Agricultural machine operation fault prediction method based on multi-source data fusion

Abstract

The present application relates to the technical field of agricultural machinery failure, and particularly relates to a kind of agricultural machinery operation failure prediction method based on multi-source data fusion, the method includes appearance detection, current tentative, internal inferiority confirmation, fault distinction, risk quantification, life estimation and threshold self-adaption.The present application calculates apparent abnormality degree by using the linear relationship of outlet pressure and angular velocity, actively adjusts driving current when abnormal and verifies the effect, confirms internal deterioration when adjustment is invalid, avoids misjudgment, according to the physical law that internal leakage must increase oil temperature and blockage must increase vacuum degree, distinguishes sealing ring wear and filter blockage through the combination characteristics of pressure drop rate, oil temperature and vacuum degree, respectively calculates aggravation risk coefficient and evolution risk degree, outputs the remaining number and safety duration of two types of faults, and simultaneously corrects preset abnormal threshold value according to adjustment frequency and angular velocity change, to realize adaptive adjustment of detection reference.

Description

Technical Field

[0001] This invention relates to the field of agricultural machinery failure technology, and in particular to a method for predicting agricultural machinery operation failures based on multi-source data fusion. Background Technology

[0002] With the rapid development of agricultural mechanization towards intelligence, the operation of hydraulic systems in agricultural machinery exhibits complex characteristics such as multi-parameter coupling, highly time-varying operating conditions, and gradual evolution of faults. Traditional diagnostic methods relying on single-variable threshold comparisons or empirical criteria are insufficient to comprehensively characterize the system's health status, effectively separate external load fluctuations from internal performance degradation, resulting in low sensitivity for early anomaly identification, inaccurate fault location, and a lack of quantitative prediction methods for remaining usable life. Against this backdrop, how to integrate multi-source monitoring data to achieve adaptive detection of early hydraulic system degradation, rapid identification of internal and external causes, and quantitative prediction of degradation trends has become a core technical problem urgently needing to be solved to improve the intelligent operation and maintenance level of agricultural machinery.

[0003] Chinese Patent Publication No. CN119982729A discloses a protection and control method, system, storage medium, and tractor for an electric lift. The method includes: acquiring the working mode of the electric lift and the working status of the solenoid valve; determining whether there is a continuous pressure build-up in the hydraulic system of the electric lift based on the working mode of the electric lift and the working status of the solenoid valve; and activating a power-off protection strategy if a continuous pressure build-up is detected in the hydraulic system of the electric lift.

[0004] Therefore, the existing technology has the following problems: the method relies on pattern combination to judge pressure buildup, which is prone to missing early degradation and cannot quantify lifespan; the method relies on preset fixed time thresholds, which is prone to misjudgment or missed judgment due to changes in operating conditions; the method adopts passive detection logic, which is prone to early warning delay and failure to identify performance deviations in the early stage. Summary of the Invention

[0005] To address this, the present invention provides a method for predicting agricultural machinery operation faults based on multi-source data fusion. This method overcomes the problems of low sensitivity and poor prediction accuracy in early fault identification caused by the coupling of external operating condition fluctuations and internal performance degradation, as well as passive static analysis, in the prior art through decoupling diagnosis and dynamic threshold adjustment.

[0006] To achieve the above objectives, this invention provides a method for predicting agricultural machinery operation faults based on multi-source data fusion, comprising: The apparent abnormality is determined based on the apparent abnormality degree and the preset abnormality threshold. The apparent abnormality degree is determined based on the obtained outlet pressure of the agricultural machinery hydraulic pump and the average angular velocity of the lifting arm. Based on the result of the apparent anomaly determination, the effectiveness of the adjustment is determined by the outlet pressure and the average angular velocity after adjusting the drive current of the hydraulic pump drive motor, wherein the adjustment of the drive current is determined based on the apparent anomaly degree. Based on the determination of ineffective adjustment, the hydraulic system is judged to have deteriorated internally according to historical change values. The historical change values ​​are determined based on the pressure drop rate and oil temperature of the hydraulic cylinder. Based on the internal deterioration determination results, the abnormality type is determined to be filter blockage or seal wear according to the deterioration performance value. The deterioration performance value is determined based on the obtained hydraulic pump inlet vacuum, the oil temperature, and the pressure drop rate. The evolution risk level is determined based on the filter blockage, the inlet vacuum, and the pressure drop rate, and the aggravation risk factor is determined based on the seal wear, the pressure drop rate, and the oil temperature. The remaining number of filtration cycles and the safe filtration duration are determined and output based on the filter blockage, the inlet vacuum, and the evolution risk level; the remaining number of sealing cycles and the safe sealing duration are determined and output based on the seal wear, the pressure drop rate, and the aggravated risk coefficient. The preset abnormal threshold is adjusted according to the adjustment frequency of the driving current and the average angular velocity within the preset operating cycle.

[0007] Furthermore, the apparent abnormality is determined based on the apparent abnormality degree and a preset abnormality threshold. The apparent abnormality degree is determined based on the obtained outlet pressure of the agricultural machinery hydraulic pump and the average angular velocity of the lifting arm. This process includes: The pressure-velocity correlation is determined based on the time-series correlation characteristics of the outlet pressure and the average angular velocity; The apparent anomaly degree is determined based on the relative deviation of the pressure-velocity correlation. An apparent abnormality is determined based on the comparison results of the threshold of apparent abnormality.

[0008] Further, the effectiveness of the adjustment is determined based on the outlet pressure and the average angular velocity after adjusting the drive current of the hydraulic pump drive motor. The process of determining the drive current based on the apparent anomaly includes: The drive current is adjusted according to the drive current and the apparent anomaly. The effectiveness of the adjustment is determined by the relative deviations between the adjusted average pressure and the pre-adjustment average pressure, and the relative deviations between the adjusted average speed and the pre-adjustment average speed. The adjusted average pressure is determined based on the outlet pressure after adjusting the drive current. The average pressure before adjustment is determined based on the outlet pressure before adjusting the drive current; The adjusted average speed is determined based on the average angular velocity after adjusting the drive current. The mean velocity before adjustment is determined based on the mean angular velocity before adjusting the drive current.

