Methods and Systems for Monitoring and Maintaining Tobacco Agricultural Machinery Throughout its Life Cycle

Abstract

This invention relates to the field of industrial data processing technology, specifically to a method and system for monitoring the entire lifecycle of tobacco agricultural machinery. The method includes the following steps: collecting acceleration, simulated voltage, and rotational speed pulses and calculating vibration amplitude and rotational speed; aligning mileage by time to generate operational data; truncating data at fixed mileage intervals to form interval characteristics and calculating density peaks; converting topology data elements in high-load intervals to generate nonlinear paths; identifying abrupt changes through window counting and density gradient difference; truncating to form operational intervals and outputting maintenance warning commands. In this invention, traditional operation and maintenance monitoring focuses on operational time, mileage, fuel consumption, and maintenance record archiving. Data is mostly periodically entered and statically classified, making it difficult to correlate load fluctuations, vibration anomalies, and rotational speed changes. Maintenance judgments rely on maintenance cycles and fault statistics, easily overlooking fatigue accumulation in high-intensity intervals, resulting in delayed warnings, coarse maintenance criteria, and insufficient identification of deterioration trends.

Description

Technical Field

[0001] This invention relates to the field of industrial data processing technology, and in particular to a method and system for monitoring the entire life cycle of tobacco agricultural machinery. Background Technology

[0002] Industrial data processing technology refers to a technical system that focuses on industrial equipment, production objects, and their operation processes. It collects, records, stores, and analyzes operational data, status data, and maintenance data to manage and schedule information throughout the entire equipment process. Its core aspects include continuous acquisition of equipment operating parameters, structured storage of historical data, periodic recording of equipment status, archiving and management of maintenance records, and time-series-based data comparison and statistical processing. Overall, it involves the organization, processing flow, and application models of data generated during the use, maintenance, and management of industrial equipment, and is widely used in the operation and management scenarios of manufacturing equipment, engineering machinery, and various types of electromechanical equipment.

[0003] The traditional tobacco agricultural machinery full life cycle operation and maintenance monitoring method and system refers to the whole process management of tobacco agricultural machinery from procurement, commissioning, daily operation, maintenance and scrapping stages. It usually involves setting up speed sensors, fuel consumption metering devices and positioning terminals on agricultural machinery equipment to collect data on operating time, mileage and fuel consumption. Maintenance time, types of replaced parts and repair content are manually recorded and entered into spreadsheets or relational databases on a local computer. The equipment usage records and maintenance records are then classified and organized according to a preset cycle. By comparing the equipment running time with the established maintenance cycle table, it is determined whether repair or parts replacement is required. At the same time, the number of equipment failures and repair records are statistically archived to complete the operation and maintenance monitoring and management of tobacco agricultural machinery from commissioning to retirement.

[0004] Traditional operation and maintenance monitoring focuses on the archiving of operation time, mileage, fuel consumption and manual maintenance records. The data sources are mainly periodic entry and static classification. In actual operation, load fluctuations, abnormal vibrations and speed changes are difficult to form a continuous correlation. Maintenance judgment mainly relies on the established maintenance cycle and the number of failures. It is easy to ignore the fatigue accumulation in the high-intensity operation range, resulting in delayed early warning, rough maintenance basis, and insufficient identification of component deterioration trend, which affects the accuracy of operation and maintenance management of tobacco agricultural machinery throughout the entire life cycle. Summary of the Invention

[0005] To address the technical problems existing in the prior art, this invention provides a method for monitoring and maintaining tobacco agricultural machinery throughout its entire lifecycle, comprising the following steps: S1: Collect the analog voltage and engine speed pulse signals from the accelerometer of the tobacco farm machinery and calculate the vibration amplitude and speed respectively. Obtain the running mileage of the tobacco farm machinery and align it with the vibration amplitude and speed over time to generate initial running data. S2: Based on the initial running data, the running mileage is truncated according to the preset fixed mileage step size to obtain a discrete mileage interval set, the vibration amplitude and rotation speed are combined into interval feature data and the probability density is calculated to generate density peak data. S3: Compare the density peak data with a preset load intensity threshold, and convert the discrete mileage interval set and interval feature data that exceed the preset load intensity threshold into discrete topological data elements to generate a nonlinear path; S4: Based on the preset window length, count the data entries corresponding to the discrete topological data elements in the nonlinear path and calculate the density gradient difference by difference. Combine the results with the preset mutation judgment threshold and output the path reconstruction sequence. S5: Call the path reconstruction sequence to truncate the nonlinear path to construct independent operating interval segments, extract vibration amplitude and probability density to analyze the fatigue state throughout the entire life cycle and compare it with the preset fatigue safety value, and output operation and maintenance monitoring instructions.

[0006] As a further embodiment of the present invention, the initial operating data includes accelerometer analog voltage, engine speed, vibration amplitude, and operating mileage; the density peak data includes discrete mileage intervals, interval feature data, and probability density; the nonlinear path includes overload mileage intervals, interval feature data, and discrete topology data elements; the path reconstruction sequence includes data entry counts, density gradient differences, and mutation judgment identifiers; and the operation and maintenance monitoring instructions include independent operating interval segments, full life cycle fatigue status, and fatigue safety comparison results.

[0007] As a further aspect of the present invention, the specific steps of S1 are as follows: S101: Acquire the analog voltage signal from the accelerometer of the tobacco agricultural machinery and the engine speed pulse signal, perform linear normalization on the analog voltage signal, perform time counting and pulse frequency conversion on the pulse signal, match and align the vibration voltage value and the speed pulse value, and generate a time-aligned dataset. S102: Based on the time-aligned dataset, the amplitude of the analog voltage signal is calculated, the vibration amplitude sequence is extracted from the voltage amplitude at each time point, the engine speed pulse signal is converted by pulse counting to obtain the speed sequence, and then re-aggregated with the vibration amplitude sequence according to the timestamp to obtain the vibration amplitude and speed sequence. S103: Call the vibration amplitude and speed sequence and combine them to obtain the running mileage information of the tobacco agricultural machinery. Integrate each data according to the timestamp, and serialize and combine each record. Analyze the correspondence between vibration amplitude, speed value and running mileage to generate initial running data.

