing multivariable collaborative control is carried out, and the target cylinder speed and concave screen opening are solved;

[0008] According to the solving result, the control instruction is outputted, and the online regulation and control of the cylinder speed and the concave screen opening are realized.

[0009] Optionally, the grain harvesting and threshing related data includes crop feeding amount F(t), cylinder speed ω(t), concave clearance d(t), crop moisture content H(t) and entrainment loss rate η(t), and breakage rate b(t).

[0010] Optionally, the soybean low-breakage threshing multivariable collaborative control specifically comprises the following steps:

[0011] According to the threshing dynamics model and the conveying-feeding dynamic model, a threshing mechanism model is constructed, which is used to construct the constraint condition mode extrusion breakage;

[0012] The data-driven model is constructed by training the LSTM network according to historical grain harvesting and threshing related data. First, the abnormal values are removed, the multi-sensor data is synchronized according to the time stamp, and the working conditions are labeled according to the crop type, drum speed, concave clearance, crop moisture content, and feeding level. Then, the input / output variables are standardized by Z-Score; then, 10 consecutive time steps are taken as the input window, and the broken rate and entrainment loss rate of the next time step are taken as the output label to generate sample sequences in the supervised learning format; the parameters such as the number of LSTM layers, the number of neurons, the Dropout ratio, and the time window length are adjusted through Bayesian optimization; finally, the model is trained, and the time series data is used to construct the model for multivariate collaborative optimization control.

[0013] The state variables and control inputs are defined, and a discrete-time state space model is constructed.

[0014] The multivariate collaborative optimization control optimization problem of soybean low-breakage threshing is defined, and at each time t, a finite-time domain optimization problem is solved.

[0015] Optionally, the threshing dynamics model is used to describe the relationship between the crop feeding amount F(t), the drum speed ω(t), the concave clearance d(t), the crop moisture content H(t), and the entrainment loss rate η(t) and the broken rate b(t):

[0016]

[0017]

[0018] Where k1, g1, d1, k2, g2, d2 are the correction factors of the model; the conveying-feeding dynamic model is used to describe the relationship between the conveying chain speed v(t), the volume V of the threshing drum of the harvester, the density ρ of the crop to be harvested, and the crop filling rate φ(t) of the system, and the correction factor τ of the drum is as follows:

[0019]

[0020] The constraint φ(t) is within the safety threshold to prevent crushing.

[0021] Optionally, the state variables and control inputs are defined, and a discrete-time state space model is constructed.

[0022]

[0023]

[0024] The state variable is x(t):

[0025]

[0026] The control input is u(t):

[0027] u(t) = [ω(t), d(t), v(t)] T

[0028] The output variable is y(t):

[0029] y(t) = [ω(t), d(t)] T

[0030] is the process noise and is the measurement noise.

[0031] Optionally, a soybean low-breakage threshing multivariable collaborative optimization control optimization problem is defined, and a finite time domain optimization problem is solved at each time t, and the specific calculation process is as follows:

[0032]

[0033] s.t.x(t+k+1|t) = Ax(t+k|t) + Bu(t+k|t), k = 0, …, N p -1

[0034] y min ≤ y(t+k|t) ≤ y max

[0035] u min ≤ u(t+k|t) ≤ u max

[0036] N p is the prediction horizon, N c is the control horizon, b ref , η ref are reference target values of the breakage rate and the loss-in-transport rate, the weight coefficients Q1 and Q2 adjust the priority of the breakage rate and the loss-in-transport rate, R d is the input variation rate penalty term, b(t+k|t) is the breakage rate of the grain harvester at the t+k|t time, η(t+k|t) is the loss-in-transport rate of the grain harvester at the t+k|t time, u(t+k+1|t) is the model error at the t+k|t+1|t time, x(t+k+1|t) is the parameter state of the harvester at the t+k|t+1|t time, x(t+k|t) is the parameter state of the harvester at the t+k|t time, y min is the minimum value of the target output, y max is the maximum value of the target output, u min is the minimum value of the system error, u max is the maximum value of the system error.