[0009] Furthermore, the process of determining the effectiveness of the adjustment based on the relative deviations between the adjusted average pressure and the pre-adjustment average pressure, and the relative deviations between the adjusted average speed and the pre-adjustment average speed, includes: The adjustment is deemed invalid based on a threshold comparison of the pressure change rate or velocity change rate. The rate of change of pressure is determined based on the change in the mean pressure after adjustment relative to the mean pressure before adjustment. The rate of change of speed is determined based on the change of the mean speed after adjustment relative to the mean speed before adjustment.

[0010] Furthermore, determining whether the hydraulic system has deteriorated based on historical change values, wherein the process of determining historical change values ​​based on obtaining the pressure drop rate and oil temperature of the hydraulic cylinder includes: The high pressure signal is triggered based on the historical frequency of exceeding the pressure drop rate. Based on the high pressure signal, the internal deterioration of the hydraulic system is determined according to the rising trend of the oil temperature.

[0011] Furthermore, the abnormality type is determined to be filter blockage or seal wear based on the degradation performance value. The process of determining the degradation performance value based on the obtained hydraulic pump inlet vacuum, the oil temperature, and the pressure drop rate includes: Based on the vacuum deviation, oil temperature deviation, and pressure transformer deviation, the abnormality type is determined to be either filter blockage or seal ring wear. Vacuum deviation, which is determined based on the historical average deviation level of the inlet vacuum; Oil temperature deviation, which is determined based on the historical average deviation level of the oil temperature; Pressure deviation is determined based on the historical average deviation level of the pressure drop rate.

[0012] Furthermore, the process of determining the evolution risk level based on the filter blockage, the inlet vacuum, and the pressure drop rate, and determining the aggravated risk factor based on the seal wear, the pressure drop rate, and the oil temperature includes: The evolution risk level is determined based on the vacuum change rate and trend consistency, wherein, The rate of change of vacuum is determined by the average value of the basis vacuum difference; Vacuum differential, which is determined based on the historical instantaneous changes in the inlet vacuum level; The trend consistency is determined based on the vacuum differential and voltage drop differential. The pressure drop differential is determined based on historical instantaneous changes in the pressure drop rate; The risk aggravation coefficient is determined based on the oil temperature difference, the proportion of synchronous changes, and the proximity of heat and oxygen. The oil temperature differential is determined based on the historical instantaneous changes in the oil temperature; The synchronous change ratio is determined based on the synchronous characteristics of the pressure drop difference and the oil temperature difference; The thermal oxygen proximity is determined based on the oil temperature.

[0013] Furthermore, the process of determining and outputting the remaining filtration cycles and filtration safety duration based on the filter clogging, the inlet vacuum, and the evolution risk includes: The remaining number of filtration cycles is determined based on the inlet vacuum level, the smooth change rate, and the evolution risk level, wherein... The smoothed rate of change is determined based on the smoothed rate of change of the inlet vacuum. The safe filtration duration is determined based on the time consumed in a single filtration cycle and the remaining number of filtration cycles.

[0014] Furthermore, the process of determining and outputting the remaining number of seal cycles and the safe sealing time based on the wear of the sealing ring, the pressure drop rate, and the aggravated risk coefficient includes: The remaining number of sealing cycles is determined based on the pressure drop rate, the pressure drop deterioration rate, and the aggravated risk coefficient, wherein... The pressure drop deterioration rate is determined based on the rate of change of the smoothed pressure drop rate between adjacent boost cycles; The sealing safety duration is determined based on the time consumed in a single lifting cycle and the remaining number of seal cycles.

[0015] Furthermore, the process of adjusting the preset abnormal threshold based on the adjustment frequency of the drive current and the average angular velocity within the preset operating cycle includes: The preset abnormal threshold is adjusted based on the comparison results of the adjustment frequency and the periodic change rate thresholds, wherein, The periodic rate of change is determined based on the change in the average angular velocity within the preset working cycle. The frequency adjustment is determined based on the ratio of the number of current adjustments to the preset operating cycle, where, The number of current adjustments is determined based on the number of times the drive current is adjusted within a preset operating cycle.

[0016] Compared with existing technologies, the advantages of this invention lie in calculating the apparent anomaly degree through the linear correlation between outlet pressure and lifting arm angular velocity, actively adjusting the drive current and verifying its effect; if the adjustment is ineffective, internal degradation is confirmed, avoiding misjudgments due to load fluctuations. Based on the increase in oil temperature caused by internal leakage and the increase in vacuum degree caused by filter blockage, the pressure drop rate, oil temperature, and vacuum degree are used to distinguish between seal wear and filter blockage. The risk coefficient for wear is calculated to exacerbate the problem, and the risk degree for blockage evolution is calculated, predicting the remaining number of seals and the safe duration for each. The preset anomaly threshold is dynamically corrected based on the frequency and angular velocity changes of the drive current, making the detection benchmark adaptable to the operating environment, thus solving the problems of low sensitivity and poor prediction accuracy in early fault identification. Attached Figure Description

[0017] Figure 1 This is a flowchart of the agricultural machinery operation fault prediction method based on multi-source data fusion in this embodiment; Figure 2 This is a logic diagram for determining the occurrence of apparent anomalies in this embodiment; Figure 3 This is the logic diagram for determining the high pressure signal in this embodiment; Figure 4 This is the logic diagram for determining whether the abnormality type in this embodiment is filter blockage. Detailed Implementation

[0018] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0019] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0020] Please see Figure 1 This is a flowchart of the agricultural machinery operation fault prediction method based on multi-source data fusion in this embodiment. In this embodiment, the agricultural machinery operation fault prediction method based on multi-source data fusion is applied to the hydraulic lifting system of a tractor. The system includes a drive motor, a hydraulic pump, a filter, a cylinder, a lifting arm, and seals. The drive motor drives the hydraulic pump to draw oil through the filter and pressurize it to the cylinder, pushing the lifting arm to lift the implement. Among them, the outlet pressure is directly related to the angular velocity of the lifting arm, the inlet vacuum reflects the unobstructedness of the filter, the pressure drop rate reflects the degree of seal wear, and the oil temperature characterizes the system's heating state. Filter blockage leads to increased vacuum and decreased angular velocity; seal wear leads to increased internal leakage and increased pressure drop rate, while throttling heat raises the oil temperature. High temperature accelerates oil oxidation and seal aging, forming a vicious cycle of performance degradation.

