A tractor test data acquisition and fault diagnosis system

Abstract

The tractor test data acquisition and fault diagnosis system belongs to the technical field of agricultural machinery, and comprises a multi-source heterogeneous data acquisition module, a dynamic state evaluation module and a fault diagnosis module.The multi-source heterogeneous data acquisition module acquires real-time data of the working state of the whole machine through sensors arranged at key positions of the tractor, acquires vibration signals measured by an engine body acceleration sensor, acquires real-time rotating speed measured by an engine rotating speed sensor, acquires load torque measured by a power output shaft torque sensor, and acquires current instruction rotating speed obtained by a vehicle-mounted control unit.The acquired original data are transmitted to the dynamic state evaluation module in real time.The dynamic state evaluation module receives multi-dimensional data, performs fusion analysis on the multi-dimensional data based on an instantaneous impact index model, and calculates the instantaneous impact index.The system has comprehensive and robust perception ability for risks, can accurately identify early and complex impact risks, and greatly reduces the misjudgment rate caused by single index fluctuation.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of agricultural machinery, in particular to a tractor test data acquisition and fault diagnosis system. BACKGROUND

[0002] In the actual application of tractors, especially in the complex and changeable field operation environment, there is an inherent contradiction between the running stability and the operation efficiency. When monitoring the state of the tractor, the prior art usually adopts a method of setting a single physical quantity static threshold, for example, monitoring the engine speed or the body vibration amplitude. When a certain index exceeds the preset limit, the system will trigger an alarm or take protective measures.

[0003] The limitations of this traditional method are:

[0004] Response lag and incomplete judgment. This method is essentially a lagging and non-comprehensive judgment mode. It is difficult to effectively distinguish between benign transient fluctuations caused by normal working conditions and potentially destructive systematic shocks, often leading to delayed response or false alarms.

[0005] Cannot fundamentally solve the core contradiction. Since the judgment basis is single and the response mechanism is rough, the prior art cannot fundamentally solve the contradiction between how to suppress sudden impact loads and maintain high operation efficiency under complex working conditions.

[0006] Lack of predictive intervention ability. The traditional method is based on post-response and lacks the ability to predict future impact risks, so it cannot take preventive measures before the impact has a significant impact, thereby making it difficult to ensure the smoothness and reliability of the tractor operation.

[0007] The above shortcomings are mainly due to the limitations of state monitoring methods and data processing technologies, which result in insufficient and timely understanding of the working state of the tractor, making it impossible for the control strategy to consider both stability and efficiency. SUMMARY

[0008] The purpose of the present application is to provide a tractor test data acquisition and fault diagnosis system to solve the problems raised in the background art.

[0009] In order to achieve the above-mentioned purpose of the application, the technical solution adopted by the present application is as follows:

[0010] A tractor test data acquisition and fault diagnosis system, comprising:

[0011] A multi-source heterogeneous data acquisition module acquires real-time data of the working state of the whole machine through a sensor group deployed at key positions of the tractor; acquires vibration signals measured by an engine body acceleration sensor; acquires real-time speed measured by an engine speed sensor The load torque measured by the torque sensor on the power output shaft is collected. ; Collect the current command speed obtained from the vehicle control unit The collected raw data is transmitted to the dynamic status assessment module in real time.

[0012] The dynamic state assessment module receives multidimensional data; it performs fusion analysis on the multidimensional data based on the instantaneous impact index model; and it calculates the instantaneous impact index. ;

[0013] The predictive control strategy generation module receives the instantaneous shock index. ; Calculate engine control intervention amount based on adaptive control command adjustment model ;

[0014] The control execution and feedback module, based on the amount of control intervention... and the current command speed Calculate the final control command The engine operating status is adjusted by executing the final control command through the tractor electronic control unit; the actual status parameters of the whole machine after execution are collected; and the actual status parameters are fed back to the multi-source heterogeneous data acquisition module.

[0015] Preferably, the instantaneous impact index model in the dynamic state assessment module is a multi-factor weighted prediction model established based on first-principles analysis of the system's dynamic response; instantaneous impact index The calculations include:

[0016] Calculate the rate of change of load torque ; Calculate the rate of change of load torque Rate of change of maximum expected load torque The ratio; calculate the real-time rotational speed. With commanded speed absolute value of deviation ; Calculate the absolute value of the deviation and the upper limit of the speed deviation. The ratio; calculate the critical band energy change rate. ; Calculate the rate of change of energy in the critical frequency band With respect to the maximum expected critical band energy change rate The ratios; multiply the three ratios by their corresponding weights respectively. , , The instantaneous impact index is obtained by summing the results. .

[0017] Preferably, the critical band energy change rate in the dynamic state assessment module The calculations include:

[0018] The time spectrum is obtained by performing a short-time Fourier transform on the original vibration signal. ; determine a preset critical band range ; integrate the time-frequency spectrum amplitude square in the critical band range to obtain the critical band power at the time instant ; take time derivative of the critical band power to obtain the critical band energy rate of change ; the critical band range is determined by modal analysis of the tractor powertrain.

[0019] Preferably, the adaptive control command adjustment model in the predictive control strategy generation module is a nonlinear control law for realizing minimum intervention; the calculation of the control intervention amount includes:

[0020] When the instantaneous impact index is less than or equal to the warning threshold , the control intervention amount is set to zero; when the instantaneous impact index is greater than the warning threshold , the difference between the instantaneous impact index and the warning threshold is calculated; the ratio of the difference to the limit threshold and the difference between the warning threshold is calculated; the ratio is raised to the power of ; the result of the operation is multiplied by the maximum allowed control adjustment amount to obtain the control intervention amount .

[0021] Preferably, the determination method of each parameter in the predictive control strategy generation module is as follows:

[0022] The maximum allowed control adjustment amount is preset according to the performance curve of the tractor engine; the warning threshold and the limit threshold are determined by statistical analysis of historical field operation data; the maximum expected load torque change rate , the upper limit of the speed deviation , and the maximum expected critical band energy rate of change are determined by taking the 99th percentile of the corresponding physical quantities in the historical data; the weights , , satisfy the normalization constraint condition ; the control response sensitivity index is set to a preset constant greater than 1.

