A method and device for monitoring the operating efficiency of a self-propelled silage harvester

Abstract

A method and device for monitoring the operating efficiency of a self-propelled silage harvester are disclosed. The method includes: real-time data collection from the field; continuous collection of field operation data via operating condition sensors and data acquisition devices; and real-time transmission of the monitoring data to an on-board processing terminal via an on-board communication device. The on-board processing terminal obtains the power consumption of the operating components based on the received monitoring data and filters the operating condition data. A data time lag analysis and correction model is established using a material flow sequence model of the header, feed roller, chopping roller, grain crushing roller, throwing fan, and throwing cylinder. Data anomalies are detected; missing data values ​​are filled; the correlation between various variables and operating efficiency is detected, and a univariate linear model is constructed; an operating efficiency model is constructed, establishing a data model of operating efficiency and the power of each component; the main factors affecting operating efficiency are identified, and a multiple linear regression operating efficiency detection model is established. This invention also discloses a device for monitoring the operating efficiency of a self-propelled silage harvester.

Description

Technical Field

[0001] This invention relates to agricultural machinery field operation condition monitoring and online evaluation technology, and in particular to a method and device for monitoring the operating efficiency of a self-propelled silage harvester based on data screening. Background Technology

[0002] Silage is one of the most important forage crops for ruminant farming, and timely and efficient harvesting of silage crops is a key factor in producing high-quality silage. To maintain optimal moisture and nutrient content in silage and achieve high-quality fermentation later, the optimal harvest period for silage crops is generally required to be within one week. For silage harvesting operations with extremely tight time constraints, the harvesting efficiency of self-propelled silage harvesters is an important indicator for evaluating their performance.

[0003] The ideal field operation state for a self-propelled silage harvester is to maintain optimal operating efficiency within its rated range while ensuring operational reliability and chopping effect. If the field operation efficiency is too low, it will not only directly affect the timely harvesting of crops but may also affect the overall chopping effect, leading to the average length or uniformity of the silage shreds not meeting expectations. Conversely, if the field operation efficiency of the self-propelled silage harvester is too high, the crop stalks will be fed in too quickly, easily causing blockages in the machine's header, feeding section, and conveying section. Subsequent manual clearing of these blockages will further delay the overall harvesting progress. Furthermore, prolonged operation under overload conditions can easily cause fatigue damage to the working parts and may shorten the engine's lifespan. Therefore, to achieve effective evaluation and intelligent harvesting control of the overall performance of the silage harvester, real-time monitoring of the harvesting process's efficiency is necessary.

[0004] The verification of the operational efficiency of existing silage harvesters still mainly relies on manual verification through multi-season field trials of prototypes, which is cumbersome and inefficient. The operational efficiency of self-propelled silage harvesters is often monitored by tracking single operating conditions (such as speed, displacement, torque, and capacitance), achieving online efficiency monitoring through indirect modeling of single-factor, full-process data. However, self-propelled silage harvesters use a full-feed harvesting method, feeding the entire crop plant. This makes the machine-material interaction mechanism more complex, prone to large detection errors and poor stability, making them difficult to apply to actual field testing. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to address the above-mentioned deficiencies of the prior art by providing a data-based method and device for monitoring the operating efficiency of a self-propelled silage harvester, so as to improve the stability and model detection accuracy of field operating efficiency monitoring of the self-propelled silage harvester, and make it applicable to the intelligent field evaluation of the operating efficiency performance of the self-propelled silage harvester.

[0006] To achieve the above objectives, the present invention provides a method for monitoring the operating efficiency of a self-propelled silage harvester, comprising the following steps:

[0007] S100: Real-time field data collection. During the field operation of the self-propelled silage harvester, the field operation data is continuously collected through the working condition sensor and data acquisition device, and the monitoring data is transmitted to the vehicle processing terminal in real time through the vehicle communication device.

[0008] S200. Obtain the power of the working component. The vehicle-mounted processing terminal obtains the power consumption of the working component based on the received monitoring data.

[0009] S300. Screening working condition data: The vehicle-mounted processing terminal uses the Mann-Kendall screening method to identify the rising and falling abrupt change points in the harvest section data area, and uses the monitoring data of the chopping roller as the basis for classification and screening.

[0010] S400, the vehicle-mounted processing terminal establishes a data time delay analysis and correction model based on the material flow sequence conveying model of the header, feeding roller, chopping roller, grain crushing roller, throwing fan and throwing cylinder in the self-propelled silage harvester;

[0011] S500: Detect outliers in the data by using the confidence level Grubbs criterion, and determine outliers by examining and removing data one by one.

[0012] S600. Fill missing values ​​in the data. Use cubic spline interpolation to fill missing values. Interpolate the removed outliers with the missing values ​​to increase the sample data.

[0013] S700. Analyze the correlation between main factors, examine the correlation between each variable and work efficiency, apply the Pearson correlation coefficient to calculate the correlation between variables and obtain the correlation coefficient, construct a univariate linear model to obtain the influence relationship of each work component on work efficiency; and

[0014] S800. Construct an operational efficiency model, establish a data model of operational efficiency and the power of each component, and conduct single-factor analysis and coefficient of determination analysis. Based on the above single-factor regression analysis and correlation factor analysis, determine the main factors affecting operational efficiency and establish a multiple linear regression operational efficiency detection model.

[0015] The above-mentioned method for monitoring the operating efficiency of a self-propelled silage harvester includes the following operating condition sensors: header speed sensor, header torque sensor, shredder roller speed sensor, shredder roller torque sensor, throwing fan speed sensor, throwing fan torque sensor, feeder hydraulic pump flow sensor, feeder hydraulic outlet pressure sensor, and feeder hydraulic inlet pressure sensor.

[0016] The above-mentioned method for monitoring the operating efficiency of a self-propelled silage harvester includes step S200, which involves power conversion of mechanical components and hydraulic components. The power consumption of the rotating mechanical components is as follows:

[0017]

[0018] Among them, P r For the power consumed by mechanical parts, n r T is the rotational speed of the mechanical component. r Let R be the rotational torque of the mechanical component, R be the rotational radius of the mechanical component, F be the rotational force on the mechanical component, and v be the rotational torque of the mechanical component. r The linear velocity of the mechanical component;

[0019] The power consumption of the hydraulic drive components is:

[0020]

[0021] Among them, P h p0 is the power consumed by the hydraulic components, p0 is the inlet pressure of the hydraulic pump, ... power consumed by the hydraulic components, p0 is the power consumed by the hydraulic pump, p0< / i η is the hydraulic pump outlet pressure, q is the hydraulic pump output flow rate, n0 is the engine real-time speed, k is the hydraulic pump transmission ratio, V0 is the hydraulic pump rated flow rate, and η is the hydraulic pump rated flow rate. v This refers to the volumetric efficiency of the hydraulic pump.