[0021] This embodiment provides a method for predicting agricultural machinery operation faults based on multi-source data fusion, including: Whether an appearance is abnormal is determined based on the degree of apparent abnormality and a preset abnormality threshold. Apparent anomaly is determined based on the real-time acquisition of the outlet pressure of the agricultural machinery hydraulic pump and the average angular velocity of the lifting arm; The outlet pressure is obtained by installing a pressure sensor, such as a strain gauge or piezoresistive pressure sensor, at the outlet pipeline of the hydraulic pump to reflect the hydraulic pump's ability to deliver oil to the cylinder. The average angular velocity is the average rotation rate of the lifting arm during a single lifting process. It is obtained by measuring and calculating the ratio of the rotation angle to the time taken by installing an angular displacement sensor, such as a rotary encoder or a Hall effect angle sensor, at the shaft of the lifting arm. The preset abnormal threshold is determined based on the inherent fluctuation range of the Pearson correlation coefficient between outlet pressure and average angular velocity under normal operating conditions. It is usually set between 5% and 15%, and in this embodiment it is set to 10%. This value matches the inherent fluctuation range of the correlation coefficient to determine the benchmark value for apparent abnormalities, which can avoid frequent false triggers or delayed warnings. Based on the results of the apparent anomaly determination, the effectiveness of the adjustment is determined by the outlet pressure and average angular velocity after adjusting the drive current of the hydraulic pump drive motor. The adjustment of the drive current is determined based on the degree of apparent anomaly. Based on the determination of ineffective adjustment, the hydraulic system is judged to have deteriorated internally based on historical change values. The historical change values ​​are determined based on the pressure drop rate and oil temperature of the hydraulic cylinders acquired in real time. Oil temperature is obtained by measuring the temperature of the hydraulic oil in real time by installing temperature sensors such as thermocouples or thermistors inside the hydraulic oil tank or in the return oil line, so as to reflect the thermal state of the system. The pressure drop rate is calculated by continuously collecting pressure signals using a pressure sensor installed in the rodless chamber of the hydraulic cylinder, reflecting the pressure drop per unit time during the pressure holding phase of the hydraulic cylinder. Based on the internal deterioration judgment results, the abnormality type is determined to be filter blockage or seal wear according to the deterioration performance value. The deterioration performance value is determined based on the real-time acquisition of the inlet vacuum, oil temperature and pressure drop rate of the agricultural machinery hydraulic pump. The inlet vacuum is measured by a vacuum pressure sensor installed between the hydraulic pump suction port and the filter, which measures the pressure difference below atmospheric pressure to indicate the degree of filter blockage. The evolution risk level is determined based on the filter blockage, the inlet vacuum, and the pressure drop rate; and the aggravation risk factor is determined based on the seal wear, the pressure drop rate, and the oil temperature. The remaining number of filtration cycles and the safe filtration duration are determined and output based on filter clogging, basic filtration cycles, and evolution risk level. Similarly, the remaining number of sealing cycles and the safe filtration duration are determined and output based on seal wear, basic sealing cycles, and aggravation risk coefficient. The number of filtration cycles is determined based on the inlet vacuum level; The number of sealing cycles is determined based on the pressure drop deterioration rate; The preset abnormal threshold is adjusted based on the adjustment frequency of the drive current and the average angular velocity within the preset operating cycle. The preset operating cycle is determined based on the probability distribution of the number of times the drive current is actively adjusted within 100 consecutive lifting cycles of five tractors of the same model during normal operation. It is usually set between 50 and 200 lifting cycles. In this embodiment, it is set to 100 lifting cycles. This value matches the median of the adjustment frequency distribution, which can avoid misjudgment or response lag due to insufficient samples.

[0022] Apparent anomaly is calculated by using the linear correlation between outlet pressure and lifting arm angular velocity. The drive current is actively adjusted and its effectiveness verified; if the adjustment is ineffective, internal degradation is confirmed, avoiding misjudgments due to load fluctuations. Based on the principle that internal leakage causes oil temperature rise and filter blockage causes vacuum rise, pressure drop rate, oil temperature, and vacuum are used to differentiate between seal wear and filter blockage. The risk factor for wear is calculated to exacerbate the problem, and the risk factor for blockage evolution is calculated, predicting the remaining number of seal cycles and safe operating time for both. The preset anomaly threshold is dynamically corrected based on the drive current adjustment frequency and angular velocity changes, making the detection benchmark adaptable to the operating environment and solving the problems of low sensitivity and poor prediction accuracy in early fault identification.