[0037] Real-time solving optimization problem, using quadratic programming solver OSQP real-time calculation of optimal control sequence u * (t|t),…,u * (t+N c -1|t), only the first control action u * (t|t) is executed, the rest of the sequence is discarded, the state x(t+1) is measured again at the next time t+1 and the optimization process is repeated to realize the rolling horizon update;

[0038] Constraint processing, hard constraints are directly embedded in the optimization problem, including the speed range of the roller, the gap value range of the concave plate, the speed limit of the conveyor; soft constraints, for constraints that are difficult to strictly satisfy, introduce a penalty term through a slack variable:

[0039] ω min ≤ω(t)≤ω max ,d min ≤d(t)≤d max ,v min ≤v(t)≤v max .

[0040] Where, ω min is the minimum value of the roller speed, ω max is the maximum value of the roller speed, d min is the minimum value of the gap value of the concave plate, d max is the maximum value of the gap value of the concave plate, v min is the minimum value of the conveyor speed, v max is the maximum value of the conveyor speed.

[0041] Optionally, according to the solving result, output control instruction, realize online regulation and control of roller speed and concave plate opening, specifically including the following steps:

[0042] The soybean low-breakage threshing multivariable collaborative optimization control system rolls the target roller speed ω * (t+k|t) and the target concave plate gap d * (t+k|t) are solved, and high-precision and low-delay control is realized through the actuator; ω * (t+k|t) is converted into the PWM duty cycle command I * (t+k|t) of the electro-hydraulic servo motor, d * (t+k|t) is converted into the displacement command S * (t+k|t) of the electric push rod:

[0043] I(t+k|t)=k3ω(t+k|t)+d3

[0044] S(t+k|t)=k4d(t+k|t)+d4

[0045] The synergistic control is realized by the following methods: priority allocation, taking the drum rotation speed as the priority adjustment variable and the recessed plate gap as the auxiliary adjustment variable; when the optimization result requires reducing the rotation speed and increasing the gap at the same time, the rotation speed adjustment is performed first, and the gap adjustment is started after a delay of Δt=0.2s; cross-coupling compensation, a rotation speed-gap coupling model is established:

[0046] Δd comp =k p ·Δω

[0047] k p is a proportional coefficient, and the gap change amount is pre-compensated when the rotation speed is adjusted.

[0048] A soybean low-breaking threshing multivariable synergistic optimization control system uses any one of the soybean low-breaking threshing multivariable synergistic optimization control methods, and comprises:

[0049] A data acquisition module is used to acquire real-time data of a grain harvesting and threshing system of a multivariable synergistic optimization control system of the grain harvesting and threshing system, to acquire the entrainment loss rate, the breaking rate, the feeding amount, the moisture content, the drum rotation speed and the recessed plate screen gap in real time through data communication with an entrainment loss rate sensor, a breaking rate sensor, a feeding amount sensor, a moisture content sensor, a displacement sensor and a rotation speed sensor;

[0050] A controller is used to run a control operation module of a multivariable synergistic optimization control method, to establish a mixed dynamic model of a threshing process, including a threshing mechanism model and a data-driven model; a system overall state space model is defined to define state variables and control inputs, and to construct a discrete-time state space model; based on an optimization objective function and a constraint condition, a finite-time-domain multivariable synergistic optimization control optimization problem is solved;

[0051] An electro-hydraulic servo system is used to control the opening degree of a proportional valve through a PWM signal, to adjust the hydraulic oil flow to drive a drum motor, and to adjust the drum rotation speed in real time;

[0052] An electric push rod is used to drive a ball screw through a step motor to convert rotary motion into linear displacement, to drive the recessed plate to move, and to adjust the recessed plate screen gap in real time;

[0053] A drum variable-speed execution mechanism is used to connect the electro-hydraulic servo system and the mechanical mechanism of the threshing drum;

[0054] A recessed plate screen gap adjustment execution mechanism is used to connect the electric push rod and the mechanical mechanism of the recessed plate gap;

[0055] Data monitoring sensors are used to measure various sensors of the entrainment loss rate, the breaking rate, the feeding amount, the moisture content, the drum rotation speed and the recessed plate screen gap.