[0023] ​The specific structure of the multi-source heterogeneous data acquisition module preferably comprises:

[0024] The engine body acceleration sensor measures three-axis vibration acceleration signals at a preset sampling frequency; the engine speed sensor monitors the crankshaft speed and calculates the real-time speed ; the power output shaft torque sensor measures the transmission system load torque ; the CAN bus interface of the vehicle-mounted control unit obtains the command speed corresponding to the current operation command ; the data synchronization unit synchronizes the above sensor data according to a unified timestamp; and the data transmission unit forms a multi-dimensional time sequence data stream and transmits it to the dynamic state evaluation module.

[0025] The specific structure of the control execution and feedback module preferably comprises:

[0026] The control instruction generation unit sends the control intervention amount to the tractor electronic control unit through the CAN bus; the execution unit causes the electronic control unit to adjust the engine fuel injection amount according to the final control instruction ; the state acquisition unit acquires the adjusted engine actual operating parameters; the feedback transmission unit re-enters the actual operating parameters into the multi-source heterogeneous data acquisition module; and the closed-loop control unit realizes dynamic evaluation and decision-making based on the latest actual state in each control cycle.

[0027] The sensor group preferably comprises:

[0028] An engine body acceleration sensor for measuring three-axis vibration acceleration signals of the engine; an engine speed sensor for monitoring the real-time speed of the crankshaft; a power output shaft torque sensor for measuring the transmission system load torque; and a vehicle-mounted control unit for obtaining the command speed corresponding to the operation command and performing data communication through the CAN bus.

[0029] The dynamic state evaluation module preferably further comprises:

[0030] A data receiving unit for receiving multi-dimensional time sequence data from the multi-source heterogeneous data acquisition module; a data preprocessing unit for standardizing the format and filtering noise of the received data; a risk calculation unit for calculating the instantaneous impact index based on the instantaneous impact index model ; and a result output unit for transmitting the calculated instantaneous impact index to the predictive control strategy generation module.

[0031] The present application provides a tractor test data acquisition and fault diagnosis system by improvement, compared with the prior art, has the following improvements and advantages:

[0032] 1. This scheme constructs a system that can jointly judge the severity of the impact from multiple perspectives such as load, speed, vibration, etc. For example, a load impact that may be ignored in the prior art due to its small amplitude will be captured in this scheme due to its extremely high rate of change. Similarly, a speed fluctuation that is within the allowable range alone will be amplified if it occurs simultaneously with a sharp increase in vibration power, and their combined effect will be identified by the system. This fusion mechanism enables the system to have a comprehensive and robust perception of risk, accurately identifying early and complex impact risks, and greatly reducing the misjudgment rate caused by fluctuations in a single indicator;

[0033] 2. The integration of the frequency band is physically equivalent to the collection of all vibration energy contributions in a specific frequency band, and has higher tolerance to small shifts in noise and frequency. The critical frequency band power change rate calculated in this way can more stably and truly reflect the system vibration state transition caused by structural impact, providing high-quality and high signal-to-noise ratio input for the calculation of the instantaneous impact index;

[0034] 3. The nonlinear control of this scheme realizes intelligent response. When the impact index just exceeds the warning threshold, the control intervention grows very slowly, avoiding overreaction to minor disturbances and ensuring smooth operation. However, once the impact index approaches the limit threshold, the intervention will grow exponentially, ensuring decisive and powerful suppression of malignant impacts. This strategy achieves a better balance between safety and efficiency;

[0035] 4. This closed-loop structure ensures that the system can perform dynamic evaluation and decision-making based on the latest actual state at each control period. This ability enables the system not only to issue control instructions but also to immediately know the actual effect of the instructions and correct the behavior at the next time based on this. This enables the tractor to maintain a dynamic balance between stability and efficiency when facing the variable and unpredictable work load in the field, significantly enhancing the robustness, environmental adaptability, and overall performance of the system. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 is a flow chart of the system of the present application. DETAILED DESCRIPTION

[0037] To make the purpose, technical scheme and advantages of the present application clearer, the present application is further described in detail below with specific examples.

[0038] Example 1:

[0039] Please refer to Figure 1 The present application provides a tractor test data acquisition and fault diagnosis system, comprising:

[0040] The multi-source heterogeneous data acquisition module captures real-time data on the overall machine's operating status through sensor arrays deployed at key parts of the tractor; it acquires vibration signals measured by the engine block acceleration sensor; and it acquires real-time speed data measured by the engine speed sensor. The load torque measured by the torque sensor on the power output shaft is collected. ; Collect the current command speed obtained from the vehicle control unit The collected raw data is transmitted to the dynamic status assessment module in real time.

[0041] The dynamic state assessment module receives multidimensional data; it performs fusion analysis on the multidimensional data based on the instantaneous impact index model; and it calculates the instantaneous impact index. ;

[0042] The predictive control strategy generation module receives the instantaneous shock index. ; Calculate engine control intervention amount based on adaptive control command adjustment model ;

[0043] The control execution and feedback module, based on the amount of control intervention... and the current command speed Calculate the final control command The tractor's electronic control unit executes final control commands to adjust the engine's operating status; collects the actual status parameters of the entire machine after execution; and feeds back the actual status parameters to the multi-source heterogeneous data acquisition module.

[0044] This embodiment discloses a tractor test data acquisition and fault diagnosis system. The purpose of the system provided by this invention is to solve the inherent contradiction between the ability of tractors to suppress sudden impact loads and maintain high operating efficiency in complex field operation environments. By constructing a feedforward-feedback closed-loop intelligent control system, predictive intervention on the engine operating status is achieved, thereby improving the vibration stability and operational reliability of the whole machine without affecting normal operating efficiency.