[0008] As a further aspect of the present invention, the specific steps of S2 are as follows: S201: Based on the initial running data, the running mileage is numerically truncated according to a preset fixed mileage step size, the continuous mileage values ​​are segmented and mapped according to the step size values, and the multi-segment intervals are indexed and numbered to obtain the discrete interval sequence structure and perform boundary value verification to obtain a discrete mileage interval set. S202: Based on the set of discrete mileage intervals, the vibration amplitude and rotational speed parameters within the corresponding intervals are paired and combined. The amplitude sequence and rotational speed sequence are aligned and spliced ​​according to the interval index. The spliced ​​records are vectorized and converted. The consistency of amplitude and rotational speed of the multi-vector terms is checked to generate interval feature data. S203: For the interval feature data, perform probability distribution numerical fitting on multiple feature vectors, perform frequency statistics and normalization on feature values ​​to construct a continuous density distribution sequence, extract and aggregate the feature elements corresponding to the positions that meet the preset ratio threshold conditions, and obtain density peak data.

[0009] As a further aspect of the present invention, the ratio threshold is determined by subtracting the maximum amplitude and distribution mean of the feature vectors included in the collected interval feature data, obtaining the deviation value by counting the total number of feature vectors included in the distribution sequence, and dividing the deviation value by the total number of feature vectors.

[0010] As a further aspect of the present invention, the specific steps of S3 are as follows: S301: Based on the density peak data, extract the data index corresponding to the position where the value exceeds the preset load intensity threshold, retrieve the corresponding discrete mileage interval set and interval feature data according to the data index, and splice and align the extracted mileage interval record and feature record to generate the over-threshold feature set. S302: Call the super-threshold feature set, perform graph structure topology mapping on the internal interval parameters and feature vectors, convert the interval center value into topology endpoints, calculate the feature evolution difference between multiple topology endpoints and convert it into connection weights, and at the same time split and encapsulate the associated topology graph structure to establish discrete topology data elements. S303: Based on the discrete topological data elements, extract the internal feature attribute parameters and topological association states of multiple data elements, perform serialization and combination sorting according to the connection weights between data elements, perform continuous curve numerical fitting on the coordinate items of the rearranged data elements, map the discrete distributed coordinate items, and generate a nonlinear path.

[0011] As a further aspect of the present invention, the load strength threshold is obtained by extracting the vibration amplitude and rotation speed parameters within the density peak data, multiplying the two to obtain the strength value, statistically analyzing the median and standard deviation parameters of multiple strength values, multiplying the variance parameter with a preset adjustment coefficient to generate an offset value, and then determining the median and offset value by addition.

[0012] As a further aspect of the present invention, the specific steps of S4 are as follows: S401: Construct a sliding window based on a preset window length and perform step-by-step traversal along the nonlinear path. At each step position, accumulate the total number of discrete topological data element associated data entries within the window. Then, splice the total number of entries obtained from multiple positions according to the window sliding sequence to generate a topological data counting sequence. S402: Call the topological data counting sequence, extract the total number of adjacent entries and perform a subtraction operation on the discrete difference term, divide the discrete difference term with the sliding step to obtain the single-point density change rate, and combine and splice multiple single-point density change rates according to the arrangement order to obtain the density gradient difference set. S403: For the density gradient difference set, extract multiple single-point density change rates and compare them with a preset mutation judgment threshold. Extract the associated index at the location where the mutation judgment threshold is exceeded and remove it. Based on the retained index, perform curve reconnection fitting on the original sequence coordinate points to generate a path reconstruction sequence.

[0013] As a further aspect of the present invention, the specific steps of S5 are as follows: S501: Call the path reconstruction sequence to truncate the nonlinear path to construct independent work interval segments, extract coordinate index items and cut and split the path continuity equation to obtain multiple sets of discrete curve trajectory items, assemble the multiple sets of discrete curve trajectory items with the associated mileage range, and establish an independent work interval set. S502: Based on the set of independent work intervals, extract the vibration amplitude parameter and probability density parameter within the same interval, perform multiplication to obtain the interval fatigue cumulative value, and perform overall superposition and summation of the interval fatigue cumulative values ​​obtained from all interval records to obtain the full life cycle fatigue state. S503: For the fatigue state of the entire life cycle, if the preset fatigue safety value is exceeded, extract the corresponding equipment identification code and the over-limit value item of the operating machinery, perform time sequence arrangement and protocol conversion encoding on the equipment identification code and the over-limit value item, and generate operation and maintenance monitoring instructions.

[0014] The tobacco agricultural machinery full life cycle operation and maintenance monitoring system includes: The data analysis module collects the analog voltage and engine speed pulse signals from the accelerometer of the tobacco farm machinery and calculates the vibration amplitude and speed respectively. It also obtains the running mileage of the tobacco farm machinery and aligns it with the vibration amplitude and speed over time to generate initial running data and transmit it to the density analysis module. The density analysis module, based on the initial running data, truncates the running mileage according to a preset fixed mileage step size to obtain a set of discrete mileage intervals, calls the vibration amplitude and rotation speed combination as interval feature data and calculates the probability density, generates density peak data and transmits it to the path generation module; The path generation module compares the density peak data with a preset load intensity threshold, converts the discrete mileage interval set and interval feature data that exceed the preset load intensity threshold into discrete topological data elements, generates a nonlinear path, and transmits it to the sequence reconstruction module. The sequence reconstruction module counts the data entries corresponding to the discrete topological data elements in the nonlinear path based on a preset window length and calculates the density gradient difference by difference. It then compares the results with a preset mutation judgment threshold, outputs the path reconstruction sequence, and transmits it to the operation and maintenance monitoring module. The operation and maintenance monitoring module calls the path reconstruction sequence to truncate the nonlinear path and construct independent operation intervals, extracts vibration amplitude and probability density to analyze the fatigue state throughout the entire life cycle and compares it with the preset fatigue safety value, and outputs operation and maintenance monitoring instructions.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, by simulating voltage through acceleration, aligning the rotational speed pulse with the mileage data time, a data foundation that can reflect the continuous changes in the working state is formed. Then, the vibration amplitude and rotational speed combination features are extracted according to a fixed mileage interval, and the probability density peak value is calculated. The high-load interval is transformed into discrete topological data elements to construct a nonlinear path. By combining window counting and density gradient difference, the location of abrupt changes is identified. Then, independent working intervals are truncated and global fatigue state indicators are extracted, so that the maintenance warning is aligned with the changes in the working intensity, improving the accuracy of deterioration identification and maintenance instructions. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the steps of the present invention; Figure 2 This is a detailed schematic diagram of S1 of the present invention; Figure 3 This is a detailed schematic diagram of S2 of the present invention; Figure 4 This is a detailed schematic diagram of S3 of the present invention; Figure 5 This is a detailed schematic diagram of S4 of the present invention; Figure 6 This is a detailed schematic diagram of S5 of the present invention; Figure 7 This is a system module diagram of the present invention. Detailed Implementation