[0056] Compared with the prior art, the soybean low-breaking threshing multivariable collaborative optimization control method and system can effectively improve the intelligent level of the grain combine harvester, solve the parameter self-adaptive regulation and control of the grain combine harvester in a complex harvesting environment, realize high-quality and low-loss harvesting of the grain, achieve the goal of grain harvesting with less loss and saving of grain, and effectively improve the agricultural economic benefits. BRIEF DESCRIPTION OF DRAWINGS

[0057] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor on the basis of the provided drawings.

[0058] Figure 1 The figure is a multivariable collaborative optimization control system diagram of the grain harvesting and threshing system of the present application.

[0059] Figure 2 The figure is a multivariable collaborative optimization control flowchart of the grain harvesting and threshing system of the present application.

[0060] Figure 3 The figure is a multivariable collaborative optimization control effect diagram of the grain harvesting and threshing system of the present application. DETAILED DESCRIPTION

[0061] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the protection scope of the present application.

[0062] Embodiment 1

[0063] The embodiments of the present application disclose a soybean low-breaking threshing multivariable collaborative optimization control method, as shown in the figure, comprising the following steps: Figure 2 The figure is a multivariable collaborative optimization control system diagram of the grain harvesting and threshing system of the present application.

[0064] Obtaining the grain harvesting and threshing related data at the current time;

[0065] Based on the grain harvesting and threshing related data, the soybean low-breaking threshing multivariable collaborative control is performed, and the target roller speed and the concave screen opening are solved.

[0066] According to the solving result, the control instruction is output, and the online regulation and control of the roller speed and the concave screen opening are realized.

[0067] Further, the grain harvesting threshing related data includes crop feeding amount F(t), cylinder speed ω(t), concave clearance d(t), crop moisture content H(t), entrainment loss rate η(t), and breakage rate b(t).

[0068] Further, the soybean low-breakage threshing multivariable collaborative control specifically includes the following steps:

[0069] According to the threshing dynamics model and the conveying-feeding dynamic model, a threshing mechanism model is constructed, which is used to construct the constraint condition mode extrusion breakage;

[0070] According to the historical grain harvesting threshing related data, an LSTM network is trained to construct a data-driven model, which is used for multivariable collaborative optimization control;

[0071] State variables and control inputs are defined, and a discrete-time state space model is constructed;

[0072] The soybean low-breakage threshing multivariable collaborative optimization control optimization problem is defined, and at each time t, a finite time domain optimization problem is solved.

[0073] The threshing dynamics model describes the relationship between the crop feeding amount F(t) (kg / s), the cylinder speed ω(t) (r / min), the concave clearance d(t) (mm), the crop moisture content H(t) (%), the entrainment loss rate η(t) (%), and the breakage rate b(t) (%):

[0074] η(t)=f1(ω(t),d(t),F(t),H(t))

[0075] b(t)=f2(ω(t),d(t),F(t),H(t))

[0076] Where f1, f2 are nonlinear functions, which are obtained by fitting experimental data:

[0077]

[0078]

[0079] The conveying-feeding dynamic model describes the relationship between the conveying chain speed v(t) (m / s) and the crop filling rate φ(t) (%) in the cylinder, which avoids accumulation or empty load, and the model is as follows:

[0080]

[0081] φ(t) is constrained in the safe interval [70%, 80%] to prevent extrusion breakage.