[0023] Please see Figure 2 This is the logic diagram for determining the appearance abnormality in this embodiment. In this embodiment, whether the appearance is abnormal is determined based on the appearance abnormality degree and a preset abnormality threshold. The appearance abnormality degree is determined based on the real-time acquired outlet pressure of the agricultural machinery hydraulic pump and the average angular velocity of the lifting arm. The Pearson correlation coefficients of outlet pressure and mean angular velocity at each moment within the preset identification time period are calculated to obtain the pressure-velocity correlation, which characterizes the degree of linear correlation between outlet pressure and mean angular velocity. The preset identification time is determined based on the statistical sample size required to calculate the time of a single boosting cycle and the correlation coefficient. It is usually set between 20 and 40 seconds. In this embodiment, it is set to 30 seconds. This value matches the total time of ten consecutive boosting cycles, which can ensure statistical significance and detect anomalies early. The absolute value of the relative deviation between the pressure-velocity correlation and a preset correlation threshold is calculated to obtain the apparent anomaly degree, which characterizes the degree of deviation from the linear correlation between outlet pressure and mean angular velocity. A preset relevant threshold is determined based on the average correlation coefficient of 500 cycles under normal operating conditions. It is usually set between 0.85 and 0.95. In this embodiment, it is set to 0.90. This value matches the 95% lower confidence limit of the average correlation coefficient under normal operating conditions, which can achieve a balance between false alarm suppression and early warning. When the apparent abnormality exceeds a preset abnormality threshold, an apparent abnormality is determined to have occurred.

[0024] Under normal operating conditions, the outlet pressure and the average angular velocity of the lifting arm maintain a linear correlation. With load changes, both increase and decrease proportionally, and the Pearson correlation coefficient is close to 1. When the energy transfer efficiency between drive and response begins to decline, the correlation between pressure and angular velocity is disrupted, and the correlation coefficient decreases accordingly. Using the absolute value of the relative deviation between this correlation coefficient and a preset correlation threshold as the apparent anomaly level allows for the detection of early abnormalities in the decoupling of pressure and angular velocity before they exceed conventional alarm limits.

[0025] Specifically, the effectiveness of the adjustment is determined by the outlet pressure and average angular velocity after adjusting the drive current of the hydraulic pump drive motor. The process of determining the drive current based on apparent anomalies includes: The driving current is increased proportionally to the apparent anomaly; the greater the apparent anomaly, the greater the adjustment range of the driving current. The effectiveness of the adjustment is determined by the relative deviations between the adjusted average pressure and the pre-adjustment average pressure, and the relative deviations between the adjusted average speed and the pre-adjustment average speed. The adjusted average pressure is determined based on the average outlet pressure over a preset first number of boost cycles after adjusting the drive current, to characterize the average recovery level of the outlet pressure after drive current adjustment. The preset first number is determined based on the dynamic response cycle required for the hydraulic system to stabilize from a current step to an angular velocity, as well as the effect of the moving average on suppressing random fluctuations. It is usually set between 3 and 7 times. In this embodiment, it is set to 5 times. This value matches the minimum number of cycles required to stabilize from a current step to a new angular velocity steady state, and can balance noise suppression and response delay. The average pressure before adjustment is determined based on the average value of the outlet pressure during the preset first number of boost cycles before adjusting the drive current, to characterize the reference level of the outlet pressure before the drive current is adjusted. The average speed after adjustment is determined based on the average angular velocity of the average angular velocity of the lifting arm after adjusting the drive current for a preset first number of lifting cycles, so as to characterize the average lifting level of the average angular velocity of the lifting arm after the drive current is adjusted. The average speed before adjustment is determined based on the average angular velocity of the average angular velocity of the lifting arm in a preset first number of lifting cycles before adjusting the drive current, so as to characterize the reference level of the average angular velocity of the lifting arm before the drive current is adjusted.

[0026] Specifically, the process of determining the effectiveness of the adjustment based on the relative deviations between the adjusted average pressure and the pre-adjustment average pressure, and the relative deviations between the adjusted average speed and the pre-adjustment average speed, includes: When the rate of pressure change is less than a preset pressure-transformer threshold, or the rate of speed change is less than a preset speed-transformer threshold, it is determined that the increment of the drive current has failed to be effectively converted into a lifting response, thus obtaining a determination that the adjustment is ineffective. The pressure change rate is determined based on the relative deviation between the average pressure after adjustment and the average pressure before adjustment, in order to characterize the degree of pressure recovery after the drive current is adjusted. The preset pressure change threshold is determined based on the statistical distribution of the relative deviation of the average pressure of the circulation outlet during three consecutive increases in operation. It is usually set between 3% and 8%, and in this embodiment it is set to 5%. This value matches the 99% confidence upper limit of the relative deviation of the average pressure, which can eliminate random fluctuations and capture early blockages. The rate of change of speed is determined based on the relative deviation between the mean speed after adjustment and the mean speed before adjustment, in order to characterize the degree of speed improvement after the drive current is adjusted. The preset speed threshold is determined based on the statistical distribution of the coefficient of variation of the average angular velocity of three consecutive acceleration cycles during operation. It is usually set between 2% and 5%, and in this embodiment it is set to 3%. This value matches the 99% confidence limit of the coefficient of variation of the average angular velocity, which can filter load noise and identify the real speed increase. Otherwise, it is determined that the increment of the drive current has been effectively converted into a lifting response, so as to obtain a determination result that the adjustment is effective.

[0027] By actively increasing the drive current, the oil supply flow and pressure of the hydraulic pump are directly increased. The fault type can be distinguished by using the current increment to test the pressure energy conversion efficiency. When the outlet pressure and lifting arm angular velocity rise significantly and synchronously after increasing the current, it indicates that the current increment has been completely transmitted to the lifting arm through the hydraulic pump and cylinder, and the energy transmission chain is uninterrupted. This indicates that the abnormality originates from increased external oil supply resistance. When the pressure rises but the angular velocity increases only slightly, it indicates that the pressure energy has been dissipated through the sealing gap in the cylinder and cannot drive the piston. In this case, adjusting the drive current is ineffective, and further analysis is needed to determine if there is an internal abnormality.