[0018] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0019] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0020] Please see Figure 1 This invention provides a method for monitoring and maintaining tobacco agricultural machinery throughout its entire lifecycle, including the following steps: S1: Collect the analog voltage and engine speed pulse signals from the accelerometer of the tobacco farm machinery and calculate the vibration amplitude and speed respectively. Obtain the running mileage of the tobacco farm machinery and align it with the vibration amplitude and speed over time to generate initial running data. S2: Based on the initial running data, the running mileage is truncated according to the preset fixed mileage step size to obtain a discrete mileage interval set. The vibration amplitude and rotation speed are combined into interval feature data and the probability density is calculated to generate density peak data. S3: Compare the density peak data with the preset load intensity threshold, and convert the discrete mileage interval set and interval feature data that exceed the preset load intensity threshold into discrete topological data elements to generate nonlinear paths; S4: Based on the preset window length, count the data entries corresponding to the discrete topological data elements in the nonlinear path and calculate the density gradient difference by difference. Combined with the preset mutation judgment threshold, output the path reconstruction sequence. S5: Call the path reconstruction sequence to truncate the nonlinear path and construct independent operation intervals, extract vibration amplitude and probability density to analyze the fatigue state throughout the entire life cycle and compare it with the preset fatigue safety value, and output operation and maintenance monitoring instructions.

[0021] Initial operating data includes accelerometer analog voltage, engine speed, vibration amplitude, and operating mileage. Density peak data includes discrete mileage intervals, interval characteristic data, and probability density. Nonlinear paths include overload mileage intervals, interval characteristic data, and discrete topology data elements. Path reconstruction sequences include data entry counts, density gradient differences, and mutation judgment indicators. Operation and maintenance monitoring instructions include independent work intervals, full life cycle fatigue status, and fatigue safety comparison results.

[0022] Please see Figure 2 The specific steps of S1 are as follows: S101: Acquire the analog voltage signal from the accelerometer of the tobacco agricultural machinery and the engine speed pulse signal, perform linear normalization on the analog voltage signal, perform time counting and pulse frequency conversion on the pulse signal, match and align the vibration voltage value and the speed pulse value, and generate a time-aligned dataset. The analog voltage signal from the accelerometer installed on the chassis of the tobacco farm machinery is acquired in real time via a communication bus, and the rotational speed pulse signal is acquired through the crankshaft measurement node. The acquired analog voltage signal is continuous waveform data, which undergoes a linear normalization transformation. The specific transformation logic is as follows: extract the real-time voltage value; calculate half of the sum of the highest and lowest range voltage values ​​as the intermediate reference voltage value; calculate the voltage deviation by subtracting the real-time voltage value from the reference voltage value; then calculate the half-range span value by halving the difference between the highest and lowest range values; finally, calculate the normalized voltage value by the ratio of the voltage deviation to the half-range span value. For example, if the real-time voltage is 3.5 volts, the lowest range is 0 volts, and the highest range is 5 volts, half of the sum of the highest and lowest values ​​yields a reference voltage of 2.5 volts; the difference between 3.5 volts and 2.5 volts yields a deviation of 1.0 volt; half of the difference between the highest and lowest values ​​yields a span value of 2.5 volts; and dividing 1.0 volt by 2.5 volts gives a normalized voltage of 0.4. Subsequently, the pulse signal undergoes time counting and pulse frequency conversion operations. A 100-millisecond time window is set using the timing process, and the cumulative number of falling edges of the pulse signal is counted as the pulse count value. Dividing the count value by the time window duration yields the pulse frequency value. For example, counting 30 times within a 0.1-second window, dividing 30 by 0.1 gives a pulse frequency of 300 Hz. Next, the global absolute clock timestamp is retrieved, and the voltage sampling and pulse acquisition time points are compared. When the difference in microseconds between the two clocks is less than a tolerance threshold of 500 microseconds, the normalized voltage and the corresponding pulse frequency are bundled to generate a time-aligned dataset. This threshold is obtained by doubling the measured extreme value of the bus data transmission delay. The execution node of this operation logic is triggered by the comparison operation based on the clock difference, which then triggers the write operation of the alignment array.

[0023] S102: The amplitude of the analog voltage signal is calculated based on the time-aligned dataset. The vibration amplitude sequence is extracted from the voltage amplitude at each time point. The engine speed pulse signal is converted by pulse counting to obtain the speed sequence. The vibration amplitude and speed sequence are then re-aggregated according to the timestamp and the vibration amplitude sequence to obtain the vibration amplitude and speed sequence. The amplitude of analog voltage signals is calculated based on a pre-constructed time-aligned dataset. All normalized voltage values ​​within a 1000-millisecond time span are continuously extracted from the dataset as analysis samples. The root mean square (RMS) value of the voltage data within this time span is calculated to extract the vibration amplitude sequence. Specifically, the calculation logic is as follows: each extracted normalized voltage value is squared to obtain a single voltage square value. All voltage square values ​​within the current span are summed to obtain a total square. This total square is divided by the total number of sampling points to obtain the RMS value. Finally, the square root of this RMS value is taken to obtain the vibration amplitude. Combining the previously acquired data, three consecutive normalized voltage points were extracted with values ​​of 0.2, 0.4, and 0.4. These were squared to obtain 0.04, 0.16, and 0.16 respectively. The sum of these three square values ​​is 0.36. Dividing 0.36 by the sampling number 3 yields a RMS value of 0.12. Taking the square root of 0.12 yields the final vibration amplitude of 0.346. Simultaneously, the speed pulse signals in the dataset are converted into speed sequences by pulse counting. The conversion rule is to multiply the preceding pulse frequency value by 60 seconds (corresponding to 1 minute) to obtain the total number of pulses per minute, and then divide this number by the inherent number of mechanical pulses occurring per revolution of the crankshaft to obtain the number of revolutions per minute. Substituting the preceding pulse frequency of 300 Hz into the calculation, multiplying it by 60 gives a total number of pulses of 18,000. Setting the number of mechanical pulses occurring per revolution to a fixed constant of 4, 18,000 is divided by 4 to obtain the corresponding speed value of 4,500 revolutions per minute. Next, the data is re-aggregated according to the timestamp and vibration amplitude sequence, retrieving the corresponding time identifier codes, and concatenating the vibration amplitude and speed values ​​of the same time window into the data dictionary sub-items. Arranging and combining them according to time sequence yields the vibration amplitude and speed sequences. The execution node of this operation logic is triggered by the feature matching result based on the same timestamp identifier code, which triggers the heterogeneous data merging operation. Table 1 shows the data structure obtained by the above execution process. Table 1. Characteristics of Time-Aligned Vibration and Rotation Speed 1623548900 0.346 4500 125.5 1623548901 0.352 4510 125.6 1623548902 0.360 4550 125.7 Table 1 shows the specific numerical correlation patterns after time alignment.