[0082] Due to the complexity of soybean material properties (e.g. uneven spatial distribution of moisture content) and field environment (e.g. feeding fluctuation caused by terrain undulation), the LSTM network is trained by the past 10 seconds of F(t), ω(t)d(t), H(t), η(t), b(t) sequences to predict the breaking rate trend b(t+1:t+5) in the next 5 seconds as the reference of the multivariable collaborative optimization control optimization. pred (t+1:t+5) as the reference of the multivariable collaborative optimization control optimization.

[0083] Define state variables and control inputs, construct a discrete-time state-space model:

[0084]

[0085]

[0086] State variables are x(t):

[0087]

[0088] Control inputs are u(t):

[0089] u(t) = [ω(t), d(t), v(t)] T

[0090] Output variables are y(t):

[0091] y(t) = [ω(t), d(t)] T

[0092] Process noise and measurement noise.

[0093] Define the soybean low-breaking threshing multivariable collaborative optimization control optimization problem, at each time t, the controller solves the following finite-time domain optimization problem:

[0094]

[0095] s.t.x(t+k+1|t) = Ax(t+k|t) + Bu(t+k|t), k = 0, …, N p -1

[0096] y min ≤ y(t+k|t) ≤ y max

[0097] u min ≤ u(t+k|t) ≤ u max

[0098] N p is the prediction horizon, N c is the control horizon, b ref , η refReference target values of breakage rate and entrainment loss rate, respectively, weight coefficients Q1, Q2 adjust the priority of breakage rate and entrainment loss rate, R penalizes control input variation, R d Input variation rate penalty term, avoid frequent adjustment.

[0099] Real-time solution of optimization problem, quadratic programming solver OSQP is used to calculate the optimal control sequence u * (t|t),…,u * (t+N c -1|t), only the first control action u * (t|t) is executed, the rest of the sequence is discarded, the state x(t+1) is measured again at the next time t+1 and the optimization process is repeated, realizing the rolling horizon update.

[0100] Constraint processing, hard constraints are directly embedded in the optimization problem, including roller speed range, concave plate gap value range, conveying speed limit; soft constraints, for constraints that are difficult to strictly satisfy (such as instantaneous breakage rate exceeding limit), introduce penalty term through slack variable to avoid feasibility loss.

[0101] ω min ≤ω(t)≤ω max ,d min ≤d(t)≤d max ,v min ≤v(t)≤v max

[0102] Soybean low-breakage threshing multivariable collaborative optimization control system Rolling optimization solves the target roller speed ω * (t+k|t) and target concave plate gap d * (t+k|t) in future time, through the actuator to realize high precision, low delay control. Convert ω * (t+k|t) to PWM duty cycle command I * (t+k|t) of electro-hydraulic servo motor, convert d * (t+k|t) to displacement command S * (t+k|t) of electric push rod.

[0103] I(t+k|t)=k3ω(t+k|t)+d3

[0104] S(t+k|t)=k4d(t+k|t)+d4

[0105] To avoid the instability of threshing caused by the asynchronization between the roller speed and the concave gap adjustment (such as the blockage caused by the lag of the speed reduction behind the gap increase), the following methods are used to realize the collaborative control of the two: priority allocation, with the roller speed as the priority adjustment variable and the concave gap as the auxiliary adjustment variable; when the optimization results require both speed reduction and gap increase, the speed adjustment is performed first, and the gap adjustment is started after a delay Δt = 0.2s. Cross-coupling compensation, a speed-gap coupling model is established

[0106] Δd comp = k p · Δω

[0107] k p is a proportional coefficient, which pre-compensates the gap change amount when adjusting the speed.

[0108] The soybean low-breakage threshing multivariable collaborative optimization control system is shown in Figure 1 , mainly including a data acquisition module, a controller, an electro-hydraulic servo system, an electric push rod, a roller variable speed execution mechanism, a concave screen gap adjustment execution mechanism, a data monitoring sensor and the like:

[0109] The data acquisition module is used for the real-time data acquisition of the grain harvesting and threshing system of the multivariable collaborative optimization control system of the grain harvesting and threshing system, and is in data communication with the entrainment loss rate sensor, the breakage rate sensor, the feeding amount sensor, the moisture content sensor, the displacement sensor and the speed sensor to acquire the entrainment loss rate, the breakage rate, the feeding amount, the moisture content, the roller speed and the concave screen gap in real time, thereby providing data support for the multivariable collaborative optimization control.