[0022] The above-mentioned method for monitoring the operating efficiency of a self-propelled silage harvester, wherein step S300, using the operating data of the shredding roller as a classification criterion, further includes classifying and filtering the monitoring data, and includes:

[0023] S301. Define the set of task data point categories D = {d1, d2, d3, d4} = {1, 2, 3, 0};

[0024] S302. Based on the shredding roller speed, preliminary classification is performed to obtain the test section and the non-test data section:

[0025]

[0026] Among them, D i For the i-th sampled data, n c n is the real-time crushing speed of the shredding roller. ce The rated speed of the shredding roller;

[0027] S303. Use the Mann-Kendall nonparametric test to detect data mutation points and extract valid work segment data;

[0028] The test data segments in the original torque detection sequence are decomposed and used independently to detect the load loading change point and the load disappearance change point; based on the data sequence X = {x1, x2, x3, ..., x...} of n torque detection values... n The sequence is forward traversal to calculate the cumulative number of times the torque value at time point i is greater than the torque value at time point j, thus constructing a forward-order sequence S. k :

[0029]

[0030]

[0031]

[0032] Calculate the positive sequence statistic UF k :

[0033]

[0034] Assuming that the detection values ​​of sequence X are random, independent, and follow the same distribution, then sequence S k The mean E(S) k ) and variance var(S k )for:

[0035]

[0036]

[0037] S304, Based on the reverse detection sequence XB={x n ,x n-1 ,…,x1}, repeat the above process in reverse order to construct the reversed order sequence S. Kb Calculate the positive sequence statistic UB k After taking the negative value, the inversion statistic sequence UB is obtained. k ;

[0038] S305. Based on the law of large numbers, taking a significance level of 0.05, within the confidence interval U... 0.05 Solving the following equation under constraints yields an effective solution that identifies the points of abrupt increase and decrease in the torque data of the working component:

[0039]

[0040] In the above-mentioned method for monitoring the operating efficiency of a self-propelled silage harvester, in step S400, the effective data ranges for the header, feed section, shredding roller, and throwing fan are respectively X g X w X q X p The corresponding data time delay analysis correction models are as follows:

[0041] X g =[x(t) cs -Δt1-Δt2),x(t cs -Δt1-Δt2+T),...,x(t cs -Δt1-Δt2+iT),...,x(t ce -Δt1-Δt2)];

[0042] X w =[x(t) cs -Δt2),x(t cs -Δt²+T),...,x(t) cs -Δt2+iT),...,x(t ce -Δt2)];

[0043] X q =[x(t) cs ),x(t cs +T),...,x(t cs +iT),...,x(t ce )];

[0044] X p =[x(t) bs ),x(t bs +T),...,x(t bs +iT),...,x(t be )];

[0045] Among them, t cs , t ce The points where the torque of the shredder roller suddenly increases and decreases; t bs , t be Δt1 represents the points where the torque of the throwing blower increases and decreases; T represents the data collection interval of the sample; Δt1 represents the time lag between the increase in the load of the feeding section and the increase in the crop feed; Δt2 represents the time lag between the increase in the load of the chopping roller and the increase in the load of the feeding section; Δt3 represents the time lag between the increase in the load of the throwing blower and the increase in the load of the chopping roller.

[0046] The above-mentioned method for monitoring the operating efficiency of a self-propelled silage harvester, wherein step S500 further includes:

[0047] S501. Singular data points include outliers and missing values. Sample data X = [x(t)] collected at a fixed frequency T. cs +T),…,x(t cs +iT),…,x(t cs +pT)], according to Bessel's formula, calculate the sample mean and standard deviation statistics:

[0048]

[0049] in, Let σ be the mean of the data sample, and σ be the standard deviation of the data sample.

[0050] S502. Sort the samples by value and reorganize them into {x(t)1, x(t)2, ..., x(t)p}, where x(t)1 ≤ x(t)2 ≤ ... ≤ x(t)p. Determine the residual v of the suspected singularities according to the following formula. i :

[0051]

[0052] S503, regarding the suspicious residual v i Compared with the critical value G0 of the Grubbs coefficient, if the value of the suspicious residual satisfies the following formula, then the point is determined to be an outlier, removed from the sequence, and re-examined:

[0053]

[0054] Where G0 can be obtained from a table, and α is the significance level;

[0055] S504. Repeat steps S501-S503 until there are no singularities in the data sequence.

[0056] In the above-mentioned method for monitoring the operating efficiency of a self-propelled silage harvester, in step S600, cubic spline interpolation is used to fill in missing values, and the function S(x) satisfies the following condition:

[0057] n points (x) on the plane i ,y i (i = 1, 2, ..., n), where (x1 < x2 < ... < xn) n ), are called sample points, and the function passes through these sample points: S(x) i )=y i (i = 1, 2, ..., n);

[0058] S(x) in each subinterval [x i ,x i+1 The above is a cubic polynomial:

[0059] S(x)=c1(xx i) 3 +c2(xx i ) 2 +c3(xx i )+c4;

[0060] S(x) is valid over the entire interval [x1, x2]. n It has continuous first and second derivatives.

[0061] The above-mentioned method for monitoring the operating efficiency of a self-propelled silage harvester, wherein step S700 further includes:

[0062] S701, Variables include the operating power P of the cutting table. g The operating power P of chopping arsenic q The operating power P of the throwing blower p The operating power P of the hydraulic pump in the feed section w The correlation coefficients between the five variables, along with the work efficiency TH, are calculated using the following formula:

[0063]

[0064] S702. Based on the real-time power consumption of each key operating component during the harvesting process of the self-propelled silage harvester, a univariate linear regression model between single power data and operating efficiency is established as follows:

[0065]

[0066] In the above-mentioned method for monitoring the operating efficiency of self-propelled silage harvesters, step S800 identifies the top three factors affecting operating efficiency and establishes a multiple linear regression model for detecting operating efficiency as follows:

[0067]

[0068] To better achieve the above objectives, the present invention also provides a self-propelled silage harvester operation efficiency monitoring device, wherein the self-propelled silage harvester operation efficiency monitoring method described above is used.