[0024] S103: Call the vibration amplitude and speed sequence and combine them to obtain the running mileage information of tobacco agricultural machinery. Integrate each data according to the timestamp, and serialize and combine each record. Analyze the correspondence between vibration amplitude, speed value and running mileage to generate initial running data. The system retrieves the generated vibration amplitude and rotational speed sequences and reads the accumulated mileage information of the tobacco agricultural machinery via the communication bus. The acquired vibration amplitude and rotational speed sequences are integrated according to their corresponding timestamps, and the corresponding mileage values ​​are appended. Each record is serialized and arranged in a predefined field order to form a planar data sequence containing four items: timestamp, vibration amplitude, rotational speed, and mileage. Based on this, the relationship between vibration amplitude, rotational speed, and mileage is analyzed. Specifically, the underlying analysis logic involves extracting the vibration amplitude from the sequence and multiplying it by a scaling factor to obtain a vibration mapping value. This vibration mapping value is then divided by the rotational speed value in the same sequence to obtain a unit rotational speed oscillation index, which is compared with an empirical benchmark value. Substituting the previously mentioned vibration amplitude of 0.346 and rotational speed of 4500 into the calculation, and using a scaling factor of 10000 (set based on the data precision alignment requirement of mapping floating-point numbers to a 4-bit integer field), 0.346 is multiplied by 10000 to obtain a vibration mapping value of 3460. 3460 is then divided by the rotational speed of 4500 to obtain an oscillation index of 0.768. For the baseline value, an average of 3000 test indices collected from 10 normal devices was calculated to obtain 0.500. This 0.500 was then multiplied by a tolerance factor of 1.2 to arrive at the final baseline value of 0.600. The judgment logic then executes a condition where 0.768 is greater than the baseline value of 0.600, triggering a technical action to assign a high-frequency anomaly marker to the corresponding data sequence and generate initial operating data. The execution node of this calculation logic lies in the comparison between the oscillation index and the empirical baseline value, triggering the issuance of a status marker command.

[0025] Please see Figure 3 The specific steps of S2 are as follows: S201: Based on the initial running data, the running mileage is numerically truncated according to the preset fixed mileage step size, the continuous mileage values ​​are segmented and mapped according to the step size value, and the multi-segment intervals are indexed and numbered to obtain the discrete interval sequence structure and perform boundary value verification to obtain the discrete mileage interval set. The initial running data is obtained, and the running mileage value is extracted. The running mileage is then truncated according to a preset fixed mileage step size. Since continuously changing mileage cannot be directly used for discrete state statistics, it needs to be converted into a stepped, fixed-span interval. The truncation conversion logic is as follows: extract the current running mileage value, divide the current running mileage value by the fixed mileage step size to obtain a mileage quotient, round the quotient down to obtain an integer baseline value for the mileage, multiply this integer baseline value by the fixed mileage step size to obtain the lower bound value of the discrete interval, and then sum the lower bound value with the fixed mileage step size to obtain the upper bound value of the discrete interval. The preceding operating mileage value of 125.5 km is obtained. A fixed mileage step size of 10 km, derived from previous field tests and based on the ratio of the average round-trip distance of the agricultural machinery to its wear cycle, is used. 125.5 km is divided by 10 km to obtain a mileage quotient of 12.55. This quotient is rounded down to obtain an integer base value of 12. 12 is multiplied by 10 km to obtain a lower bound value of 120 km. The sum of 120 km and 10 km yields an upper bound value of 130 km. After segmenting and mapping the continuous mileage values, indexing and encoding are performed on the multi-segment intervals. The integer base value is summed with a constant 1 to obtain the interval index number. The sum of the integer base value 12 and 1 is directly assigned to the interval index number 13. Next, the discrete interval sequence structure is obtained and boundary value verification is performed. The logic is to determine whether the current running mileage value is greater than or equal to the lower boundary value and strictly less than the upper boundary value. 125.5 kilometers is greater than 120 kilometers and less than 130 kilometers, which meets the verification condition, confirming that the value belongs to the interval. If it does not meet the condition, it is marked as out-of-bounds data and removed. After each verification and filtering, a set of discrete mileage intervals containing the lower boundary value, the upper boundary value, and the interval index number is obtained. The execution node of this operation logic is that, based on the boundary comparison operation between the running mileage and the upper and lower boundary values ​​of the discrete interval, the discretization and truncation storage operation of the corresponding continuous data is triggered.