[0028] Please see Figure 3 This is the logic diagram for determining the triggering of the high pressure signal in this embodiment. In this embodiment, the determination of whether the hydraulic system has deteriorated is based on historical change values. The process of determining the historical change values ​​based on the real-time acquisition of the hydraulic cylinder pressure drop rate and oil temperature includes: When the number of times the pressure drop exceeds the preset threshold, a high pressure signal is triggered. The number of times the pressure drop exceeds the limit is determined based on the frequency of the pressure drop rate being greater than a preset drop threshold in the preset first number of boost cycles prior to the current moment, in order to characterize the cumulative degree of pressure drop rate exceeding the limit in multiple cycles; The preset number threshold is determined based on the probability distribution of the pressure drop rate during a preset first number of boosting cycles under the condition of no internal leakage. It is usually set to between 60% and 80% of the preset first number. In this embodiment, it is set to 4 times. This value corresponds to a statistical confidence level of 99.4% for at least four exceedances in five boosting cycles, which can eliminate single accidental interference and capture continuous leakage. The preset pressure drop threshold is determined based on the historical average of the pressure drop rate under no internal leakage conditions plus three times the standard deviation. It can control the fluctuation probability within 99.7%. It is usually set between 0.02 MPa per second and 0.08 MPa per second. In this embodiment, it is set to 0.05 MPa per second. This value corresponds to the statistical confidence limit of 99.7%, which can control the false judgment probability within 0.3%. Based on the aforementioned high pressure signal, when the historical temperature change rate is positive, it is determined that the hydraulic system is internally deteriorating. The historical temperature change rate is determined based on the slope of the linear trend of the oil temperature during the first preset number of increase cycles prior to the current moment, to characterize the monotonically increasing trend of the oil temperature in the cycle sequence.

[0029] By statistically analyzing the frequency of pressure drop rate exceeding limits in multiple increase cycles and examining whether the slope of the linear trend of oil temperature with the cycle sequence is positive, the internal degradation judgment can be transformed into a dual-condition verification of repeated pressure exceeding limits and monotonous oil temperature rise after actively increasing the drive current proves ineffective. Internal leakage causes repeated pressure exceeding limits, while single, occasional fluctuations are unlikely to be sustained. The heat generated by leakage accumulates through oil backflow, inevitably causing a monotonous rise in oil temperature in the tank. These two factors are independent and time-matched. The joint judgment avoids misjudgment caused by heat transfer delays and prevents missed judgments due to single-parameter threshold interference from operating condition fluctuations, enabling high-confidence confirmation of internal performance degradation in the hydraulic system.

[0030] Please see Figure 4 This is the logic diagram for determining whether the abnormality type is filter blockage in this embodiment. In this embodiment, the abnormality type is determined to be filter blockage or seal wear based on the deterioration performance value. The process of determining the deterioration performance value based on the real-time acquired inlet vacuum of the agricultural machinery hydraulic pump, the oil temperature, and the pressure drop rate includes: When the vacuum deviation exceeds a preset deviation threshold, and is simultaneously greater than both the oil temperature deviation and the pressure transformer deviation, the abnormality type is determined to be filter blockage. Conversely, when both the oil temperature deviation and the pressure transformer deviation exceed preset deviation thresholds, and are simultaneously greater than the vacuum deviation, the abnormality type is determined to be seal wear. The preset deviation threshold is determined based on the statistical fluctuation range of the relative deviation of each parameter under fault-free conditions. It is usually set between 5% and 15%, and in this embodiment it is set to 10%. This value matches the 99% confidence limit of the relative deviation under fault-free conditions and retains a safety margin. It can eliminate random noise and ensure that the real deviation participates in the judgment. Vacuum deviation is determined based on the relative deviation between the average value of the inlet vacuum degree in the preset first number of lifting cycles prior to the current moment and a preset vacuum threshold, to characterize the degree to which the inlet vacuum degree deviates from the threshold. The preset vacuum threshold is determined based on the statistical distribution of the inlet vacuum degree of the hydraulic system under the condition of no filter blockage. It is usually set between 0.005 MPa and 0.015 MPa. In this embodiment, it is set to 0.01 MPa. This value matches the 99.7 percent confidence upper limit of the vacuum degree when there is no blockage, which can control the probability of normal fluctuations being misjudged as blockage. The oil temperature deviation is determined based on the relative deviation between the average oil temperature in the preset first number of boost cycles prior to the current moment and a preset oil temperature threshold, to characterize the degree of oil temperature deviation from the threshold. The preset oil temperature threshold is determined based on the statistical distribution of oil temperature in the hydraulic system under no internal leakage conditions. It is usually set between 40°C and 60°C. In this embodiment, it is set to 50°C. This value matches the 99.7% confidence upper limit of oil temperature under no leakage conditions, which can eliminate the interference of ambient temperature changes and identify the temperature rise caused by internal leakage. The pressure drop deviation is determined based on the relative deviation between the average pressure drop rate in the preset first number of boost cycles prior to the current moment and the preset drop threshold, in order to characterize the degree to which the pressure drop rate deviates from the threshold. Otherwise, the current anomaly type is determined to be too complex, and a shutdown check is required.

[0031] Based on the structural characteristics of the hydraulic system, namely that the filter is installed on the suction side of the hydraulic pump, when it is blocked, the vacuum at the pump inlet increases independently while the cylinder seal is not damaged, so the pressure drop rate and oil temperature will not rise synchronously; the sealing ring is installed between the piston and the cylinder wall, and after wear, high-pressure oil leaks from the gap, which increases the pressure drop rate. At the same time, the leaked oil throttling and heating causes the oil temperature to rise synchronously. It is possible to distinguish the abnormal type by comparing the relative magnitudes of vacuum deviation, oil temperature deviation and pressure deviation, thus avoiding the conflict and calibration difficulties of multi-parameter independent threshold judgment.