[0069] The technical effects of this invention are as follows:

[0070] This invention achieves real-time acquisition of operating data such as machine speed, machine torque, hydraulic flow, hydraulic pressure, and operating power throughout the entire operation process of a self-propelled silage harvester, including the cutting, feeding, chopping, and throwing sections, by installing a non-destructive power monitoring sensor. A data preprocessing algorithm combining Mann-Kendall boundary detection and time lag correction is applied to effectively filter the operating data. Outliers and missing values ​​are corrected and supplemented using Globus detection and cubic spline interpolation. Single-factor and multi-factor least-squares operating efficiency measurement models are constructed based on Pearson correlation analysis. By combining operating data from multiple components for fusion regression, accurate measurement of the operating efficiency of the self-propelled silage harvester is achieved. This provides an effective technical solution for rapidly and accurately measuring the operating efficiency of self-propelled silage harvesters and for intelligent field assessment in complex operating environments.

[0071] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments, but this is not intended to limit the present invention. Attached Figure Description

[0072] Figure 1 This is a schematic diagram illustrating the working principle of an embodiment of the present invention;

[0073] Figure 2 This is a schematic diagram showing the installation position of the working condition sensor on a self-propelled silage harvester according to an embodiment of the present invention;

[0074] Figure 3A , 3B This is a schematic diagram of the measurement principle of a resistance strain gauge torque sensor according to an embodiment of the present invention;

[0075] Figure 4 This is a schematic diagram of the measurement principle of a photoelectric speed sensor according to an embodiment of the present invention;

[0076] Figure 5 This is a schematic diagram of an ultrasonic hydraulic flow sensor according to an embodiment of the present invention;

[0077] Figure 6 This is a schematic diagram of the installation position of a hydraulic pressure sensor according to an embodiment of the present invention;

[0078] Figure 7 This is a schematic diagram of a harvester logistics conveying model according to an embodiment of the present invention.

[0079] Among them, the attached reference numerals

[0080] 1. Cutting table speed sensor

[0081] 2. Throwing fan speed sensor

[0082] 3. Chopping roller speed sensor

[0083] 4. Hydraulic pump flow sensor

[0084] 5. Outlet pressure sensor

[0085] 6. Inlet pressure sensor

[0086] 7. Hydraulic pump

[0087] 8. Resistance strain gauge torque sensor

[0088] 9. Photoelectric speed sensor

[0089] 91 Photoelectric Speed ​​Measuring Disc

[0090] 92 working axes

[0091] 93 Photoelectric sensor probe

[0092] 10 Stainless Steel Straight Pipe

[0093] 11. Ultrasonic hydraulic flow sensor

[0094] 12 Flexible shafts Detailed Implementation

[0095] The structural and working principles of the present invention will be described in detail below with reference to the accompanying drawings:

[0096] See Figure 1 and Figure 2 The self-propelled silage harvester operation efficiency monitoring device of the present invention monitors the operation efficiency of the silage harvester based on data screening, and includes: operating condition sensors, a data acquisition device, a vehicle-mounted communication device, and a vehicle-mounted processing terminal. The data acquisition device is connected to the operating condition sensors and the vehicle-mounted communication device respectively, and is used to collect and forward operating condition data of key components in the field. The vehicle-mounted communication device is installed on the self-propelled silage harvester and is used to construct a whole-vehicle bus-type data communication network to realize the interconnection of data of the entire self-propelled silage harvester. The vehicle-mounted processing terminal is installed in the cab of the self-propelled silage harvester and is connected to the vehicle-mounted communication device, and is used to receive and process field operating condition data and display the operation efficiency detection results. The operating condition sensors include a header torque sensor, a header speed sensor 1, a shredder roller speed sensor 3, a shredder roller torque sensor, a blower speed sensor 2, a blower roller torque sensor, a feed section hydraulic pump flow sensor 4, a feed section hydraulic outlet pressure sensor 5, and a feed section hydraulic inlet pressure sensor 6, etc. The self-propelled forage harvester operation efficiency monitoring device monitors the forage harvester operation efficiency through the following method.

[0097] The method for monitoring the operating efficiency of a self-propelled silage harvester of the present invention includes the following steps:

[0098] Step S100: Real-time collection of field data. During the field operation of the self-propelled silage harvester, field operation data is continuously collected through working condition sensors and data acquisition devices, and the monitoring data is transmitted to the vehicle processing terminal in real time through the vehicle communication device.

[0099] Step S200: Obtain the power of the working component. The vehicle-mounted processing terminal obtains the power consumption of the working component based on the received monitoring data.

[0100] Step S300: Screening working condition data. The vehicle-mounted processing terminal uses the Mann-Kendall screening method to identify the rising and falling abrupt change points in the harvest section data area. By extracting the monitoring data of the chopping roller as the classification basis, the monitoring data is classified and screened.

[0101] Step S400: The vehicle-mounted processing terminal establishes a data time delay analysis and correction model based on the material flow sequence model of the header, feeding roller, chopping roller, grain crushing roller, throwing fan and throwing cylinder in the self-propelled silage harvester.

[0102] Step S500: Detect outliers in the data. Use the confidence level Grubbs criterion to determine outliers by checking and removing data one by one.

[0103] Step S600: Fill in missing data values. Use cubic spline interpolation to fill in missing values. Interpolate the removed outliers with the missing values ​​to increase the sample data.

[0104] Step S700: Analyze the correlation of main factors, detect the correlation between each variable and work efficiency, apply the Pearson correlation coefficient to calculate the correlation between variables and obtain the correlation coefficient, construct a univariate linear model, and obtain the influence relationship of each work component on work efficiency; and

[0105] Step S800: Construct an operational efficiency model, establish a data model of operational efficiency and the power of each component, and conduct single-factor analysis and coefficient of determination analysis; based on the above single-factor regression analysis and correlation factor analysis, determine the main factors affecting operational efficiency, and establish a multiple linear regression operational efficiency detection model.

[0106] Step S200 includes power conversion for mechanical components and power conversion for hydraulic components. The power consumption of the rotating mechanical components is as follows:

[0107]

[0108] Among them, P r For the power consumed by mechanical parts, n r T is the rotational speed of the mechanical component. rLet R be the rotational torque of the mechanical component, R be the rotational radius of the mechanical component, F be the rotational force on the mechanical component, and v be the rotational torque of the mechanical component. r The linear velocity of the mechanical component;

[0109] The power consumption of the hydraulic drive components is:

[0110]

[0111] Among them, P h p0 represents the power consumption of the hydraulic components, p0 represents the inlet pressure of the hydraulic pump 7, and p i η is the outlet pressure of hydraulic pump 7, q is the hydraulic flow rate output by hydraulic pump 7, n0 is the real-time engine speed, k is the transmission ratio of hydraulic pump 7, V0 is the rated flow rate of hydraulic pump 7, and η is the output pressure of hydraulic pump 7. v The volumetric efficiency of hydraulic pump 7.