[0026] S202: Based on the discrete mileage interval set, the vibration amplitude and rotational speed parameters within the corresponding interval are paired and combined. The amplitude sequence and rotational speed sequence are aligned and spliced ​​according to the interval index. The spliced ​​record is vectorized and the consistency of amplitude and rotational speed is checked for multiple vector terms to generate interval feature data. Based on the constructed and verified set of discrete mileage intervals, the vibration amplitude and speed parameters within the corresponding intervals are paired and combined. The discrete mileage interval set sub-item with interval index number 13 is extracted, and the corresponding running record sequence belonging to this interval is retrieved from the initial running data to extract the vibration amplitude and speed parameters. The specific alignment and splicing process involves establishing a data dictionary with interval index numbers as keys, and sorting all vibration amplitude and speed parameters under the same key value in ascending order by absolute time in seconds. After aligning and splicing the amplitude sequence and speed sequence according to the interval index, a vectorization transformation operation is performed on the spliced ​​records. The transformation operation logic is as follows: the current rated maximum engine speed is obtained as the speed normalization benchmark; the speed value is divided by the speed normalization benchmark to obtain the relative speed load rate; then, the vibration amplitude and relative speed load rate are combined into a two-dimensional feature vector. Substituting the previously obtained vibration amplitude of 0.346 and rotational speed of 4500 rpm, the rated maximum rotational speed of the agricultural machinery, 6000 rpm, was retrieved as the speed normalization benchmark. Dividing 4500 rpm by 6000 rpm yielded a relative speed load rate of 0.750. The vibration amplitude of 0.346 and the speed load rate of 0.750 were combined to construct a two-dimensional spliced ​​record feature vector. Next, the amplitude and rotational speed of the multi-vector items were checked for consistency. The logic was to divide the vibration amplitude in the two-dimensional feature vector by the relative speed load rate to derive the unit load vibration ratio. This unit load vibration ratio was then compared with a set benchmark consistency judgment interval. The judgment interval was defined by the percentage of the extreme values ​​of the equipment operating data under no-wear conditions fluctuating within a 10% tolerance range, set to 0.300 to 0.600. Dividing 0.346 by 0.750 yields a unit load vibration ratio of 0.461. Comparing 0.461 with the judgment interval, and finding it to be within the range of 0.300 to 0.600, the consistency check passes, generating interval feature data containing verification status markers. The execution node of this logic is triggered by the persistent storage of the verified feature vector based on the comparison result between the unit load vibration ratio and the judgment interval.

[0027] S203: For interval feature data, perform probability distribution numerical fitting on multiple feature vectors, perform frequency statistics and normalization on feature values ​​to construct a continuous density distribution sequence, extract and aggregate the feature elements corresponding to the positions that meet the preset ratio threshold conditions, and obtain density peak data. For the generated interval feature data, probability distribution numerical fitting is performed on multiple feature vectors. Since individual feature values ​​exhibit fluctuations and spikes, their distribution patterns need to be discretized and aggregated for extraction. All unit load vibration ratios that pass consistency verification are extracted from the interval feature data as feature value samples. A fixed distribution statistical class interval is set, dividing the total data span of the feature values ​​into multiple continuous distribution statistical intervals. Frequency statistics and normalization are performed on the feature values ​​to construct a continuous density distribution sequence. The calculation logic is as follows: the cumulative number of feature values ​​falling within a single distribution statistical interval is counted to obtain the frequency statistical value. This frequency statistical value is divided by the total number of feature values ​​participating in the statistics to calculate the probability density value reflecting the degree of data aggregation. The previously generated unit load vibration ratio of 0.461 is used, and the distribution statistical class interval is set to 0.020. This constructs a distribution statistical interval with a lower limit of 0.450 and an upper limit of 0.470, and the vibration ratio 0.461 is assigned to this statistical interval. Assuming the total number of feature values ​​collected in this batch is 5000, and the frequency of values ​​falling within the 0.450 to 0.470 range is 1200, dividing 1200 by 5000 yields a probability density value of 0.240 for this range. Subsequently, feature elements corresponding to locations meeting a preset proportion threshold are extracted and aggregated. The extraction logic involves setting a preset proportion threshold representing a high-density data area, iterating through the probability density values ​​of all ranges, and extracting the median of the distribution for a range greater than the preset proportion threshold as the density peak data. The preset proportion threshold is set to 0.200, established based on large-scale statistical derivation of historical fault characteristics. The probability density value of 0.240 is compared with the preset proportion threshold of 0.200; since 0.240 is greater than 0.200, the extraction condition is met. The sum of 0.470 and 0.450 is divided by a constant 2 to obtain the median of the distribution, 0.460. The aggregated output is then used to obtain the median of the distribution, 0.460, as the density peak data. The execution node of this operation logic is to trigger the extraction operation of the median of the high-frequency feature distribution based on the comparison operation between the probability density value and the preset ratio threshold.

[0028] Please see Figure 4 The specific steps of S3 are as follows: S301: Based on the density peak data, extract the data index corresponding to the position where the value exceeds the preset load intensity threshold, retrieve the corresponding discrete mileage interval set and interval feature data according to the data index, and splice and align the extracted mileage interval record and feature record to generate the over-threshold feature set. The extracted density peak data is obtained and substituted into the previously obtained distribution median of 0.460 as the density peak data. After extracting this data, a preset load intensity threshold is set. This threshold is established based on the minimum value obtained by time-series sampling of the operating load of the same model of agricultural machinery chassis 30 minutes before the failure, specifically set to 0.420. The density peak data 0.460 is compared with the preset load intensity threshold 0.420. 0.460 is greater than 0.420, satisfying the trigger condition of the value exceeding the threshold. The corresponding data index is extracted, and the previous interval index number 13 belonging to this density peak data is obtained as the data index. Based on data index 13, pointer matching operation is performed in the storage space to retrieve the corresponding discrete mileage interval set and interval feature data. The lower limit value of 120 km and the upper limit value of 130 km generated in the previous sequence are read, and the unit load vibration ratio of 0.461, vibration amplitude of 0.346, and speed load rate of 0.750 within this interval are read. The extracted mileage interval records and feature records are spliced ​​and aligned to establish a horizontal data table structure with the unique equipment identification serial number as the retrieval prefix. The lower limit value of 120 km and the upper limit value of 130 km are placed in the mileage interval record column. The unit load vibration ratio of 0.461, vibration amplitude of 0.346, and speed load rate of 0.750 are merged and placed in the feature record column. The values ​​under the same serial number and interval index number are horizontally merged and stored using row pointers to generate a feature set exceeding the threshold. The execution node of this operation logic is triggered by the comparison result between the density peak data and the preset load intensity threshold, which triggers the feature matrix splicing and storage operation of the high-risk mileage interval.