[0045] The system is built as a complete technical closed loop, and its implementation includes the following four collaborative modules:

[0046] The multi-source heterogeneous data acquisition module is used to comprehensively and in real-time capture key physical quantities reflecting the dynamic working state of the tractor. In this embodiment, the module performs data acquisition through a series of sensors deployed at key locations in the tractor powertrain; it acquires raw vibration signals characterizing the overall vibration state through an engine block acceleration sensor; and it acquires the actual crankshaft speed through an engine speed sensor, denoted as... The real-time load torque of the transmission system is collected by a torque sensor mounted on the power output shaft and recorded as follows: Meanwhile, the target rotating speed corresponding to the current operation instruction, i.e., the instruction rotating speed, is obtained by reading the vehicle controller local area network bus data After all the collected raw data are given a unified time stamp, multi-dimensional time sequence data flow is formed and transmitted to the dynamic state evaluation module for next step processing;

[0047] The dynamic state evaluation module is used for fusion analysis of the collected multi-dimensional data to generate a unified index capable of quantitatively evaluating the instantaneous impact risk of the tractor. In this embodiment, the module receives multi-dimensional time sequence data from the previous module; a preset instantaneous impact index model is integrated inside, and the core function of the model is to fuse data of different physical dimensions and change characteristics into a dimensionless and standardized risk index; the module calculates the input data through the model, and finally outputs an instantaneous impact index changing with time ;

[0048] The predictive control strategy generation module is used for dynamically generating an optimal engine control intervention strategy according to the evaluated risk level. In this embodiment, the module receives the instantaneous impact index calculated by the dynamic state evaluation module ; based on the built-in self-adaptive control instruction adjustment model, the input risk index is converted into a specific engine control intervention amount ; the adjustment logic of the model aims to realize minimum intervention, i.e., intervention is only performed when the risk reaches a certain level, and the intervention intensity is nonlinearly related to the risk level;

[0049] The control execution and feedback module is used for implementing the generated control strategy and verifying the control effect, thereby forming a closed-loop control. In this embodiment, the module receives the control intervention amount ; according to the current instruction rotating speed and the control intervention amount , the final control instruction is calculated:

[0050]

[0051] wherein, : the final control instruction at time t; : the instruction rotating speed, i.e., the target rotating speed corresponding to the current operation instruction; : the control intervention amount at time t, indicating the adjustment amount to be subtracted from the current instruction rotating speed;

[0052] The command is sent to the tractor electronic control unit via the CAN bus. The ECU adjusts the engine's fuel injection quantity and other actuators to change the engine's actual operating state. After the control is executed, the module immediately collects the adjusted actual state parameters of the whole machine, such as the new speed and vibration signal, and uses the actual state parameters as the latest input to feed back to the initial end of the system - the multi-source heterogeneous data acquisition module.

[0053] Through the sequential execution and closed-loop feedback of the above four modules, this system can monitor the tractor's load, vibration, and speed status in real time, and predict the risk of severe vibration or stall caused by sudden load changes. When the system anticipates the impact risk, it can proactively and smoothly adjust the engine command and appropriately reduce the speed to absorb the impact, avoiding overshoot and oscillation caused by the delayed response in traditional systems. This predictive intervention mechanism significantly improves the smoothness and reliability of the tractor's operation under harsh conditions. At the same time, because the intervention is precise and minimal, it maximizes the guarantee of work efficiency, thereby achieving synergistic optimization of vibration stability and work efficiency.

[0054] The tractor test data acquisition and fault diagnosis system disclosed in this invention has achieved substantial technical progress in improving the operational stability and efficiency of tractors during operation compared with the prior art; the progress is rooted in its unique system architecture and inherent algorithm logic.

[0055] Existing technologies for monitoring tractor status typically rely on setting static thresholds for single physical quantities, such as vibration amplitude or engine speed. When a certain indicator exceeds the limit, an alarm is triggered or coarse protective measures are taken. The inherent drawback of this approach is that it is a delayed and non-comprehensive judgment, which makes it difficult to distinguish between benign transient fluctuations and potentially destructive systemic shocks. This often leads to false alarms or delayed responses and cannot fundamentally resolve the contradiction between stability and efficiency under complex operating conditions.

[0056] Example 2

[0057] The instantaneous impact index model in the dynamic state assessment module is a multi-factor weighted prediction model established based on first-principles analysis of the system's dynamic response; instantaneous impact index The calculations include:

[0058] Calculate the rate of change of load torque ; Calculate the rate of change of load torque Rate of change of maximum expected load torque The ratio; calculate the real-time rotational speed. With commanded speed absolute value of deviation ; Calculate the absolute value of the deviation and the upper limit of the speed deviation. The ratio; calculate the critical band power change rate. ; calculate the critical band power change rate with the maximum expected critical band power change rate ; multiply the above three ratios by the corresponding weight , , and then sum to obtain the instantaneous impact index ;

[0059] As a specific embodiment, the generation method of the instantaneous impact index in the dynamic state evaluation module of Example 1 is further limited; the instantaneous impact index model used by the module is designed based on the first principle analysis of the system dynamic response to establish a multi-factor weighted prediction model, aiming to fuse multiple key precursor indicators related to impact risk;

[0060] The calculation process of the instantaneous impact index is defined as a three-factor weighted summation mathematical expression as follows:

[0061]

[0062] wherein, : instantaneous impact index, a dimensionless scalar value, used to quantify the comprehensive impact risk level of the tractor at time t, calculated and output by this module;

[0063] : load torque change rate, unit N·m / s, derived from the time derivative of the load torque sequence collected by the multi-source heterogeneous data collection module , and its role is to represent the severity of external load change;

[0064] : maximum expected load torque change rate, with the same unit as , derived from statistical analysis of a large amount of historical operation data, taking the 99th percentile value of , and its role is to serve as a normalization denominator to convert the load torque change rate into a dimensionless ratio;

[0065] : real-time speed, unit r / min, derived from real-time collection by the engine speed sensor;

[0066] : command speed, unit r / min, derived from the vehicle-mounted control unit CAN bus interface;

[0067] : upper limit of speed deviation, unit r / min, also derived from historical data statistics, taking the 99th percentile value of the load rate of change, the role is as a normalization denominator, the absolute value of the speed deviation is converted into a dimensionless ratio;

[0068] : critical band power rate of change, the calculation method is described in the next embodiment, the role is to represent the increase and decrease rate of core vibration power related to impact and failure;

[0069] : maximum expected critical band power rate of change, the unit is the same as , the source is also the 99th percentile value set in the historical data , the role is to convert the vibration power rate into a dimensionless ratio as a normalization denominator;