[0110] The controller is used for running the control operation module of the multivariable collaborative optimization control method, establishing a mixed dynamic model of the threshing process, including a threshing mechanism model and a data-driven model; a system overall state space model, defining state variables and control inputs, constructing a discrete-time state space model; based on the optimization objective function and the constraint condition, solving the finite time domain multivariable collaborative optimization control optimization problem.

[0111] The electro-hydraulic servo system controls the opening of the proportional valve through the PWM signal, adjusts the flow of hydraulic oil to drive the roller motor, and adjusts the roller speed in real time.

[0112] The electric push rod converts rotary motion into linear displacement through a ball screw driven by a stepping motor, drives the concave plate to move, and adjusts the concave screen gap in real time.

[0113] The roller variable speed execution mechanism is used for connecting the mechanical mechanism of the electro-hydraulic servo system and the threshing cylinder.

[0114] The concave screen gap adjustment execution mechanism is used for connecting the mechanical mechanism of the electric push rod and the concave screen gap.

[0115] Data monitoring sensors, various sensors for measuring entrainment loss rate, crushing rate, feeding amount, moisture content, roller speed and concave screen gap.

[0116] Example 2

[0117] A soybean low-crushing threshing multi-variable collaborative optimization control system was installed on a CM100 Valley grain harvester of Weichai-Laver Intelligent Agricultural Technology Co., Ltd. The data acquisition module was GHK01 harvesting parameter online acquisition module designed by Nanjing Agricultural Mechanization Institute of the Ministry of Agriculture and Rural Affairs, the crushing rate sensor was SH-002 harvester crushing and impurity detection sensor designed by Nanjing Agricultural Mechanization Institute of the Ministry of Agriculture and Rural Affairs, the loss rate sensor was SS-002 harvester loss rate detection sensor designed by Nanjing Agricultural Mechanization Institute of the Ministry of Agriculture and Rural Affairs, the speed sensor was M5M8M12M18 Hall proximity switch of Chongqing Shengyi New Electronics Technology Co., Ltd., the displacement sensor was WEP50 pull wire displacement sensor of Shanghai Jiangjingxiang Electronics Co., Ltd. The moisture sensor was Skc01 grain moisture content sensor designed by Nanjing Agricultural Mechanization Institute of the Ministry of Agriculture and Rural Affairs, the feeding amount sensor was WRC01 feeding amount detection module designed by Nanjing Agricultural Mechanization Institute of the Ministry of Agriculture and Rural Affairs, the push rod motor was Jiechuan JC8 push rod of Changzhou Jiechuan Mechanical and Electrical Equipment Co., Ltd., the electro-hydraulic servo system was MS series servo motor of Xiamen Jingyan Automation Component Co., Ltd., and the controller was SPC-FFMC-Y1620 controller of Changsha Shubo Electronics Technology Co., Ltd.

[0118] The data acquisition module of the soybean low-crushing threshing multi-variable collaborative optimization control system collects the data of each sensor through the CAN bus, obtains the information of real-time roller speed, soybean moisture content, crushing rate, loss rate, concave screen gap, feeding amount and over-bridge chain speed at the current time, and delivers all the data to the controller.

[0119] The controller loads the threshing mechanism model obtained by fitting the experimental data, and the threshing dynamics model is:

[0120]

[0121]

[0122] g(H(t))=-0.1714·H(t) 2 +5.9086·H(t)-49.277

[0123] The delivery-feeding dynamic model is:

[0124]

[0125] Constrain φ(t) in the safe interval [70%, 80%] to prevent crushing breakage.