[0029] S302: Call the over-threshold feature set, perform graph structure topology mapping on the internal interval parameters and feature vectors, convert the interval center value into topology endpoints, calculate the feature evolution difference between multiple topology endpoints and convert it into connection weights, and at the same time split and encapsulate the associated topology graph structure to establish discrete topology data elements. The generated over-threshold feature set is used to perform graph topological mapping on the internal interval parameters and feature vectors. The lower bound value of 120 km and the upper bound value of 130 km generated by splicing are read. The sum of 120 km and 130 km yields a total span of 250 km. Dividing the total span of 250 km by a constant 2, the interval center value of 125 km is derived. The interval center value of 125 km is converted into the horizontal axis topological endpoint coordinates of the topological graph structure. Then, the feature evolution difference between multiple topological endpoints is calculated and converted into connection weights. The unit load vibration ratio of the previous adjacent valid data element (0.430) is retrieved, and the unit load vibration ratio of the current endpoint (0.461) is extracted. The difference between 0.461 and 0.430 is calculated to obtain the absolute value of the feature evolution difference (0.031). To avoid zero-value overflow, a smoothing compensation coefficient of 0.009 is set. The feature evolution difference of 0.031 and 0.009 are summed to obtain a denominator of 0.040. Dividing the denominator by a constant 1 yields a connection weight of 25.000. Next, the associated topology graph structure is split and encapsulated. A lower limit for graph structure splitting and truncation is set to 20.000. The calculated connection weight of 25.000 is compared with the truncation lower limit of 20.000. If 25.000 is greater than 20.000, the connection edges between the topology endpoints are retained. All topology endpoints that pass the truncation comparison and their corresponding connection edges are extracted into the global graph network, forming an independent locally connected subgraph network structure. Memory encapsulation and allocation are performed on each point and edge attribute, and weak connections with values ​​less than the truncation lower limit are disconnected, thus establishing a complete discrete topology data element. The execution node of this operation logic is the independent memory encapsulation operation of the target topology subgraph triggered by the comparison and filtering operation based on the connection weight and the truncation lower limit.

[0030] S303: Based on discrete topological data elements, extract the internal feature attribute parameters and topological association status of multiple data elements, perform serialization and combination sorting according to the connection weight between data elements, perform continuous curve numerical fitting on the coordinate items of the rearranged data elements, map the discrete distributed coordinate items, and generate nonlinear paths. The system acquires the established discrete topology data elements and extracts the internal characteristic attribute parameters and topology association status of multiple data elements. It reads the interval center value (125 km), unit load vibration ratio (0.461), and connection weight (25.000) of the current data element. Simultaneously, it extracts the center value (115 km), vibration ratio (0.430), and connection weight (22.000) of directly adjacent data elements. Based on the connection weights between data elements, it performs serialization and sorting, placing all extracted connection weight values ​​in a one-dimensional array. It performs a pairwise comparison and swapping of adjacent elements, comparing connection weights 25.000 and 22.000. Since 25.000 is greater than 22.000, the data element node containing 25.000 is moved to the front of the memory queue, completing the descending serialization rearrangement. A continuous curve numerical fitting was performed on the rearranged data element coordinates. The difference between the center value of 125 km and 115 km yielded a mileage difference of 10 km. The difference between the vibration ratios of 0.461 and 0.430 yielded a feature difference of 0.031. Dividing the feature difference of 0.031 by the mileage difference of 10, the local linear slope value of 0.0031 was derived. Multiplying the center value of 125 km by the slope of 0.0031 yielded an offset product of 0.3875. Subtracting 0.3875 from the vibration ratio of 0.461 yielded an intercept term of 0.0735. The discrete coordinates were mapped, and a point-by-point coherent polynomial parameter set was established based on the local linear slope of 0.0031 and the intercept term of 0.0735. According to this parameter set, the ordinate values ​​were written pixel by pixel in a two-dimensional coordinate system according to the function expression, generating a nonlinear path containing a direction vector. The execution node of this operation logic is to trigger the mapping and drawing operation of discrete feature nodes to continuous trajectory curves based on the derivation results of polynomial slope and intercept.

[0031] Please see Figure 5 The specific steps of S4 are as follows: S401: Construct a sliding window based on the preset window length and perform step traversal along a nonlinear path. At each step position, accumulate the total number of discrete topological data element associated data entries within the window. Then, splice the total number of entries obtained from multiple positions according to the window sliding order to generate a topological data counting sequence. The generated nonlinear path data containing direction vectors is acquired, and the statistical spatial span average of structural microcracks occurring in the agricultural machinery chassis during its historical operating cycle is extracted and set as the preset window length, specifically 50 kilometers. The sliding step distance of the sliding window along the centerline of the nonlinear path is set to 10 kilometers. A step-by-step traversal operation is performed along the nonlinear path, with the initial starting position set to path mileage 0 kilometers. When the first sliding step reaches 10 kilometers, all discrete topological data elements within the window mileage range of 10 kilometers to 60 kilometers are extracted. Before the accumulation operation, a null value removal preprocessing operation is performed on the extracted sample records to filter out invalid entries with missing fields. The number of record entries encapsulated within each valid data element is extracted, and the number of entries of all data elements within this range is summed. For example, if the first data element contains 12 records, the second data element contains 18 records, and the third data element contains 15 records, then the records 12, 18, and 15 are continuously added together, resulting in a total of 45 entries at the first step position. At the second step to 20 kilometers, data elements within the mileage range of 20 to 70 kilometers are extracted, resulting in a total of 52 entries. At the third step to 30 kilometers, the total number of entries is 61. The total number of entries obtained from multiple positions is then horizontally concatenated and merged according to the sliding time sequence. The records 45, 52, and 61 are sequentially written into a linear memory array structure to generate a topological data counting sequence. The execution node of this operation logic is triggered by the sequential storage operation of the serialized array based on the accumulated node results within the window.