[0070] : weight coefficient, three preset constants of dimensionless, set to meet the normalization constraint condition , the role is to adjust the relative importance of load, speed and vibration in comprehensive risk assessment according to different working conditions or diagnostic targets;

[0071] The technical effect of the embodiment is that by normalizing and weighting the three physically different but internally related indicators of load rate of change, speed deviation and vibration power rate of change, a more comprehensive and robust risk assessment system is constructed; Compared with the traditional method relying on a single threshold, the index can comprehensively reflect the performance of impact events in multiple dimensions, effectively avoid the misjudgment caused by the fluctuation of single indicator, significantly improve the accuracy and reliability of risk assessment, and provide high-quality decision basis for subsequent accurate control;

[0072] The technical scheme of the application constructs a complete closed-loop control system from pre-prediction to post-feedback, and its beneficial effects are reflected in the following aspects:

[0073] Its core advantage lies in the introduction of an instantaneous impact index calculated based on a multi-factor weighted prediction model; The physical meaning of the index is to construct a scalar field with dimension one that can represent the systematic risk of the powertrain; The mathematical expression

[0074]

[0075] Among them, : weight, : torque rate of change, : speed, : power, subscript These represent the maximum value, the command, and the upper limit of deviation, respectively. It's not a simple numerical addition, but rather a weighted fusion of three observations—the rate of change of load torque, the deviation of engine speed, and the rate of change of vibration power in the key frequency band—which are physically different but highly correlated in terms of risk causes. These observations are normalized and then weighted. This fusion mechanism allows the system to collaboratively assess the severity of the impact from multiple perspectives. A load impact that might be ignored in existing technologies, with a small amplitude but an extremely high rate of change, will be considered significant in this invention due to its impact on… The significant contribution of vibration is captured; a single fluctuation in rotational speed that seems within acceptable limits, when occurring simultaneously with a sharp increase in vibrational power, will have a combined effect. The above is amplified; the system gains a comprehensive and robust perception of risks, greatly reducing the misjudgment rate of fluctuations in a single indicator, and achieving accurate identification of early and complex shock risks.

[0076] Critical band power change rate in dynamic state assessment module The calculations include:

[0077] The time spectrum is obtained by performing a short-time Fourier transform on the original vibration signal. Determine the preset critical frequency band range. Within the critical frequency band The inner time spectrum amplitude squared is obtained by integrating. Critical band power at time For critical band power The critical band power change rate is obtained by taking the time derivative. Critical frequency band range The powertrain of the tractor was determined through modal analysis;

[0078] In this embodiment, the critical band power change rate in the dynamic state assessment module is further defined. The calculation method aims to extract the vibration features most relevant to load impact or early faults from complex broadband vibration signals in a stable and accurate manner.

[0079] The calculation process is as follows: determine a preset critical frequency band range. The frequency band is not arbitrarily selected, but is determined in advance based on modal analysis of the tractor powertrain or statistical analysis of historical fault data. The principle is that when a specific structure is subjected to impact or develops early cracks, the vibration energy will be concentrated in certain specific frequency ranges that are sensitive to its damage. This range is defined as the critical frequency band.

[0080] For example, modal analysis of the powertrain of a certain type of four-cylinder diesel engine revealed that when its transmission system is subjected to torsional impact, its first-order torsional resonance frequency is mainly concentrated in the frequency band of 150Hz to 250Hz. Therefore, in this embodiment, the critical frequency band range can be preset to [specific value]. Hz, specifically designed to capture energy changes associated with key vibrational modes;

[0081] A short-time Fourier transform is performed on the raw vibration signal input from the multi-source heterogeneous data acquisition module to obtain a two-dimensional time spectrum. The spectrum shows the distribution of signal power over time and frequency;

[0082] In the computational implementation, in order to achieve a balance between time-domain and frequency-domain resolution, the Hanning window can be used as the window function for the short-time Fourier transform. The window length is set to 1024 sampling points, and the overlap rate between windows is 50%. For a sampling frequency of 20kHz, such parameter settings can provide a frequency resolution of about 19.5Hz and a time resolution of 25.6ms, which is sufficient for effective analysis of the critical frequency band of interest.

[0083] Based on the aforementioned determined critical frequency band Within the range, the critical band power at time t is calculated by integrating the square of the time-frequency amplitude. The mathematical model is as follows:

[0084]

[0085] in:

[0086] Critical frequency band power, its physical meaning is the total vibration power within a preset frequency band at time t, and its dimension is... This is calculated in this step;

[0087] The time-frequency spectrum is the result of the original vibration signal undergoing a short-time Fourier transform, and its dimensions are: ;

[0088] The lower and upper limits of the critical frequency band, in Hz, are preset parameters;

[0089] By analyzing the calculated power time series By taking the derivative over time, the rate of change of critical band power can be obtained. ; :time;

[0090] Considering the unavoidable noise in sensor signals, directly performing differential calculations would amplify the noise and affect the accuracy of the rate of change calculation; therefore, when calculating the torque sequence... and critical band power When time derivation is performed, the embodiment preferably adopts a Savitzky-Golay filter; the filter can realize data smoothing filtering and high-order derivative calculation at the same time by least square fitting of data in a sliding window with a polynomial; for example, a S-G filter with a window length of 11 and a polynomial order of 3 can be selected to robustly calculate the load torque change rate and critical band power change rate ;

[0091] The result will be the calculation of the above exponent

[0092] The technical effect brought by the embodiment is to significantly improve the stability and accuracy of vibration feature extraction; in the traditional vibration analysis method, the main frequency will jump and drift under the load of variable working conditions, resulting in unstable features; the method avoids the difficulty of tracking a single frequency by integrating the power of the preset frequency band with clear physical meaning; this power integration idea is not sensitive to the small drift of frequency, and can more robustly capture the systematic vibration power change caused by load impact, so that the risk index is more accurate and reliable in describing the impact response of the real physical world;

[0093] In the representation of the vibration signal, the application discards the traditional method of tracking a single and easy-to-jump main frequency, and instead calculates the vibration power in the preset critical frequency band :

[0094]

[0095] wherein, : time-frequency spectrum, : frequency, : time; the critical frequency band at t is determined in advance by modal analysis of the power assembly, which corresponds to the frequency range most sensitive to load impact or most relevant to early fault features; this integration of the frequency band is physically equivalent to gathering all vibration energy contributions in a certain frequency range, and the result has higher tolerance to noise and frequency drift compared to point frequency measurement; the calculated critical band power change rate can more stably and truly reflect the system vibration state transition caused by structural impact, providing high-quality and high signal-to-noise ratio input for the calculation of the instantaneous impact index .