[0112] In step S300, the monitoring data is classified and filtered based on the operating data of the shredding roller, further including:

[0113] Step S301: Define the set of job data point categories D = {d1, d2, d3, d4} = {1, 2, 3, 0};

[0114] Step S302: Based on the shredding roller speed, perform preliminary classification to obtain the test section and the non-test data section:

[0115]

[0116] Among them, D i For the i-th sampled data, n c n is the real-time crushing speed of the shredding roller. ce The rated speed of the shredding roller;

[0117] Step S303: Use the Mann-Kendall nonparametric test to detect data mutation points and extract valid work segment data;

[0118] The test data segments in the original torque detection sequence are decomposed and used independently to detect the load loading change point and the load disappearance change point; based on the data sequence X = {x1, x2, x3, ..., x...} of n torque detection values... n The sequence is forward traversal to calculate the cumulative number of times the torque value at time point i is greater than the torque value at time point j, thus constructing a forward-order sequence S. k :

[0119]

[0120]

[0121]

[0122] Calculate the positive sequence statistic UF k :

[0123]

[0124] Assuming that the detection values ​​of sequence X are random, independent, and follow the same distribution, then sequence S k The mean E(S) k ) and variance var(S k )for:

[0125]

[0126]

[0127] Step S304: Based on the reverse detection sequence XB = {x n ,x n-1 ,…,x1}, repeat the above process in reverse order to construct the reversed order sequence S. Kb Calculate the positive sequence statistic UB k After taking the negative value, the inversion statistic sequence UB is obtained. k ;

[0128] Step S305: According to the law of large numbers, take a significance level of 0.05, and within the confidence interval U... 0.05 Solving the following equation under constraints yields an effective solution that identifies the points of abrupt increase and decrease in the torque data of the working component:

[0129]

[0130] In step S400, the effective data ranges for the cutting table, feeding section, shredding roller, and throwing fan are respectively X g X w X q X p The corresponding data time delay analysis correction models are as follows:

[0131] X g =[x(t) cs -Δt1-Δt2),x(t cs -Δt1-Δt2+T),...,x(t cs -Δt1-Δt2+iT),...,x(t ce -Δt1-Δt2)];

[0132] X w =[x(t) cs -Δt2),x(t cs -Δt²+T),...,x(t)cs -Δt2+iT),...,x(t ce -Δt2)];

[0133] X q =[x(t) cs ),x(t cs +T),...,x(t cs +iT),...,x(t ce )];

[0134] X p =[x(t) bs ),x(t bs +T),...,x(t bs +iT),...,x(t be )];

[0135] Among them, t cs , t ce The points where the torque of the shredder roller suddenly increases and decreases; t bs , t be Δt1 represents the points where the torque of the throwing blower increases and decreases; T represents the data collection interval of the sample; Δt1 represents the time lag between the increase in the load of the feeding section and the increase in the crop feed; Δt2 represents the time lag between the increase in the load of the chopping roller and the increase in the load of the feeding section; Δt3 represents the time lag between the increase in the load of the throwing blower and the increase in the load of the chopping roller.

[0136] Step S500 further includes:

[0137] Step S501: Singular data points include outliers and missing values. Sample data X = [x(t)] is collected at a fixed frequency T. cs +T),…,x(t cs +iT),…,x(t cs +pT)], according to Bessel's formula, calculate the sample mean and standard deviation statistics:

[0138]

[0139] in, Let σ be the mean of the data sample, and σ be the standard deviation of the data sample.

[0140] Step S502: Sort the samples by value and reorganize them into {x(t)1,x(t)2,…,x(t)p}, where x(t)1≤x(t)2≤…x(t)p. Determine the residual v of the suspected singularities according to the following formula. i :

[0141]

[0142] Step S503: Transfer the suspicious residual v i Compared with the critical value G0 of the Grubbs coefficient, if the value of the suspicious residual satisfies the following formula, then the point is determined to be an outlier, removed from the sequence, and re-examined:

[0143]

[0144] Where G0 can be obtained from a table, α is the significance level; and

[0145] Step S504: Repeat steps S501-S503 until there are no singularities in the data sequence.

[0146] In step S600, cubic spline interpolation is used to fill in missing values, and the function S(x) satisfies the following condition:

[0147] n points (x) on the plane i ,y i (i = 1, 2, ..., n), where (x1 < x2 < ... < xn) n ), are called sample points, and the function passes through these sample points: S(x) i )=y i (i = 1, 2, ..., n);

[0148] S(x) in each subinterval [x i ,x i+1 The above is a cubic polynomial:

[0149] S(x)=c1(xx i ) 3 +c2(xx i ) 2 +c3(xx i )+c4;

[0150] S(x) is valid over the entire interval [x1, x2]. n It has continuous first and second derivatives.

[0151] Step S700 further includes:

[0152] Step S701, variables include the operating power P of the cutting table. g The operating power P of chopping arsenic q The operating power P of the throwing blower p The operating power P of the hydraulic pump 7 in the feed section w The correlation coefficients between the five variables, along with the work efficiency TH, are calculated using the following formula:

[0153] as well as

[0154] Step S702: Using the real-time power consumption of each key operating component during the harvesting process of the self-propelled silage harvester as the model input, establish a univariate linear regression model between single power data and operating efficiency as follows:

[0155]

[0156] In step S800, the top three factors affecting work efficiency are identified, and a multiple linear regression work efficiency detection model is established as follows:

[0157]

[0158] An embodiment of the present invention provides a method for monitoring the operating efficiency of a self-propelled silage harvester, which uses the aforementioned self-propelled silage harvester operating efficiency monitoring device for detection, and includes the following steps:

[0159] Step S100: Real-time field data acquisition. During the field operation of the self-propelled silage harvester, field operation data is continuously collected through the working condition sensors and the data acquisition device, and the monitoring data is transmitted in real time to the vehicle-mounted processing terminal through the vehicle-mounted communication device. The working condition sensors mainly include a header speed sensor 1, a header torque sensor, a shredder roller speed sensor 3, a shredder roller torque sensor, a blower speed sensor 2, a blower torque sensor, a feed hydraulic pump flow sensor 4, a feed hydraulic outlet pressure sensor 5, and a feed hydraulic inlet pressure sensor 6. The installation positions of the above working condition sensors on the self-propelled silage harvester are as follows: Figure 2 As shown, the cutting table, shredding roller and throwing fan are mechanical transmission components, and the hydraulic pump 7 of the feeding section is a hydraulic transmission component; during operation, the above-mentioned working condition sensors monitor the data in real time and upload it to the vehicle terminal for data processing.