[0032] S402: Call the topological data counting sequence, extract the total number of adjacent entries and perform a subtraction operation on the discrete difference term, divide the discrete difference term with the sliding step to obtain the single-point density change rate, and combine and splice multiple single-point density change rates according to the arrangement order to obtain the density gradient difference set. Retrieve the generated topological data counting sequence and extract the total number of the two adjacent entries in the array structure according to time order. Read the total number of entries (52) at the second step position in the previous step and the total number of entries (45) at the first step position. Subtract the total number of entries (52) from the total number of entries (45) to obtain 7 discrete difference items. Extract the previously set window sliding step size (10 km). Divide the calculated discrete difference items (7) by this sliding step size (10 km) to derive a single-point density change rate of 0.70 items per kilometer. Continue extracting the total number of entries (61) at the third step position. Subtract the 61 items from the 52 items at the second position to obtain 9 discrete difference items for the next interval. Divide these 9 discrete difference items by the sliding step size (10 km) to obtain the corresponding single-point density change rate of 0.90 items per kilometer. The fourth step position entries, totaling 55, are extracted. The difference between these 55 and the previous 61 entries yields a discrete difference of -6. Dividing -6 by 10 derives a single-point density change rate of -0.60 per kilometer. Based on the order of positions in the original topological data counting sequence, multiple single-point density change rate values ​​are sequentially pushed into a floating-point data set for combination and concatenation, forming data sets containing 0.70, 0.90, and -0.60, thus obtaining the density gradient difference set. The execution node of this operation is triggered by the sequential aggregation and storage operation of the node density change gradient based on the result of the division between the difference term and the step size value.

[0033] S403: For density gradient difference set, extract multiple single-point density change rates and compare them with preset mutation judgment thresholds. Extract the associated indexes at the locations where the mutation judgment thresholds are exceeded and remove them. Based on the retained indexes, associate the original sequence coordinate points and perform curve reconnection fitting to generate path reconstruction sequence. The generated density gradient difference set is retrieved, and the 95% confidence upper limit of the load density before fatigue fracture in the historical life cycle of agricultural machinery of the same model is obtained as the preset mutation judgment threshold. This benchmark value is limited to 0.85 per kilometer. Multiple single-point density change rates are extracted from the difference set and compared one by one with the preset mutation judgment threshold of 0.85. The first single-point density change rate of 0.70 per kilometer, calculated previously, is read and compared with 0.85. Since 0.70 is less than 0.85, it is marked as normal and its first step position index is retained. The second change rate of 0.90 per kilometer is read and compared with 0.85. Since 0.90 is greater than 0.85, the judgment condition is met, and the second step position index associated with the position where this value exceeds the preset judgment threshold is extracted. This second step position index is removed from the reserved memory queue and a zero-wiring removal operation is performed, cutting off the forward and backward memory pointers of this index. The original sequence coordinates associated with the retained index are read. The x-coordinate of 15 km and the y-intercept of 0.086 corresponding to the first position index are extracted, as well as the x-coordinate of 35 km and the y-intercept of 0.098 associated with the third position index. These retained coordinates are substituted into a multivariate linear derivation matrix for curve reconnection fitting calculation. Subtracting 0.086 from 0.098 yields the y-intercept difference of 0.012, and subtracting 15 from 35 yields the y-intercept span of 20. Dividing 0.012 by 20 derives the reconstruction slope of 0.0006, generating a path reconstruction sequence. The execution node of this operation logic is triggered by the comparison between the density change rate and the judgment threshold, which eliminates high-risk discrete nodes and regenerates the coherent trajectory of the remaining nodes.

[0034] Please see Figure 6 The specific steps of S5 are as follows: S501: Call the path reconstruction sequence to truncate the nonlinear path and construct independent work interval segments, extract coordinate index items and cut and split the path continuity equation to obtain multiple sets of discrete curve trajectory items, assemble the multiple sets of discrete curve trajectory items with the associated mileage range, and establish an independent work interval set. The output path reconstruction sequence is obtained, and the transverse coordinate breakpoints recorded in the reconstruction sequence are extracted. Based on this breakpoint set, the original nonlinear path is physically truncated into segments, constructing multiple sets of mutually separated independent work intervals. During the truncation operation, the previously generated retained coordinate points are read, and the transverse coordinate of the first position (15 km) is extracted as the starting index, and the transverse coordinate of the third position (35 km) is extracted as the ending index. Based on the extracted coordinate indexes, a domain-limiting truncation operation is performed on the continuous polynomial path equation that originally covered the global mileage. The continuous equation is numerically divided within a closed interval of 15 km to 35 km, and all discrete mapping node data within this interval are extracted to obtain the first set of discrete curve trajectory terms limited to this coordinate range. Simultaneously, the starting coordinates of the next segment (35 km to 50 km) are read, and the same equation truncation operation is performed to obtain the second set of discrete curve trajectory terms. The system invokes a memory concatenation operation to merge the first set of discrete curve trajectory items with the associated mileage range of 15 km to 35 km horizontally, and merges the second set of trajectory items with the range of 35 km to 50 km. The merged data segments are then sequentially written into a hierarchical tree-structured storage space, establishing an independent set of work intervals. The execution node of this operation logic lies in the forced segmentation operation of the equation's domain based on spatial coordinate breakpoints, triggering the segmented decoupled storage operation of the trajectory data. This segmented data indicates the distribution boundaries of the driving trajectory in the local work area, and it forms a direct data mapping inheritance relationship with the path reconnection results.

[0035] S502: Based on the set of independent working intervals, the vibration amplitude parameters and probability density parameters within the same interval are extracted and multiplied to obtain the interval fatigue cumulative value. The interval fatigue cumulative values ​​obtained from all interval records are superimposed and summed to obtain the full life cycle fatigue state. The generated set of independent operating intervals is retrieved, and the vibration amplitude and probability density parameters within the same mileage range are extracted by traversing the underlying storage nodes. Before data extraction, the historical acquisition logs of the triaxial accelerometer installed on the agricultural machinery chassis suspension are used as the input source. After extracting the sampled values ​​from five consecutive time points, summing and averaging them to remove extreme noise, the smoothed vibration amplitude parameter within the mileage range of 15 km to 35 km is read as 0.346. Simultaneously, the discrete integral area calculation is performed on the load distribution histogram within this interval, and the central tendency probability density parameter of the quantized output is read as 0.820. The read vibration amplitude parameter 0.346 and probability density parameter 0.820 are placed in the arithmetic logic control unit for multiplication, and the first set of interval fatigue cumulative values ​​of 0.283 is derived by multiplying the two. Continuing to traverse the next independent operating segment, from 35 km to 50 km, the vibration amplitude of 0.450 and probability density of 0.928 are read and multiplied to obtain the second set of segment fatigue cumulative values ​​of 0.417. The internal accumulation calculator is invoked to perform a total summation of the segment fatigue cumulative values ​​recorded for all segments. The first set of 0.283 and the second set of 0.417 are summed to calculate the full life-cycle fatigue state value of 0.700. The execution node of this calculation logic lies in the calculation based on the multiplication and accumulation of fatigue characteristic values ​​across multiple segments, triggering the quantitative aggregation operation of the global mechanical stress state. This accumulated summation value indicates the quantification degree of overall material fatigue damage during the equipment's service life; its continuous increase represents a direct extension from single-segment risk to overall lifespan degradation assessment.