[0096] Embodiment 3

[0097] The adaptive control instruction adjustment model in the predictive control strategy generation module is a nonlinear control law aiming to minimize intervention; the control intervention amount The calculation includes:

[0098] When the instantaneous impact index is less than or equal to the warning threshold , the control intervention amount is set to zero; when the instantaneous impact index is greater than the warning threshold , the difference between the instantaneous impact index and the warning threshold is calculated; the ratio of the difference to the limit threshold and the difference between the warning threshold is calculated; the ratio is raised to the power of ; the result is multiplied by the maximum allowed control adjustment amount to obtain the control intervention amount ;

[0099] In this embodiment, the calculation method of the control intervention amount in the predictive control strategy generation module is limited; the adaptive control instruction adjustment model used by the module is designed from a nonlinear control law aiming to minimize intervention, and the core purpose is to minimize the disturbance to the normal operation process as much as possible under the premise of ensuring safety;

[0100] The calculation logic of the control intervention amount is designed as a segmented function, which is realized by a unified mathematical expression:

[0101]

[0102] The calculation logic and parameter definition of the expression are as follows:

[0103] The internal logic is that when the instantaneous impact index is less than or equal to the preset warning threshold , the calculation result is 0, and the control intervention amount is set to zero, which embodies the principle of no risk and no intervention; when is greater than the warning threshold , the calculation is started;

[0104] : control intervention amount, unit: r / min, indicating the adjustment amount that needs to be subtracted from the current instruction speed, calculated and output by this module;

[0105] : maximum allowable control adjustment, unit: r / min, is a preset engineering safety parameter, which limits the maximum amplitude of single control intervention, preventing the engine from stalling due to too fast speed drop;

[0106] : instantaneous impact index, input by the preceding dynamic state evaluation module;

[0107] : warning threshold, dimensionless, is a preset parameter, representing a critical point at which the risk starts to become significant and the system needs to start paying attention;

[0108] : limit threshold, dimensionless, is a preset parameter, representing a critical point at which the risk has approached the limit of the system and strong intervention is needed, and ;

[0109] : control response sensitivity index, is a preset constant greater than 1, for example, set to 2, which introduces nonlinear characteristics to the control response;

[0110] The technical effect of the embodiment is to realize an intelligent and nonlinear control response strategy; due to the existence of the index, when the risk index just exceeds the warning threshold , the growth of the control intervention amount is relatively gentle, avoiding overreaction to minor disturbances and ensuring smooth operation of the system; when approaches the limit threshold , the intervention amount will increase sharply, ensuring decisive and powerful suppression of major impact risks; this nonlinear control law makes the system have both smooth response and rapid suppression, and shows appropriate intervention strength when dealing with different levels of risks, thus achieving a better balance between safety and efficiency;

[0111] In the generation of the control strategy, the adaptive control instruction adjustment model adopted by the application has a core, which is a nonlinear control law that realizes minimum intervention:

[0112]

[0113] Among them, : control intervention amount, : maximum adjustment amount, : warning threshold, : limit threshold, : sensitivity index; the ingenious part of the control law is that the warning threshold and the control response sensitivity index greater than 1 are introduced.This achieves nonlinear control response; when the impact index Warning threshold When the fluctuation is within a small range, control the amount of intervention. The growth rate was extremely gradual, and the system demonstrated excellent disturbance suppression capabilities, avoiding unnecessary interference with normal operations; however, once the impact index... Approaching the limit threshold This leads to an exponential increase in the amount of intervention, ensuring decisive suppression of severe shocks; the methods for determining each parameter, such as the maximum permissible control adjustment amount... Based on the engine performance curve preset, warning and limit thresholds By conducting statistical analysis on historical field operation data, it was ensured that the implementation of the nonlinear control law had sufficient physical basis and data support.

[0114] The parameters in the predictive control strategy generation module are determined as follows:

[0115] The maximum permissible control adjustment amount is preset based on the tractor engine performance curve. Warning thresholds were determined through statistical analysis of historical field operation data. and limit threshold The maximum expected load torque change rate is determined by taking the 99th percentile of the corresponding physical quantity in historical data. upper limit of speed deviation and the maximum expected critical band power change rate Set weights , , Satisfy normalization constraints Set the control response sensitivity index A preset constant greater than 1;

[0116] In this embodiment, the determination methods of key parameters and thresholds in the predictive control strategy generation module and the dynamic state evaluation module are systematically explained to ensure that those skilled in the art can fully implement this technical solution without excessive experimentation.

[0117] The specific methods for determining each parameter are as follows:

[0118] Maximum permissible control adjustment amount The parameters are determined based on the tractor's engine performance curve. By analyzing the curve, a maximum speed reduction is preset to effectively absorb impact energy without causing the engine to enter an unstable operating range or stall. This value is [value to be filled in]. ;

[0119] Warning threshold and limit threshold The determination of the two thresholds is derived from statistical analysis of a large amount of historical field operation data; the method is to correlate the calculated value sequence in the historical data with the actual occurrence of adverse events such as vibration overrun or stall, and draw a relationship curve of the value and the probability of occurrence of adverse events; on the curve, the value corresponding to the inflection point of the significant increase of the probability of occurrence of adverse events in the relationship curve is selected as the warning threshold , and the value corresponding to the saturation point of the probability close to 1 is selected as the limit threshold ; ;