[0160] Step S200: Power measurement of key operating components, mainly including power conversion of mechanical components and hydraulic components. The on-board processing terminal converts the received sensor data into the power consumed by each component. The power sources of key operating components include both mechanical transmission and hydraulic transmission. For mechanically transmitted operating components, their real-time operating power can be obtained by measuring the rotational speed and torque of the rotating shaft. The basic measurement principle is as follows: Regarding torque measurement, such as... Figure 3A , 3B As shown, Figure 3A , 3B This is a schematic diagram illustrating the measurement principle of a resistance strain gauge torque sensor 8 according to an embodiment of the present invention. The resistance strain gauge torque sensor 8 is used for torque measurement. The strain gauge of the resistance strain gauge torque sensor 8 is fixed to the rotating elastic shaft 12 being measured using a special bonding process. Its advantages are simple structure, ease of manufacturing, and high accuracy. Regarding rotational speed measurement, such as... Figure 4 As shown, Figure 4 This is a schematic diagram of the measurement principle of the photoelectric speed sensor 9 according to an embodiment of the present invention. When the photoelectric speed measuring disk 91 rotates with the working shaft 92, the photoelectric sensing probe 93 detects and outputs several speed pulse signals, which are used to calculate and obtain the speed information of the working shaft 92.

[0161] According to the definition of torque in mechanics, we have:

[0162] T r =F·R; (1)

[0163] During the rotation of the working component, the conversion relationship between linear velocity and rotational speed is as follows:

[0164]

[0165] Substituting equations (1) and (2) into the equations, the power consumed by the rotating mechanical parts is:

[0166]

[0167] In the formula, P r n represents the power consumed by the mechanical components. r T represents the rotational speed of a mechanical component. r R represents the rotational torque of the mechanical component, R represents the rotational radius of the mechanical component, and F represents the rotational force on the mechanical component.

[0168] For hydraulically driven operating components, their real-time operating power can be obtained by measuring the output differential pressure and output flow rate of the hydraulic pump 7. An ultrasonic hydraulic flow sensor 11 is used, such as... Figure 5 As shown, Figure 5 This is a schematic diagram of an ultrasonic hydraulic flow sensor 11 according to an embodiment of the present invention, which is installed on the stainless steel straight pipe 10 in the middle between the feed hydraulic pump 7 and the hydraulic motor; the hydraulic pressure sensor includes an outlet pressure sensor 5 and an inlet pressure sensor 6, as shown below. Figure 6 As shown, Figure 6 This is a schematic diagram of the installation position of a hydraulic pressure sensor according to an embodiment of the present invention. The outlet pressure sensor 5 and the inlet pressure sensor 6 are respectively installed at the outlet and inlet of the feed hydraulic pump 7.

[0169] The power calculation for hydraulic pump 7 is as follows:

[0170]

[0171] In the above formula, P h This indicates the power consumption of the hydraulic components, Δp represents the pressure difference between the inlet and outlet of hydraulic pump 7, p0 represents the inlet pressure of hydraulic pump 7, and p i q represents the outlet pressure of hydraulic pump 7, and q represents the hydraulic flow rate output by hydraulic pump 7.

[0172] The inlet and outlet differential pressure Δp can be obtained by measuring the inlet pressure sensor 6 and the outlet pressure sensor 5 of the hydraulic pump 7. Since the hydraulic pump 7 is powered by the engine through a pulley, its output flow rate is calculated using the following formula:

[0173] q=n0·k·V0·η v (5)

[0174] In the above formula, q represents the hydraulic flow rate output by hydraulic pump 7, n0 represents the real-time engine speed, k represents the transmission ratio of hydraulic pump 7, V0 represents the rated flow rate of hydraulic pump 7, and η v Indicates the volumetric efficiency of hydraulic pump 7, η v =0.93

[0175] Substituting equations (4) and (5), the power consumption of the hydraulic drive component is:

[0176]

[0177] Therefore, the operating power P of the header can be obtained by substituting the data obtained from the header speed sensor 1 and the header torque sensor into formula (3). g The operating power P of the chopping bed can be obtained by substituting the data obtained from the chopping bed speed sensor and the chopping bed torque sensor into formula (3). q The operating power P of the throwing blower can be obtained by substituting the data obtained from the speed sensor 2 and the torque sensor of the throwing blower into formula (3). p The operating power P of the hydraulic pump 7 can be obtained by substituting the data obtained from the flow sensor 4, hydraulic outlet pressure sensor 5, and hydraulic inlet pressure sensor 6 of the feed section hydraulic pump into formula (6). w .

[0178] Step S300: Main component working condition data screening. The vehicle-mounted processing terminal uses the Mann-Kendall screening method to identify the rising and falling abrupt change points in the harvest section data area. By extracting the monitoring data of the chopping roller as the dividing basis, the measurement data is classified and screened.

[0179] Data acquired during the operation of a self-propelled silage harvester can be categorized according to different operating states: pre-acceleration data point d1 (operating components running unloaded), harvesting data point d2 (operating components running with load), operation stop data point d3 (operating components returning to unloaded state), and shutdown waiting data point d4 (operating components stopping operation). To construct an effective power-operation efficiency model, only data from the silage harvesting segment during the experiment should be selected as the model's original data input. The shredding roller and the blower are key operating components of the silage harvester, accounting for over 60% of the total power consumption. The shredding roller exhibits more significant data variation characteristics under field operating conditions. Therefore, this invention uses the shredding roller's operating data as the basis for classification and screening of the measured data. The main screening process is as follows.

[0180] Define the set of task data point categories D = {d1, d2, d3, d4} = {1, 2, 3, 0};

[0181] Preliminary classification based on the shredding roller speed yields test sections and non-test data sections.

[0182]

[0183] In the formula, D i Let n represent the i-th sampled data. c This indicates the real-time crushing speed n of the shredding roller. ce This indicates the rated speed of the shredding roller.