[0126] Train the LSTM network with F(t), ω(t)d(t), H(t), η(t), b(t) sequences of the past 10 seconds to predict the breakage rate trend b(t+1:t+5) for the next 5 seconds as a reference for the multivariable collaborative optimization control optimization. pred

[0127] Define state variables and control inputs, construct a discrete-time state-space model:

[0128]

[0129]

[0130] State variables are x(t):

[0131]

[0132] Control inputs are u(t):

[0133] u(t) = [ω(t), d(t), v(t)] T

[0134] Output variables are y(t):

[0135] y(t) = [ω(t), d(t)] T

[0136] Process noise and measurement noise are

[0137] Define the soybean low-breakage threshing multivariable collaborative optimization control optimization problem, at each time t, the controller solves the following finite-time optimization problem:

[0138]

[0139] s.t. x(t+k+1|t) = Ax(t+k|t) + Bu(t+k|t), k = 0, …, N p -1

[0140] y min ≤ y(t+k|t) ≤ y max

[0141] u min ≤ u(t+k|t) ≤ u max

[0142] N p = 10, corresponding to 1 second in the future, assuming the sampling time Δt = 0.1 s, N c ​= 5, b ref ≤ 4.0%, η ref ≥ 95%, Q1 = 0.6, Q2 = 0.4.

[0143] The optimal control sequence u is calculated in real time by the quadratic programming solver OSQP * (t|t),…,u * (t+N c -1|t) is executed only the first control action u * (t|t) is discarded, the state x(t+1) is measured again at the next time instant t+1 and the optimization process is repeated, implementing a receding horizon update.

[0144] The constraints are handled, 450 r / min ≤ ω(t) ≤ 550 r / min, 15 mm ≤ d(t) ≤ 25 mm, 1.0 m / s ≤ v(t) ≤ 1.5 m / s

[0145] The soybean threshing multivariable collaborative optimization control system calculates the target drum speed ω * (t+k|t) and the target concave clearance d * (t+k|t) and implements high-precision and low-delay control through the actuator. The ω * (t+k|t) is converted into the PWM duty cycle command I of the electro-hydraulic servo motor * (t+k|t), the d * (t+k|t) is converted into the displacement command S of the electric push rod * (t+k|t).

[0146] I(t+k|t) = 0.0037ω(t+k|t) - 1.1089

[0147] S(t+k|t) = 2.3691d(t+k|t) - 16.464

[0148] The drum speed is the priority adjustment variable, and the concave clearance is the auxiliary adjustment variable; when the optimization result requires to reduce the speed and increase the clearance at the same time, the speed adjustment is executed first, and the clearance adjustment is started after a delay Δt = 0.2 s. Cross-coupling compensation, a speed-clearance coupling model is established

[0149] Δd comp = k p ·Δω

[0150] k p = 0.05, the clearance change amount is pre-compensated when adjusting the speed.

[0151] As Figure 3As shown, through the above operation, if the soybean harvester is operating at the current time with the roller speed of 500 r / min, the concave clearance of 10 mm, and the conveying speed of 1.5 m / s; the feeding amount is detected to suddenly increase by 20% during operation, the soybean low-breaking threshing multi-variable collaborative optimization control system calculates a new control sequence within 0.1 second, the roller speed is reduced to 480 r / min, the concave clearance is increased to 10.5 mm, and the conveying speed is increased to 1.7 m / s; the new control sequence is converted into a control instruction, the duty ratio is 60%, and the push rod motor is pushed out by 25 mm, so that the soybean threshing and breaking rate is reduced from 4.5% to 2.8%, and the complete threshing rate is maintained at 96%.

[0152] The various embodiments in the specification are described in a progressive manner, and each embodiment focuses on the difference from other embodiments. The same or similar parts between the various embodiments can be referred to each other. For the device disclosed by the embodiments, since it corresponds to the method disclosed by the embodiments, the description is relatively simple, and the related parts can be referred to the method part.