[0120] The maximum expected load torque rate of change , the upper limit of the speed deviation , and the maximum expected critical frequency band power rate of change : The determination method of the three normalization denominator parameters is consistent, that is, by statistically analyzing the long-term observation values of the corresponding physical quantities in the historical operation data, i.e. , , , and taking the value of the 99th percentile as the set value; the principle of this method is to select a statistical upper limit that can contain most of the normal working condition fluctuations, so that the normalized ratio can effectively highlight abnormal events;

[0121] The weight coefficient : The three parameters are set to satisfy the normalization constraint condition ; the values can be empirically configured or optimized by an optimization algorithm according to the main operation types of the tractor, such as plowing, rotary tillage, or the fault modes that need to be focused on, to adjust the contribution of different risk factors in the final evaluation;

[0122] As an implementation manner, for the rotary tillage operation of the tractor in heavy soil, since the mutation of the load torque under the working condition is the main factor causing the impact, and the stability of the speed is crucial to ensure the quality of rotary tillage, the weight coefficient can be empirically configured as follows: , , ; this set of weights highlights the contribution of the load torque rate of change, while considering the speed deviation and vibration state, thereby realizing targeted risk assessment;

[0123] The control response sensitivity index : the parameter is set to a preset constant greater than 1, and its working principle is to introduce nonlinearity by using the power function characteristic; by setting it to a constant greater than 1, such as 2, the control response curve can be concave, realizing the control effect of slow increase at the beginning and rapid increase at the end;

[0124] ​​​ The selection of the value aims to balance the smoothness and timeliness of the control response; a smaller value, for example, close to 1.1, will make the growth of the control intervention closer to linear, the response more moderate, and suitable for working conditions where the efficiency of the operation is extremely high and too much intervention is not desired; a larger value, for example, 3 or higher, will make the intervention curve steeper near the limit threshold, the response more aggressive, and suitable for critical tasks that require absolute safety and prevention of stall; in this embodiment, the value of 2 is selected as an engineering balance between smoothness and response sensitivity, ensuring that the intervention grows slowly at the beginning of the risk, and provides sufficient decisive suppression when the risk is significant. The selection of the value aims to balance the smoothness and timeliness of the control response; a smaller value, for example, close to 1.1, will make the growth of the control intervention closer to linear, the response more moderate, and suitable for working conditions where the efficiency of the operation is extremely high and too much intervention is not desired; a larger value, for example, 3 or higher, will make the intervention curve steeper near the limit threshold, the response more aggressive, and suitable for critical tasks that require absolute safety and prevention of stall; in this embodiment, the value of 2 is selected as an engineering balance between smoothness and response sensitivity, ensuring that the intervention grows slowly at the beginning of the risk, and provides sufficient decisive suppression when the risk is significant. The selection of the value aims to balance the smoothness and timeliness of the control response; a smaller value, for example, close to 1.1, will make the growth of the control intervention closer to linear, the response more moderate, and suitable for working conditions where the efficiency of the operation is extremely high and too much intervention is not desired; a larger value, for example, 3 or higher, will make the intervention curve steeper near the limit threshold, the response more aggressive, and suitable for critical tasks that require absolute safety and prevention of stall; in this embodiment, the value of 2 is selected as an engineering balance between smoothness and response sensitivity, ensuring that the intervention grows slowly at the beginning of the risk, and provides sufficient decisive suppression when the risk is significant. The selection of the value aims to balance the smoothness and timeliness of the control response; a smaller value, for example, close to 1.1, will make the growth of the control intervention closer to linear, the response more moderate, and suitable for working conditions where the efficiency of the operation is extremely high and too much intervention is not desired; a larger value, for example, 3 or higher, will make the intervention curve steeper near the limit threshold, the response more aggressive, and suitable for critical tasks that require absolute safety and prevention of stall; in this embodiment, the value of 2 is selected as an engineering balance between smoothness and response sensitivity, ensuring that the intervention grows slowly at the beginning of the risk, and provides sufficient decisive suppression when the risk is significant. The selection of the value aims to balance the smoothness and timeliness of the control response; a smaller value, for example, close to 1.1, will make the growth of the control intervention closer to linear, the response more moderate, and suitable for working conditions where the efficiency of the operation is extremely high and too much intervention is not desired; a larger value, for example, 3 or higher, will make the intervention curve steeper near the limit threshold, the response more aggressive, and suitable for critical tasks that require absolute safety and prevention of stall; in this embodiment, the value of 2 is selected as an engineering balance between smoothness and response sensitivity, ensuring that the intervention grows slowly at the beginning of the risk, and provides sufficient decisive suppression when the risk is significant.

[0125] The gain technical effect brought by this embodiment is to provide a scientific and reproducible calibration method for the core algorithm of the entire system; by deeply binding the setting of parameters with the physical characteristics of the engine and historical big data analysis, it is ensured that the threshold and coefficient of the model are not based on arbitrary speculation, but closely fit the objective operating rules of the specific tractor model in the real working environment; this greatly improves the practicality, reliability and adaptability of the system, making the technical solution easily deployed on different models of tractors and achieving the expected consistent control effect.

[0126] Embodiment 4

[0127] The specific composition of the multi-source heterogeneous data acquisition module includes:

[0128] The engine body acceleration sensor measures the three-axis vibration acceleration signal at a preset sampling frequency; the engine speed sensor monitors the crankshaft speed and calculates the real-time speed ; the power output shaft torque sensor measures the transmission system load torque ; the CAN bus interface of the vehicle-mounted control unit obtains the command speed corresponding to the current operation command ; the data synchronization unit synchronizes the above sensor data according to a unified timestamp; the data transmission unit forms a multi-dimensional time sequence data stream and transmits it to the dynamic state assessment module;

[0129] In this embodiment, the specific hardware and software composition of the multi-source heterogeneous data acquisition module is described in detail;

[0130] The module includes a group of sensors, a data synchronization unit and a data transmission unit; among them, the engine body acceleration sensor is a three-axis accelerometer that measures the three-axis vibration acceleration signal of the engine body at a preset sampling frequency that is much higher than the highest frequency required for analysis, for example, 20kHz; the engine speed sensor calculates and outputs the real-time speed ; The power output shaft torque sensor is integrated in the transmission chain to directly measure the load torque transmitted to the work implement ; The vehicle-mounted control unit provides the command speed corresponding to the current operation command through its CAN bus interface ; The data synchronization unit is used to solve the problem of different synchronization of different sensor data in time, receives independent data streams from all the above sensors, and applies a unified high-precision timestamp to each data point; The data transmission unit is responsible for integrating the time-synchronized data streams into structured multi-dimensional time series data streams and transmitting them to the dynamic state assessment module;

[0131] The gain technical effect of the embodiment is to ensure that the data of the input system model is high-quality, high-fidelity and time-synchronized; The setting of the data synchronization unit is crucial for the subsequent Exponential calculation is essential because the instantaneous correspondence between multiple indicators such as speed deviation and torque change rate is the basis for accurate risk judgment; Time-accurate multi-dimensional data stream is the guarantee for the whole closed-loop control system to respond in time and accurately, thereby improving the overall performance of the system.