[0184] The Mann-Kendall nonparametric test was used to detect data mutation points, thereby extracting valid work segment data.

[0185] Mann-Kendall is a nonparametric time series rank test method, which is well-suited for field operation data with missing values ​​and outlier interference, and where the distribution of the tested samples is unknown. However, it is not suitable for detecting multiple abrupt change points. Therefore, the test data segments in the original torque detection sequence are first decomposed and used independently to detect load loading and load disappearance abrupt change points. Secondly, based on the data sequence X = {x1, x2, x3, ..., x...} of n torque detection values... n The sequence is forward traversal to calculate the cumulative number of times the torque value at time i is greater than the torque value at time j, thus constructing a forward-order sequence S. k :

[0186]

[0187]

[0188]

[0189] Calculate the positive sequence statistic UF k :

[0190]

[0191] Assuming that the detection values ​​of sequence X are random, independent, and follow the same distribution, then sequence S k The mean E(S) k ) and variance var(S k )for

[0192]

[0193]

[0194] Based on the reverse detection sequence XB={x n ,x n-1 ,…,x1}, repeat the above process in reverse order to construct the reversed order sequence S. Kb Calculate the positive sequence statistic UB k After taking the negative value, the inversion statistic sequence UB is obtained. k .

[0195] Finally, based on the law of large numbers, the statistic follows a standard normal distribution. Taking a significance level of 0.05, the following formula is solved under the constraint of confidence interval U0.05. If a valid solution exists, the rising and falling abrupt change points of the torque data of the working component are obtained.

[0196]

[0197] Step S400: Screening of working condition data for all components. The vehicle-mounted processing terminal establishes a data time delay analysis and correction model based on the logistics transport model inside the self-propelled silage harvester, which is from "cutting table -> feeding roller -> chopping roller -> grain crushing roller -> throwing fan -> throwing cylinder".

[0198] Based on the material transport model of corn silage in a combine harvester, from "header -> feed roller -> shredder roller -> grain crusher roller -> throwing fan -> throwing cylinder", such as Figure 7 As shown, data time delay analysis is performed.

[0199] When silage is fed into the machine:

[0200] The crop arrives at the header at time th. The time lag caused by the change in header torque due to the change in header load is negligible, so ts = th.

[0201] After the crop is collected by the header, it arrives at the conveyor roller at time tf. The increase in load and torque of the feeding section has a time lag of Δt1 relative to the crop feeding, as shown in Equation 12.

[0202] After the crop passes through the feeding and compaction device, it reaches the chopping drum at time tc. The change in the load torque of the chopping drum has a time lag of (Δt1+Δt2) relative to the crop feeding.

[0203] After the crop is chopped and crushed, it arrives at the feeding fan at time tb. The fan load data monitoring shows a time lag (Δt1+Δt2+Δt3) relative to the crop feeding.

[0204]

[0205]

[0206] In the above formula, S1 and S2 represent the travel distance of the silage crop at the disc header, and n h ,n f Indicates the rotational speed of the rotary cutter head and the speed of the upper feed roller, R h ,R f This indicates the radius of the rotating cutter head and the radius of the upper feed roller.

[0207] Based on the Mann-Kendall nonparametric test method (1)-(3) in step S300, the torque change abrupt change points of the two components with the highest power consumption (shredding roller and throwing fan) are obtained. Assuming that the effective data intervals of the header, feed section, shredding roller and throwing fan are X g X w X q X p .

[0208] X g =[x(t) cs -Δt1-Δt2),x(t cs -Δt1-Δt2+T),...,x(t cs -Δt1-Δt2+iT),...,x(t ce -Δt1-Δt2)]; (14)

[0209] X w =[x(t) cs -Δt2),x(t cs -Δt²+T),...,x(t) cs -Δt2+iT),...,x(t ce -Δt2)]; (15)

[0210] X q =[x(t) cs ),x(t cs +T),...,x(t cs +iT),...,x(t ce(16)

[0211] X p =[x(t) bs ),x(t bs +T),...,x(t bs +iT),...,x(t be (17)

[0212] In the above formula, t cs ,t ce Indicates the points where the torque of the shredder roller suddenly increases and decreases; t bs ,t be Δt1 represents the points where the torque of the throwing blower increases and decreases; T is the data collection interval for the sample; Δt1 is the time lag between the increase in the load of the feeding section and the increase in the crop feed; Δt2 is the time lag between the increase in the load of the chopping roller and the increase in the load of the feeding section; Δt3 is the time lag between the increase in the load of the throwing blower and the increase in the load of the chopping roller.

[0213] Step S500: Data outlier detection. Outlier judgment is performed based on the principle of statistical distribution. The confidence level Grubbs criterion is adopted, and the outlier data is determined through "one-by-one test and data removal".

[0214] The actual field power data collected contains certain outlier data points. Outlier data points include both outliers and missing values. Outlier identification is generally based on statistical distribution principles, employing the Grubbs criterion and using a process of "testing and removing outliers one by one" to determine outliers. The main principles and steps are as follows:

[0215] Step S501, namely, collecting sample data X = [x(t) at a fixed frequency T] cs +T),…,x(t cs +iT),…,x(t cs +pT)], calculate the sample mean and standard deviation statistics according to Bessel's formula.

[0216]

[0217] In the above formula σ represents the mean of the data sample, and σ represents the standard deviation of the data sample.

[0218] Step S502: Sort the samples by value and reorganize them into {x(t)1,x(t)2,…,x(t)p}, where x(t)1≤x(t)2≤…x(t)p. The left and right edge values ​​are most likely to contain errors. Determine the residual v of the suspected singularity according to formula (19). i .

[0219]

[0220] Step S503: Transfer the suspicious residual v i Compare G0 with the critical value of the Grubbs coefficient. If the value of the suspicious residual satisfies the formula (20), then the point is determined to be an outlier. After removing it from the sequence, the test is repeated. Repeat the above steps until there are no outliers in the data sequence.

[0221]

[0222] In the above formula, G0 can be obtained from a table, and α is the significance level, which is generally taken as 0.05 or 0.01.

[0223] The data collected in the experiment can be processed through the above steps to remove outlier data.

[0224] Step S600: Data missing value filling. Missing values ​​are filled using cubic spline interpolation. The removed outliers are interpolated with the missing values ​​to increase the sample data.