[0153] The above description of the disclosed embodiments enables a person skilled in the art to implement or use the present application. Various modifications to the embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A soybean low breakage threshing multivariable coordinated optimization control method, characterized in that, The method comprises the following steps: Obtaining grain harvesting and threshing related data at the current time; Based on the soybean low breakage threshing multivariable collaborative control and the target cylinder speed and the concave opening degree are solved; According to the solving result, the control instruction is output, and the online regulation and control of the cylinder speed and the concave opening degree are realized.

2. The soybean low breakage threshing multivariable synergistic optimization control method according to claim 1, characterized in that, The grain harvesting and threshing related data includes crop feeding amount F(t), cylinder speed ω(t), concave gap d(t), crop moisture content H(t) and entrainment loss rate η(t), and breakage rate b(t).

3. The soybean low breakage threshing multivariable synergistic optimization control method according to claim 1, characterized in that, The soybean low breakage threshing multivariable collaborative control specifically comprises the following steps: According to the threshing dynamics model and the conveying-feeding dynamic model, a threshing mechanism model is constructed, which is used to construct a constraint condition mode extrusion breakage; According to the historical grain harvesting and threshing related data, a data-driven model is constructed by training an LSTM network. First, the abnormal values are removed, the multi-sensor data are synchronized according to the time stamp, and the working conditions are labeled according to the crop type, cylinder speed, concave gap, crop moisture content and feeding amount level. Then, the input / output variables are standardized by Z-Score. Then, 30 consecutive time steps are taken as the input window, and the breakage rate and entrainment loss rate of the next time step are taken as the output label to generate a sample sequence in the supervised learning format. The parameters such as the number of LSTM layers, the number of neurons, the Dropout ratio and the time window length are adjusted through Bayesian optimization. Finally, the model is trained, and the trained model is used to construct time series data for multivariable collaborative optimization control. Define state variables and control inputs, and construct a discrete-time state space model. Define the soybean low breakage threshing multivariable collaborative optimization control optimization problem, and solve the finite time domain optimization problem at each time t.

4. The soybean low breakage threshing multivariable synergistic optimization control method according to claim 3, characterized in that, The threshing dynamics model is used to describe the relationship between the crop feeding amount F(t), the cylinder speed ω(t), the concave gap d(t), the crop moisture content H(t) and the entrainment loss rate η(t), and the breakage rate b(t): Wherein, k1, g1, d1, k2, g2, d2 are the correction factors of the model, the conveying-feeding dynamic model is used to describe the relationship between the conveying chain speed v(t), the volume V of the threshing cylinder of the harvester and the density of the crop to be harvested ρ, and the relationship model of the system correction factor τ and the crop filling rate φ(t) is as follows: Restrict φ(t) within a safety threshold to prevent extrusion breakage.

5. The soybean low breakage threshing multivariable collaborative optimization control method according to claim 1, wherein Define state variables and control inputs, and construct a discrete-time state space model: x(t+1)=Ax(t)+Bu(t)+θ(t) The state variable is x(t): The control input is u(t): u(t) = [ω(t), d(t), v(t)] T The output variable is y(t): y(t) = [ω(t), d(t)] T θ(t), is the process noise and the measurement noise.