[0132] The specific composition of the control execution and feedback module includes:

[0133] The control instruction generation unit calculates the final control instruction based on the control intervention amount sent to the tractor electronic control unit through the CAN bus; The execution unit makes the electronic control unit adjust the engine fuel injection amount according to the final control instruction ; The state acquisition unit acquires the actual operating parameters of the adjusted engine; The feedback transmission unit re-enters the actual operating parameters into the multi-source heterogeneous data acquisition module; The closed-loop control unit realizes dynamic assessment and decision-making based on the latest actual state in each control cycle;

[0134] In this embodiment, the internal functional unit composition and workflow of the control execution and feedback module are described in detail to illustrate the implementation mechanism of the closed-loop control;

[0135] The module includes a control instruction generation unit, an execution unit, a state acquisition unit, and a feedback transmission unit; The control instruction generation unit receives the control intervention amount , calculates the final control instruction; and encapsulates it into a CAN bus data frame for sending; the execution unit, i.e. the ECU of the tractor and the actuators controlled thereby, adjusts the fuel injection amount and other parameters of the engine after receiving the instruction, so that the actual engine speed tends to the new target; the state acquisition unit acquires the adjusted actual engine operation parameters through the sensor group immediately after the ECU executes the control instruction; the feedback transmission unit transmits the latest parameters reflecting the control effect back to the multi-source heterogeneous data acquisition module of the system as the input of the next control cycle; the cooperative working mechanism of the units ensures that the evaluation-decision-execution-re-evaluation process can be completed in each control cycle, so that the entire system can continuously evaluate and decide based on the latest actual state;

[0136] The gain technical effect of the embodiment lies in that the specific implementation path of the system closed-loop control is clarified; through the series of closely connected actions of instruction generation-execution-state acquisition-feedback, the system constitutes a truly adaptive dynamic control body; not only can it issue control instructions, but also can immediately know the actual effect brought by the instructions and correct the behavior at the next moment based thereon; this fast closed-loop feedback mechanism enables the system to effectively respond to the rapid changes of the working conditions, continuously maintain the running state of the tractor in the optimal region, and significantly enhances the robustness and environmental adaptability of the system;

[0137] Through the control execution and feedback module, the actual state parameters of the whole machine after the control intervention are fed back to the multi-source heterogeneous data acquisition module, forming a high-speed closed-loop control system; this closed-loop structure ensures that the system can continuously evaluate and decide based on the latest actual state in each control cycle, realizing continuous and dynamic optimization of the running state of the tractor; this capability enables the tractor to maintain a dynamic balance between stability and efficiency at all times when facing the variable and unpredictable working load in the field, improving the working performance and reliability of the whole machine.

[0138] The sensor group includes:

[0139] An engine body acceleration sensor for measuring three-axis vibration acceleration signals of the engine; an engine speed sensor for monitoring the real-time speed of the crankshaft; a power output shaft torque sensor for measuring the load torque of the transmission system; and a vehicle-mounted control unit for obtaining the instruction speed corresponding to the operation instruction and performing data communication through the CAN bus.

[0140] In the embodiment, the specific composition and functions of the sensor group constituting the sensing basis of the system are collectively described.

[0141] The sensor group includes:

[0142] Engine body acceleration sensor, used to monitor the three-axis vibration acceleration signals generated by the engine during operation, to provide core data for analyzing system vibration state and calculating critical frequency band power;

[0143] Engine speed sensor, used to monitor the real-time speed of the engine crankshaft , to provide key information about the engine power output state for the system;

[0144] Power output shaft torque sensor, used to measure the actual load torque borne by the transmission system , to provide direct basis for judging external impact;

[0145] Vehicle-mounted control unit, used to provide the command speed corresponding to the current operation command through its CAN bus interface , to enable the system to obtain operation commands;

[0146] The technical effect of the gain of the embodiment lies in that by explicitly specifying the cooperative combination of the four sensors, a perception system capable of comprehensively capturing the trinity information of tractor state-load-command is constructed; the acceleration, speed and torque sensors describe the objective state of the machine from the physical layer, while the vehicle-mounted control unit provides operation command data; such multi-dimensional information input provides complete and necessary data basis for deep state evaluation and predictive control of the system, and is a prerequisite for realizing advanced intelligent diagnosis and control.

[0147] The dynamic state evaluation module further includes:

[0148] A data receiving unit for receiving multi-dimensional time series data from the multi-source heterogeneous data acquisition module; a data preprocessing unit for performing format standardization and noise filtering on the received data; a risk calculation unit for calculating the instantaneous impact index based on the instantaneous impact index model ; and a result output unit for transmitting the calculated instantaneous impact index to the predictive control strategy generation module;

[0149] In the embodiment, the internal functional unit composition of the dynamic state evaluation module is further refined;

[0150] The module comprises a data receiving unit, a data preprocessing unit, a risk calculation unit and a result output unit in specific implementation. The data receiving unit is responsible for receiving multi-dimensional time series data streams from the multi-source heterogeneous data acquisition module. The data preprocessing unit processes the received original data, including format standardization, unit unification and noise filtering by using a digital filter to eliminate the influence of sensor noise or electromagnetic interference on subsequent calculation. The risk calculation unit is the core of the module, which is based on a preset instantaneous impact index model to fuse and calculate the preprocessed data to obtain a quantitative instantaneous impact index . The result output unit transmits the calculated instantaneous impact index to the predictive control strategy generation module.