[0225] During the experiment, the collected samples will contain a certain number of outliers and missing values, resulting in data loss. To ensure sufficient sample data, interpolation is used to complete the data by removing outliers and missing values. To ensure the smoothness of the obtained data, cubic spline interpolation is used. Cubic splines have continuous second derivatives, resulting in smoother curves. The basic principle is as follows:

[0226] Given n points (x) on a plane i ,y i (i = 1, 2, ..., n), where (x1 < x2 < ... < xn) n These points are called sample points. If a function S(x) satisfies the following three conditions, then S(x) is called a cubic spline function passing through these n points.

[0227] (1)S(x i )=y i (i = 1, 2, ..., n), meaning the function passes through these sample points;

[0228] (2) S(x) in each subinterval [x i ,x i+1 The above is a cubic polynomial:

[0229] S(x)=c1(xx i ) 3 +c2(xx i ) 2 +c3(xx i )+c4;

[0230] (3) S(x) is valid over the entire interval [x1, x2, x3, x4, x5, x6, x7, x7, x8, x9 ...n It has continuous first and second derivatives.

[0231] Step S700: Principal factor correlation analysis, detect the correlation between each variable and work efficiency, apply Pearson correlation coefficient to calculate the correlation between variables and obtain the correlation coefficient, construct a univariate linear model, and obtain the influence relationship of each work component on work efficiency.

[0232] To detect the correlation between the variables for further data analysis, the Pearson correlation coefficient was used to calculate the relationship between the variables, the main one being the operating power P of the cutting table. g The operating power P of chopping arsenic q The operating power P of the throwing blower p The operating power P of the hydraulic pump 7 in the feed section w Five variables, including work efficiency (TH).

[0233] The formula for calculating the Pearson correlation coefficient is as follows:

[0234]

[0235] The correlation coefficients between each pair of the five variables can be calculated using the above formula.

[0236] Using the real-time power consumption of key operating components during the harvesting process of a self-propelled silage harvester as model input, a univariate linear regression model between single power data and operating efficiency is established:

[0237]

[0238] Step S800: Construction of work efficiency model. Establish a data model of work efficiency and the power of each component, and conduct single-factor analysis and coefficient of determination analysis. Based on the above single-factor regression analysis and correlation factor analysis, determine the top three factors affecting work efficiency, and establish a multiple linear regression work efficiency detection model.

[0239] Based on the correlation coefficient analysis in step S700, the influence relationship of each working component on the working efficiency and the univariate linear function model between each working component and the working efficiency are obtained. Combining the above correlation factor analysis and univariate regression analysis, the top three factors affecting the working efficiency are determined, and a multiple linear regression working efficiency detection model is established. The multiple linear regression equation is constructed as follows.

[0240] By screening the main factors affecting the operating efficiency of self-propelled silage harvesters, a least squares multiple regression model of multiple power data and operating efficiency was established, as shown in the figure below, to achieve multi-parameter correction and fusion of operating efficiency detection.

[0241]

[0242] Further residual and accuracy analysis was performed on the constructed multivariate linear model of work efficiency. If no abnormal detection points were found, it indicates that the detection model can well match the original data.

[0243] Operational efficiency is a key factor affecting the harvesting performance of self-propelled silage harvesters. This invention utilizes a power monitoring sensor, a key component installed non-destructively, to achieve real-time acquisition of operating data such as machine speed, torque, hydraulic flow, hydraulic pressure, and operating power throughout the entire operation process of the self-propelled silage harvester, including the cutting, feeding, chopping, and throwing sections. A data preprocessing algorithm combining Mann-Kendall boundary detection and time lag correction is applied to effectively filter the operating data. Outliers and missing values ​​are corrected and supplemented using Globus detection and cubic spline interpolation. Single-factor and multi-factor least-squares operational efficiency measurement models are constructed based on Pearson correlation analysis. By combining the operating data from multiple components for fusion regression, accurate measurement of the operational efficiency of the self-propelled silage harvester is achieved. This invention provides an effective technical means to solve the problem of rapid and accurate measurement of the operational efficiency of self-propelled silage harvesters in complex operating environments and for intelligent field assessment.

[0244] Of course, the present invention may have other various embodiments. Without departing from the spirit and essence of the present invention, those skilled in the art can make various corresponding changes and modifications according to the present invention, but these corresponding changes and modifications should all fall within the protection scope of the appended claims.

Claims

1. A method of monitoring the operating efficiency of a self-propelled silage harvester, characterised in that, The method comprises the following steps: S100, real-time acquisition of field data, during field operation of the self-propelled silage harvester, field operation data are continuously acquired by a working condition sensor and a data acquisition device, and the monitoring data are transmitted in real time to a vehicle-mounted processing terminal through a vehicle-mounted communication device; S200, acquisition of working component power, the vehicle-mounted processing terminal acquires the consumption power of the working component according to the received monitoring data; S300, screening of working condition data, the vehicle-mounted processing terminal identifies the rising and falling mutation points of the data region of the harvesting section by using a Mann-Kendall screening method, and classifies and screens the monitoring data by extracting the monitoring data of the chopping roller as the division basis; S400, the vehicle-mounted processing terminal establishes a data time delay analysis correction model through a material flow sequence conveying model of a header, a feeding roller, a chopping roller, a grain crushing roller, a throwing fan and a throwing cylinder in the self-propelled silage harvester; S500, detection of data abnormal points, a confidence level Grubbs criterion is used to determine the singular data by one-by-one inspection and data elimination; S600, filling of data missing values, a cubic spline interpolation method is used to fill the missing values, and the removed abnormal values and missing values are interpolated to increase the sample data; S700, analysis of main factor correlation, the correlation between each variable and the operation efficiency is detected, a Pearson correlation coefficient is used to calculate the correlation between variables and obtain a correlation coefficient, a linear model is constructed, and the influence relationship of each working component on the operation efficiency is obtained; and S800, construction of an operation efficiency model, a data model of the operation efficiency and each component power is established, single factor analysis and a determination coefficient analysis are performed, the main factors affecting the operation efficiency are determined according to the single factor regression analysis and the related factor analysis, and a multi-linear regression operation efficiency detection model is established.

2. The self-propelled silage harvester work-efficiency monitoring method according to claim 1, characterized in that, The working condition sensor comprises a header rotating speed sensor, a header torque sensor, a chopping roller rotating speed sensor, a chopping roller torque sensor, a throwing fan rotating speed sensor, a throwing fan torque sensor, a feeding part hydraulic pump flow sensor, a feeding part hydraulic outlet pressure sensor and a feeding part hydraulic inlet pressure sensor.