[0032] Specifically, the process of determining the evolution risk level based on the filter blockage, the inlet vacuum, and the pressure drop rate, and determining the aggravated risk factor based on the seal wear, the pressure drop rate, and the oil temperature includes: The product of the vacuum change rate and the trend consistency is calculated as a ratio to a preset vacuum change threshold to obtain the evolution risk level, which characterizes the rate of filter clogging progression. A preset vacuum change threshold is determined based on the statistical upper limit of the random fluctuation of the vacuum rise rate under unblocked conditions. Its value is usually set between 0.001 MPa per cycle and 0.003 MPa per cycle. In this embodiment, it is set to 0.002 MPa per cycle. This value corresponds to the 99.7% confidence limit of normal fluctuations and can distinguish between real deterioration and random jitter. If the value is too low, it is easy to make false alarms, and if it is too high, it is easy to miss the judgment. The vacuum change rate is determined based on the average of the first-order forward difference between the inlet vacuum degree of each of the preset second number of lifting cycles prior to the current moment and the inlet vacuum degree of the adjacent previous cycle, to characterize the average rate of increase of the inlet vacuum degree. The second preset quantity is determined based on the minimum number of cycles required for parameter changes to stabilize during the fault evolution process. Its value is usually set between 6 and 12 times. In this embodiment, it is set to 10 times. This value matches the average number of cycles required for the vacuum degree to slowly rise to stable deterioration. It can capture trend changes and avoid excessively long windows to smooth out early anomalies. The trend consistency is determined by the ratio of the number of positive pressure drop differential values ​​to the number of vacuum values ​​in a preset number of boost cycles after the vacuum change increases. This ratio characterizes the rate of pressure drop following the increase in vacuum. Vacuum differential is determined based on the first-order forward difference between the inlet vacuum degree of each boost cycle and the inlet vacuum degree of the adjacent previous cycle in the second number of boost cycles before the current time, so as to characterize the instantaneous change trend of vacuum degree. The pressure drop differential is determined based on the first-order forward difference between the inlet vacuum degree of each boost cycle and the inlet vacuum degree of the adjacent previous cycle in the preset second number of boost cycles before the current moment, so as to characterize the instantaneous change trend of vacuum degree. The vacuum quantity is determined based on the number of lifting cycles with positive vacuum differences in the preset second number of lifting cycles prior to the current moment; The risk aggravation coefficient is determined by calculating the ratio of the average value of all oil temperature differences to the preset temperature change threshold, the synchronous change ratio, and the product of the thermal-oxygen proximity, to characterize the tendency of accelerated seal wear. The oil temperature difference is determined based on the first-order forward difference of the oil temperature in each of the second number of boost cycles before the current moment, and the oil temperature in the previous cycle, to characterize the direction of successive fluctuations in oil temperature. The preset temperature change threshold is determined based on the statistical upper limit of the random fluctuation of the oil temperature change rate under the condition of no internal leakage. Its value is usually set between 0.1 degrees Celsius per cycle and 0.2 degrees Celsius per cycle. In this embodiment, it is set to 0.15 degrees Celsius per cycle. This value corresponds to the 99.7% confidence upper limit of normal fluctuation and can eliminate the interference of environmental temperature drift. The synchronous change ratio is based on the ratio of the number of boost cycles in which the pressure drop difference and oil temperature difference are both positive within the same boost cycle to a preset second number, in order to characterize the trend of the pressure drop rate and oil temperature rising synchronously. Thermo-oxidative proximity, determined based on the current oil temperature's relative position within a preset temperature range, characterizes the degree to which the oil temperature approaches the oxidation acceleration zone. The preset temperature range has a lower limit corresponding to the starting temperature at which the oil begins to accelerate oxidation, and an upper limit corresponding to the oxidation rate saturation temperature. The value is usually set between 30°C and 80°C. In this embodiment, it is set to 40°C to 60°C. This range is consistent with the inflection point of the hydraulic oil oxidation rate and can quantify the aggravating effect of oil temperature on blockage within a linear range.

[0033] In this embodiment, both the evolution risk degree and the aggravation risk coefficient are limited to a value range not exceeding 1 to avoid numerical overflow that could lead to inaccurate subsequent judgments.

[0034] By quantifying the evolution risk of wear induced by blockage, the risk value only increases when the blockage worsens and causes seal damage. By quantifying the risk of wear exacerbating blockage into the product of three factors—the rate of oil temperature rise, the synchronicity of oil temperature and pressure drop rate, and the range of oil temperature—the risk value only becomes significant when wear heat reaches the accelerated oxidation temperature and rises synchronously with the pressure drop rate. At the same time, by compressing the two risk coefficients into a unit range, the cross-risk warning can be both sensitive and reliable.

[0035] Specifically, the remaining number of filtration cycles and the safe filtration duration are determined and output based on filter clogging, the baseline number of filtration cycles, and the risk of filtration progression. The process of determining the baseline number of filtration cycles based on the inlet vacuum includes: The baseline number of filtration cycles is multiplicatively reduced based on the evolution risk level, ensuring the reduced value is not less than half of the baseline number of filtration cycles. This yields the remaining filtration cycles, representing the actual number of remaining cycles after considering clogging-induced wear. The number of filtration cycles is determined by the ratio of the difference between the current inlet vacuum level and a preset vacuum threshold to the smoothed rate of change, thus representing the number of available cycles extrapolated at the current deterioration rate. The smoothing rate of change is determined based on the average of the first-order forward differences of the smoothed vacuum degree between each boost cycle and its preceding adjacent boost cycle in a preset second number of boost cycles. This smoothing rate of change characterizes the long-term rate of vacuum degree deterioration. The smoothed vacuum degree is determined by an exponentially weighted moving average of the inlet vacuum degree over a preset smoothing coefficient and a preset second number of lifting cycles, to characterize the long-term trend of the smoothed inlet vacuum degree. The preset smoothing coefficient is determined based on the time constant when the autocorrelation coefficient of random fluctuations between adjacent cycles in the inlet vacuum degree sequence or pressure drop rate sequence first decays to below 0.5. It is usually set between 0.2 and 0.4. In this embodiment, it is set to 0.3. This value matches the autocorrelation decay period of short-term fluctuations in the sequence and can take into account both noise suppression and trend response. The product of the remaining number of filtration cycles and the average time consumed is calculated to obtain the safe filtration duration, which characterizes the safe filtration replacement interval in units of time. The average time is determined based on the average time taken to obtain a single boost cycle.