[0151] The technical effect brought by the embodiment is to improve the accuracy and stability of risk assessment. The setting of the data preprocessing unit improves the signal-to-noise ratio and consistency of the input data by standardizing and denoising the original data, effectively avoiding the pollution of the risk calculation results caused by burrs and interference in the original signal. This directly improves the reliability of the instantaneous impact index , thereby reducing the possibility of system misjudgment and making the entire diagnosis and control system run more stably and accurately.

[0152] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and not to limit it. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application, which should be covered in the scope of the claims of the present application.

Claims

1. A tractor test data acquisition and fault diagnosis system, characterized by, Comprise: Multi-source heterogeneous data acquisition module, through the deployment of sensor group in the key parts of tractor to capture real-time data of the whole machine working state; Collecting vibration signals measured by engine body acceleration sensors; collecting real-time rotation speed measured by engine rotation speed sensors ; collecting load torque measured by power output shaft torque sensors ; collecting current instruction rotation speed obtained by vehicle-mounted control units ; transmitting collected original data to dynamic state assessment modules in real time Dynamic state evaluation module, receiving multi-dimensional data; fusion analysis of multi-dimensional data based on the instantaneous impact index model; calculating the instantaneous impact index ; a predictive control strategy generation module receives a transient impact index ; calculates an engine control intervention amount based on an adaptive control instruction adjustment model ; a control execution and feedback module configured to calculate a final control command based on the control intervention quantity and a current instruction rotational speed calculate a final control command adjust the engine operating state by executing the final control command through the tractor electronic control unit; collect actual state parameters of the whole machine after execution; and feed back the actual state parameters to the multi-source heterogeneous data collection module; The instantaneous impact index model in the dynamic state evaluation module is a multi-factor weighted prediction model based on first principle analysis of system dynamic response; the calculation of the instantaneous impact index includes: Calculate the rate of change of load torque ; Calculate the rate of change of load torque Rate of change of maximum expected load torque The ratio; calculate the real-time rotational speed. With commanded speed absolute value of deviation ; Calculate the absolute value of the deviation and the upper limit of the speed deviation. The ratio; calculate the critical band energy change rate. ; Calculate the rate of change of energy in the critical frequency band With respect to the maximum expected critical band energy change rate The ratios; multiply the three ratios by their corresponding weights respectively. , , The instantaneous impact index is obtained by summing the results. ,as follows: ； The adaptive control command adjustment model in the predictive control strategy generation module is a nonlinear control law for achieving minimum intervention; the control intervention amount The calculation includes: when the instantaneous jerk index is less than or equal to a warning threshold , the control intervention amount is set to zero; when the instantaneous jerk index is greater than the warning threshold , the difference between the instantaneous jerk index and the warning threshold is calculated; the ratio of the difference to the limit threshold and the difference between the warning threshold is calculated; the ratio is raised to the power of ; and the result is multiplied by the maximum allowed control adjustment to obtain the control intervention amount as follows: ； The determination method of each parameter in the predictive control strategy generation module is as follows: Pre-set maximum allowable control adjustment amount according to tractor engine performance curve ; warning threshold and limit threshold determined by statistical analysis of historical field operation data ; maximum expected load torque rate of change, upper limit of speed deviation, and maximum expected critical band energy rate of change determined by taking the 99th percentile of corresponding physical quantities in historical data ; set weights , , that satisfy normalization constraint conditions ; set control response sensitivity index to a pre-set constant greater than 1.​​​ 2. The tractor testing data acquisition and fault diagnosis system of claim 1, wherein, Dynamic state assessment module for critical band energy rate of change The calculation includes: The time spectrum is obtained by performing a short-time Fourier transform on the original vibration signal. Determine the preset critical frequency band range. Within the critical frequency band The inner time spectrum amplitude squared is obtained by integrating. Critical band power at time For critical band power The rate of change of energy in the critical band is obtained by taking the time derivative. Critical frequency band range The determination was made through modal analysis of the tractor powertrain.

3. The tractor testing data acquisition and fault diagnosis system of claim 1, wherein, The specific composition of the multi-source heterogeneous data acquisition module includes: Engine body acceleration sensor measures three-axis vibration acceleration signals at a preset sampling frequency; engine speed sensor monitors the crankshaft speed and calculates the real-time speed ; power output shaft torque sensor measures the transmission system load torque ; the CAN bus interface of the vehicle control unit obtains the command speed corresponding to the current operation command ; the data synchronization unit synchronizes the above sensor data according to a unified timestamp; the data transmission unit forms a multi-dimensional time sequence data stream and transmits it to the dynamic state evaluation module.

4. The tractor testing data acquisition and fault diagnosis system of claim 1, wherein, The specific composition of the control execution and feedback module includes: The control instruction generating unit generates a control intervention amount The control instruction generating unit generates a control intervention amount The control instruction generating unit generates a control intervention amount The state acquisition unit acquires the actual operation parameters of the engine after the fuel injection amount is adjusted; the feedback transmission unit re-inputs the actual operation parameters to the multi-source heterogeneous data acquisition module; and the closed-loop control unit realizes dynamic evaluation and decision-making based on the latest actual state in each control cycle.

5. The tractor testing data acquisition and fault diagnosis system of claim 1, wherein, The sensor group includes: Engine body acceleration sensor, for measuring engine three-axis vibration acceleration signal; Engine speed sensor, for monitoring the real-time speed of crankshaft; Power output shaft torque sensor, for measuring the load torque of transmission system; Vehicle control unit, for obtaining the instruction speed corresponding to the operation instruction and carrying out data communication through CAN bus.

6. The tractor testing data acquisition and fault diagnosis system of claim 1, wherein, The dynamic state evaluation module further includes: The data receiving unit is used for receiving multi-dimensional time-series data from a multi-source heterogeneous data collection module; the data preprocessing unit is used for standardizing the format and filtering noise of the received data; the risk calculation unit is used for calculating the instantaneous impact index based on an instantaneous impact index model ; and the result output unit is used for transmitting the calculated instantaneous impact index to a predictive control strategy generation module.

Images