6. The soybean low breakage threshing multi-variable co-optimization control method according to claim 1, characterized in that, Define the soybean low breakage threshing multivariable collaborative optimization control optimization problem, and solve the finite time domain optimization problem at each time t, and the specific calculation process is as follows: s.t. x(t+k+1|t) = Ax(t+k|t) + Bu(t+k|t), k = 0,..., N p -1 y min ≤ y(t+k|t) ≤ y max u min ≤ u(t+k | t) ≤ u max N p N c b ref ,η ref are the reference target values of the breakage rate and the entrainment loss rate, respectively, the weight coefficients Q1, Q2 adjust the priority of the breakage rate and the entrainment loss rate, R d is the input variation rate penalty term. Solve optimization problem in real-time, use quadratic programming solver OSQP to compute optimal control sequence u in real-time * (t|t),…,u * (t+N c -1|t), only the first control action u*(t|t) is executed, the rest of the sequence is discarded, the state x(t+1) is measured again at the next time t+1 and the optimization process is repeated, realizing the rolling horizon update; Constraint processing, hard constraints are directly embedded in the optimization problem, including cylinder speed range, concave gap value range and conveying speed limit; soft constraints, for constraints that are difficult to strictly satisfy, introduce a penalty term through a relaxation variable: ω min ≤ ω (t) ≤ ω max ,d min ≤ d (t) ≤ d max ,v min ≤ v (t) ≤ v max ; where ω min is the minimum value of the drum rotational speed, ω max is the maximum value of the drum rotational speed, d min is the minimum value of the recess plate gap value, d max is the maximum value of the recess plate gap value, v min is the minimum value of the conveying speed, v max is the maximum value of the conveying speed.

7. The soybean low breakage threshing multi-variable co-optimization control method according to claim 1, characterized in that, According to the solving result output control instruction, the online regulation and control of the drum rotating speed and the concave screen opening degree are realized, and the specific steps include the following steps. The soybean low-breaking threshing multivariable collaborative optimization control system solves the target roller rotating speed ω at future time through rolling optimization * (t+k|t) and target concave plate gap d * (t+k|t) is realized through the actuator; ω * (t+k|t) is converted into the PWM duty ratio instruction I of the electro-hydraulic servo motor * (t+k|t), d * (t+k|t) is converted into the displacement instruction S of the electric push rod * (t+k|t): I(t+k|t)=k3ω(t+k|t)+d3 S(t+k|t)=k4d(t+k|t)+d4 The collaborative control is realized through the following methods: priority allocation, taking the drum rotating speed as the priority adjustment variable and the concave screen gap as the auxiliary adjustment variable; when the optimization result requires reducing the rotating speed and increasing the gap at the same time, the rotating speed adjustment is performed first, and the gap adjustment is started after a delay of Δt=0.2s; cross-coupling compensation, a rotating speed-gap coupling model is established: Δd comp = k p · Δω k p is a proportionality coefficient, and is a pre-compensation amount of the gap variation when the rotation speed is adjusted.

8. A soybean low-breakage threshing multivariable collaborative optimization control system using the soybean low-breakage threshing multivariable collaborative optimization control method according to any one of claims 1 to 7, characterized by, It includes: A data acquisition module for a grain harvesting and threshing system multi-variable collaborative optimization control system, which acquires real-time data of the grain harvesting and threshing system through data communication with the entrainment loss rate sensor, the breakage rate sensor, the feed rate sensor, the moisture content sensor, the displacement sensor and the rotating speed sensor, and acquires the entrainment loss rate, the breakage rate, the feed rate, the moisture content, the drum rotating speed and the concave screen gap in real time; A controller for running the control operation module of the multi-variable collaborative optimization control method, establishing a mixed dynamic model of the threshing process, including a threshing mechanism model and a data-driven model; a system overall state space model, defining state variables and control inputs, and constructing a discrete time state space model; based on the optimization objective function and the constraint condition, a rolling solution of the finite time domain multi-variable collaborative optimization control optimization problem is performed; An electro-hydraulic servo system for controlling the opening degree of the proportional valve through the PWM signal, adjusting the hydraulic oil flow to drive the drum motor, and adjusting the drum rotating speed in real time; An electric push rod for driving the ball screw through the stepping motor, converting the rotary motion into linear displacement, driving the concave plate to move, and adjusting the concave screen gap in real time; A drum variable speed actuator for connecting the electro-hydraulic servo system and the mechanical mechanism of the threshing drum; A concave screen gap adjustment actuator for connecting the electric push rod and the mechanical mechanism of the concave screen gap; Data monitoring sensors for measuring various types of sensors for measuring the entrainment loss rate, the breakage rate, the feed rate, the moisture content, the drum rotating speed and the concave screen gap.

Images