[0036] Specifically, the remaining number of seal cycles and the safe sealing time are determined and output based on seal wear, the number of seal cycles, and the risk aggravation factor. The process of determining the number of seal cycles based on the pressure drop deterioration rate includes: The number of cycles required for sealing is multiplied and reduced based on an aggravated risk factor, ensuring the reduced number is not less than half of the original number of cycles. This yields the remaining number of cycles, representing the actual remaining number of cycles after considering accelerated oxidation due to internal leakage. The number of sealing cycles is determined based on the ratio of the difference between the current pressure drop degradation rate and a preset pressure drop threshold to the pressure drop degradation rate. The pressure drop deterioration rate is determined based on the average of the first-order forward differences of the smoothed pressure drop of each boost cycle and its preceding adjacent boost cycle in a preset second number of boost cycles, to characterize the long-term deterioration rate of the pressure drop. The pressure drop smoothing method is determined by an exponentially weighted moving average of the pressure drop rate over the preset second number of boost cycles based on a preset smoothing coefficient, in order to characterize the long-term trend of the pressure drop rate after smoothing. The product of the remaining number of seal replacements and the average time consumed is calculated to obtain the seal safety duration, which characterizes the seal safety replacement interval in time units.

[0037] Exponentially weighted moving average remaining life prediction focuses more on long-term trends. When the filter becomes clogged, the inlet vacuum gradually rises to the pump inlet ultimate vacuum; when the seal wears, the pressure drop rate gradually rises to the severe leakage threshold. The difference between the current parameter value and the physical limit is the remaining degradation space. The smoothed deterioration rate reflects the average increment per cycle increase. Dividing the two yields the number of usable cycles linearly extrapolated according to the current deterioration rate, which serves as the base remaining life. Based on this, an evolution risk factor or aggravation risk coefficient is introduced for multiplicative reduction. This ensures that the positive feedback effect of cloggedness inducing seal wear or wear accelerating oil oxidation is reflected in the prediction, while limiting the reduction to no more than half to avoid unnecessary downtime due to overestimation of cross-risk leading to an underestimated prediction value.

[0038] The process of adjusting the preset abnormal threshold based on the adjustment frequency of the drive current and the average angular velocity within the preset operating cycle includes: When the adjustment frequency is less than a preset frequency threshold and the periodic change rate is less than a preset change threshold, the preset abnormal threshold is decreased according to a preset adjustment step size. The preset frequency threshold is determined based on the statistical distribution of the natural probability of the hydraulic lifting system actively adjusting the drive current under the condition of no internal deterioration and no filter blockage. It is usually set between 2% and 5%, and in this embodiment it is set to 3%. This value corresponds to the frequency of three active adjustments within 100 cycles, which can identify threshold sluggishness under the premise of accumulating sufficient samples. The preset change threshold is determined based on the statistical upper limit of the drift rate of angular velocity caused by random load fluctuations under the condition of no internal degradation and no filter blockage. It is usually set between -0.01 radians per second per cycle and -0.005 radians per second per cycle. In this embodiment, it is set to -0.008 radians per second per cycle. This value corresponds to the 99.7 percent confidence lower limit of the angular velocity drift rate, which can distinguish between random fluctuations and continuous deterioration. The preset adjustment step size is determined based on the balance between the convergence speed and stability of the threshold adaptive adjustment. It is usually set between 5% and 15%. In this embodiment, it is set to 10%. This value matches the engineering experience of completing the full range correction in ten cycles and can make the threshold converge within three to five cycles. The rate of change of the cycle is determined by removing the average angular velocity value of the first lifting cycle adjacent to each adjustment of the drive current within the preset working cycle to eliminate transient interference, and arranging the remaining average angular velocities in chronological order to calculate the rate of change of the average angular velocity at the beginning and end of the sequence, so as to characterize the trend of continuous deterioration of angular velocity within the cycle. The frequency adjustment is determined based on the ratio of the number of current adjustments to the preset operating cycle, where, The number of current adjustments is determined based on the number of times the drive current is adjusted within a preset operating cycle.

[0039] When the adjustment frequency is too low and the angular velocity continues to deteriorate simultaneously, it indicates that the current preset anomaly threshold is higher than the actual anomaly trigger point, causing the apparent anomalies that should have been captured to be ignored for a long time. Actively reducing the threshold can realign the detection boundary with the actual degradation rate, and can restore the sensitive response to early anomalies without adding ineffective adjustments.

[0040] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for predicting agricultural machinery operation faults based on multi-source data fusion, characterized in that, include: The apparent abnormality is determined based on the apparent abnormality degree and the preset abnormality threshold. The apparent abnormality degree is determined based on the obtained outlet pressure of the agricultural machinery hydraulic pump and the average angular velocity of the lifting arm. Based on the result of the apparent anomaly determination, the effectiveness of the adjustment is determined by the outlet pressure and the average angular velocity after adjusting the drive current of the hydraulic pump drive motor, wherein the adjustment of the drive current is determined based on the apparent anomaly degree. Based on the determination of ineffective adjustment, the hydraulic system is judged to have deteriorated internally according to historical change values. The historical change values ​​are determined based on the pressure drop rate and oil temperature of the hydraulic cylinder. Based on the internal deterioration determination results, the abnormality type is determined to be filter blockage or seal wear according to the deterioration performance value. The deterioration performance value is determined based on the obtained hydraulic pump inlet vacuum, the oil temperature, and the pressure drop rate. The evolution risk level is determined based on the filter blockage, the inlet vacuum, and the pressure drop rate, and the aggravation risk factor is determined based on the seal wear, the pressure drop rate, and the oil temperature. The remaining number of filtration cycles and the safe filtration duration are determined and output based on the filter blockage, the inlet vacuum, and the evolution risk level; the remaining number of sealing cycles and the safe sealing duration are determined and output based on the seal wear, the pressure drop rate, and the aggravated risk coefficient. The preset abnormal threshold is adjusted according to the adjustment frequency of the driving current and the average angular velocity within the preset operating cycle. The determination of whether an appearance is abnormal is based on the apparent anomaly degree and a preset anomaly threshold. The apparent anomaly degree is determined based on the obtained outlet pressure of the agricultural machinery hydraulic pump and the average angular velocity of the lifting arm. The process includes: The pressure-velocity correlation is determined based on the time-series correlation characteristics of the outlet pressure and the average angular velocity; The apparent anomaly degree is determined based on the relative deviation of the pressure-velocity correlation. An apparent abnormality is determined based on the comparison result of the threshold of apparent abnormality. The process of determining and outputting the remaining filtration cycles and filtration safety time based on the filter clogging, the inlet vacuum, and the evolution risk level includes: The remaining number of filtration cycles is determined based on the inlet vacuum level, the smooth change rate, and the evolution risk level, wherein... The smoothed rate of change is determined based on the smoothed rate of change of the inlet vacuum. The safe filtration duration is determined based on the time consumed in a single filtration cycle and the remaining number of filtration cycles.