3. The self-propelled silage harvester work efficiency monitoring method according to claim 1 or 2, characterized in that, In step S200, mechanical component power conversion and hydraulic component power conversion are included, the consumption power of a mechanical rotating component is: Wherein, P r is the power consumed by the mechanical component, n r is the rotational speed of the mechanical component, T r is the rotational torque of the mechanical component, R is the rotational radius of the mechanical component, F is the rotational force of the mechanical component, and v r is the linear speed of the mechanical component. The consumption power of a hydraulic drive component is: wherein P h is the power consumption of the hydraulic component, p0 is the hydraulic pump inlet pressure, p i is the hydraulic pump outlet pressure, q is the hydraulic flow output by the hydraulic pump, n0 is the real-time engine speed, k is the hydraulic pump transmission speed ratio, V0 is the rated flow of the hydraulic pump, and η v is the volumetric efficiency of the hydraulic pump.

4. The self-propelled silage harvester work efficiency monitoring method according to claim 1 or 2, characterized in that, In step S300, the working data of the chopping roller are used as the division basis to classify and screen the monitoring data, and further comprising: S301, defining a working data point category set D={d1, d2, d3, d4}={1, 2, 3, 0}; S302, preliminary classification according to the chopping roller rotating speed to obtain a test section and a non-test data section: Wherein, D i is the i-th sampling data, n c is the real-time crushing speed of the chopping roller, n ce is the rated speed of the chopping roller; S303, detecting data mutation points by using a Mann-Kendall non-parametric test method to extract effective operation section data; The test data section in the original torque detection sequence is decomposed and used independently to detect the load loading mutation point and the load disappearance mutation point respectively; according to the data sequence X = {x1, x2, x3, …, xn} of n torque detection values, the cumulative number of torque values greater than the torque value at the jth time point at the ith time point in the sequence is calculated in a forward traversal, and a positive sequence rank sequence S n is constructed. k ​ Computing a sequence of positive order statistics UF k : Assuming that the detection values of the X sequence are randomly independent and follow the same distribution, the mean E(S k ) and variance var(S k ) of the sequence S k ) are: S304、According to the reverse detection sequence XB={x n ,x n-1 ,…,x1}, the above process is repeated in reverse order to construct the reverse order sequence S Kb , calculate the positive sequence statistics UB k , and take the negative value to get the reverse sequence statistics UB k ; S305、According to the law of large numbers, take the significance level of 0.05, in the confidence interval U 0.05 The constraint condition is solved as follows, and the effective solution is the rising and falling mutation points of the operation component torque data:

5. The self-propelled silage harvester work efficiency monitoring method according to claim 1 or 2, characterized in that, In step S400, the effective data intervals of the cutting table, the feeding part, the chopping roller and the blowing fan are X g , X w , X q , X p , and the corresponding data time delay analysis correction models are respectively: X g = [x(t cs -Δt1-Δt2), x(t cs -Δt1-Δt2+T),..., x(t cs -Δt1-Δt2+iT),..., x(t ce -Δt1-Δt2)] ; X w = [x(t cs -Δt2), x(t cs -Δt2+T),...,x(t cs -Δt2+iT),...,x(t ce -Δt2)] ; X q = [x(t cs ), x(t cs + T),..., x(t cs + iT),..., x(t ce )] ; X p = [x(t bs ), x(t bs + T),..., x(t bs + iT),..., x(t be )] ; Wherein, t cs , t ce are the rising and falling points of the torque mutation of the chopping roller; t bs , t be are the rising and falling points of the torque mutation of the blowing fan; T is the data collection interval of the sample; Δt1 is the time lag of the increase of the feeding part load relative to the feeding of the crop; Δt2 is the time lag of the increase of the chopping roller load relative to the increase of the feeding part load; Δt3 is the time lag of the increase of the blowing fan load relative to the increase of the chopping roller load.

6. The self-propelled silage harvester work efficiency monitoring method according to claim 1 or 2, characterized in that, In step S500, further comprising: S501、S501, the singular data points include outliers and missing values, the sample data X = [x(t cs +T),…,x(t cs +iT),…,x(t cs +pT)] collected at a fixed frequency T, according to the Bessel formula, the sample mean and standard deviation statistics are calculated: wherein is the mean of the data sample and σ is the standard deviation of the data sample; S502, reorganize the sample in value size order as {x(t)1, x(t)2, …, x(t)p}, wherein x(t)1≤x(t)2≤…x(t)p, determine the residual v of the suspected singular point according to the following formula i : S503、Determine whether the suspicious residual v i If the value of the suspicious residual satisfies the following formula, compared with the Grubbs coefficient critical value G0, the point is determined as an outlier, and the sequence is retested after being removed. Wherein, G0 can be obtained by table lookup, and α is a significance level; S504, repeating steps S501-S503 until there is no singular point in the data sequence.

7. The self-propelled silage harvester work efficiency monitoring method according to claim 1 or 2, characterized in that, In step S600, the missing value is filled by using cubic spline interpolation method, and the function S(x) satisfies the following conditions: n points (x i ,y i )(i = 1, 2,... n) on a plane, where (x1 n ), called sample points, a function passes through the sample points: S(x i ) = y i (i = 1, 2,..., n); S(x) is a cubic polynomial on each sub-interval [x i ,x i+1 ] S(x) = c1(x - x0) + c2(x - x0)2+ c3(x - x0)3+ c4; (1) i ) 3 +c2(x-x i ) 2 +c3(x-x i )+c4; S(x) has continuous first and second derivatives over the entire interval [xl, x n ].

8. The self-propelled silage harvester work efficiency monitoring method according to claim 1 or 2, characterized in that, In step S700, further comprising: S701、variables include the work power P of the header g , the work power P of the chopping reel q , the work power P of the thrower fan p , the work power P of the feeding portion hydraulic pump w and the work efficiency TH, the correlation coefficient between each two of the five variables is calculated by the following formula: and S702, according to the real-time power consumption of each key operation component in the self-propelled silage harvester harvesting process as the model input, a linear regression model of single power data and operation efficiency is established as follows:

9. The self-propelled silage harvester work efficiency monitoring method according to claim 1 or 2, characterized in that, In step S800, the first three factors affecting the operation efficiency are determined, and a multiple linear regression operation efficiency detection model is established as follows:

10. A self-propelled silage harvester work efficiency monitoring device, characterized by, The self-propelled silage harvester operation efficiency monitoring method for realizing any one of claims 1-9.

Images