2. The method for predicting agricultural machinery operation faults based on multi-source data fusion according to claim 1, characterized in that, The effectiveness of the adjustment is determined based on the outlet pressure and the average angular velocity after adjusting the drive current of the hydraulic pump drive motor. The process of determining the drive current based on the apparent anomaly includes: The drive current is adjusted according to the drive current and the apparent anomaly. The effectiveness of the adjustment is determined by the relative deviations between the adjusted average pressure and the pre-adjustment average pressure, and the relative deviations between the adjusted average speed and the pre-adjustment average speed. The adjusted average pressure is determined based on the outlet pressure after adjusting the drive current. The average pressure before adjustment is determined based on the outlet pressure before adjusting the drive current; The adjusted average speed is determined based on the average angular velocity after adjusting the drive current. The mean velocity before adjustment is determined based on the mean angular velocity before adjusting the drive current.

3. The method for predicting agricultural machinery operation faults based on multi-source data fusion according to claim 2, characterized in that, The process of determining the effectiveness of an adjustment based on the relative deviations between the adjusted average pressure and the original average pressure, and the relative deviations between the adjusted average speed and the original average speed, includes: The adjustment is deemed invalid based on a threshold comparison of the pressure change rate or velocity change rate. The rate of change of pressure is determined based on the change in the mean pressure after adjustment relative to the mean pressure before adjustment. The rate of change of speed is determined based on the change of the mean speed after adjustment relative to the mean speed before adjustment.

4. The method for predicting agricultural machinery operation faults based on multi-source data fusion according to claim 3, characterized in that, Determining whether a hydraulic system has deteriorated based on historical change values ​​involves a process that uses the acquisition of hydraulic cylinder pressure drop rate and oil temperature data. The high pressure signal is triggered based on the historical frequency of exceeding the pressure drop rate. Based on the high pressure signal, the internal deterioration of the hydraulic system is determined according to the rising trend of the oil temperature.

5. The method for predicting agricultural machinery operation faults based on multi-source data fusion according to claim 4, characterized in that, The abnormality type is determined to be filter blockage or seal wear based on the degradation performance value. The process of determining the degradation performance value based on the obtained hydraulic pump inlet vacuum, oil temperature, and pressure drop rate includes: Based on the vacuum deviation, oil temperature deviation, and pressure transformer deviation, the abnormality type is determined to be either filter blockage or seal ring wear. Vacuum deviation, which is determined based on the historical average deviation level of the inlet vacuum; Oil temperature deviation, which is determined based on the historical average deviation level of the oil temperature; Pressure deviation is determined based on the historical average deviation level of the pressure drop rate.

6. The method for predicting agricultural machinery operation faults based on multi-source data fusion according to claim 5, characterized in that, The process of determining the evolution risk level based on the filter blockage, the inlet vacuum, and the pressure drop rate, and determining the aggravated risk factor based on the seal wear, the pressure drop rate, and the oil temperature includes: The evolution risk level is determined based on the vacuum change rate and trend consistency, wherein, The rate of change of vacuum is determined based on the average value of the vacuum difference; Vacuum differential, which is determined based on the historical instantaneous changes in the inlet vacuum level; The trend consistency is determined based on the vacuum differential and voltage drop differential. The pressure drop differential is determined based on historical instantaneous changes in the pressure drop rate; The risk aggravation coefficient is determined based on the oil temperature difference, the proportion of synchronous changes, and the proximity of heat and oxygen. The oil temperature differential is determined based on the historical instantaneous changes in the oil temperature; The synchronous change ratio is determined based on the synchronous characteristics of the pressure drop difference and the oil temperature difference; The thermal oxygen proximity is determined based on the oil temperature.

7. The method for predicting agricultural machinery operation faults based on multi-source data fusion according to claim 6, characterized in that, The process of determining and outputting the remaining number of seal cycles and the seal safety time based on the wear of the sealing ring, the pressure drop rate, and the risk aggravation factor includes: The remaining number of sealing cycles is determined based on the pressure drop rate, the pressure drop deterioration rate, and the aggravated risk coefficient, wherein... The pressure drop deterioration rate is determined based on the rate of change of the smoothed pressure drop rate between adjacent boost cycles; The sealing safety duration is determined based on the time consumed in a single lifting cycle and the remaining number of seal cycles.

8. The method for predicting agricultural machinery operation faults based on multi-source data fusion according to claim 7, characterized in that, The process of adjusting the preset abnormal threshold based on the adjustment frequency of the drive current and the average angular velocity within the preset operating cycle includes: The preset abnormal threshold is adjusted based on the comparison results of the adjustment frequency and the periodic change rate thresholds, wherein, The periodic rate of change is determined based on the change in the average angular velocity within the preset working cycle. The frequency adjustment is determined based on the ratio of the number of current adjustments to the preset operating cycle, where, The number of current adjustments is determined based on the number of times the drive current is adjusted within a preset operating cycle.

Images