[0036] S503: For fatigue status throughout the entire life cycle, if the preset fatigue safety value is exceeded, extract the corresponding equipment identification code and the over-limit value item of the operating machinery, perform time sequence arrangement and protocol conversion encoding on the equipment identification code and the over-limit value item, and generate operation and maintenance monitoring instructions. The generated full-lifecycle fatigue state value of 0.700 is extracted and input into a comparison register for comparison with a preset fatigue safety value. This preset fatigue safety value is derived from the yield strength of the same material standard tensile fracture test sample, reduced by 15%, and is set to 0.650. During the comparison operation, if the full-lifecycle fatigue state value of 0.700 is greater than the preset fatigue safety value of 0.650, the over-limit condition is met, triggering an alarm data extraction operation. The equipment basic attribute data table is accessed, and the equipment identification code corresponding to the operating machine that generated this state value is extracted as the numeric string 320512. The difference of 0.050 is calculated by subtracting the safety value of 0.650 from the state value of 0.700, and taken as the over-limit value item. The current absolute time clock of the main board is retrieved and converted into a pure digital sequence format as a sorting reference. The equipment identification code 320512 and the over-limit value item 0.050 are time-preordered, generating a combined character sequence containing a time prefix. The rearranged sequence undergoes protocol conversion encoding. Based on communication pin level rules, the identifier and out-of-limit value are split into binary bit data streams. A parity check bit is appended to the end of the data stream for packet encapsulation, generating maintenance monitoring instructions that can be directly parsed and executed by the remote controller. The execution node of this logic is triggered by comparing the fatigue state assessment value with the baseline safety value, leading to the extraction of the out-of-limit equipment identifier and the re-issuance of the underlying communication data packet. This instruction data reflects the actual dangerous situation of irreversible damage to the equipment, providing hard-wired-level communication credentials for subsequent intervention actions such as cutting off the equipment's ignition circuit to execute a forced shutdown.

[0037] Please see Figure 7 The tobacco agricultural machinery full life cycle operation and maintenance monitoring system includes: The data analysis module collects the analog voltage and engine speed pulse signals from the accelerometer of the tobacco farm machinery and calculates the vibration amplitude and speed respectively. It also obtains the running mileage of the tobacco farm machinery and aligns it with the vibration amplitude and speed over time to generate initial running data and transmit it to the density analysis module. The density analysis module, based on the initial running data, truncates the running mileage according to the preset fixed mileage step size to obtain a set of discrete mileage intervals, calls the vibration amplitude and rotation speed combination as interval feature data and calculates the probability density, generates density peak data and transmits it to the path generation module. The path generation module compares the density peak data with the preset load intensity threshold, converts the discrete mileage interval set and interval feature data that exceed the preset load intensity threshold into discrete topological data elements, generates nonlinear paths, and transmits them to the sequence reconstruction module. The sequence reconstruction module counts the data entries corresponding to discrete topological data elements in the nonlinear path based on a preset window length and calculates the density gradient difference by difference. It then compares the results with a preset mutation judgment threshold, outputs the path reconstruction sequence, and transmits it to the operation and maintenance monitoring module. The operation and maintenance monitoring module calls the path reconstruction sequence to truncate the nonlinear path and construct independent operation intervals. It extracts vibration amplitude and probability density to analyze the fatigue state throughout the entire life cycle and compares it with the preset fatigue safety value, and outputs operation and maintenance monitoring instructions.

[0038] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A tobacco farm machine full life cycle operation and maintenance monitoring method, characterized in that, Includes the following steps: S1: Collect the analog voltage of the accelerometer and the engine speed pulse signal of the tobacco farm machinery, calculate the vibration amplitude and speed respectively, obtain the running mileage of the tobacco farm machinery, and align it with the vibration amplitude and speed time to generate initial running data; the initial running data includes the accelerometer analog voltage, engine speed, vibration amplitude, and running mileage; S2: Based on the initial running data, the running mileage is truncated according to a preset fixed mileage step size to obtain a discrete mileage interval set. The vibration amplitude and rotation speed are combined to form interval feature data and the probability density is calculated to generate density peak data. The density peak data includes discrete mileage intervals, interval feature data, and probability density. Specifically, it includes: S201: Based on the initial running data, the running mileage is numerically truncated according to a preset fixed mileage step size, the continuous mileage values ​​are segmented and mapped according to the step size values, and the multi-segment intervals are indexed and numbered to obtain the discrete interval sequence structure and perform boundary value verification to obtain a discrete mileage interval set. S202: Based on the set of discrete mileage intervals, the vibration amplitude and rotational speed parameters within the corresponding intervals are paired and combined. The amplitude sequence and rotational speed sequence are aligned and spliced ​​according to the interval index. The spliced ​​records are vectorized and converted. The consistency of amplitude and rotational speed of the multi-vector terms is checked to generate interval feature data. S203: For the interval feature data, perform probability distribution numerical fitting on multiple feature vectors, perform frequency statistics and normalization on feature values ​​to construct a continuous density distribution sequence, extract and aggregate the feature elements corresponding to the positions that meet the preset ratio threshold conditions, and obtain density peak data. S3: Compare the density peak data with a preset load intensity threshold, and convert the discrete mileage interval set and interval feature data that exceed the preset load intensity threshold into discrete topology data elements to generate a nonlinear path; the nonlinear path includes overload mileage intervals, interval feature data, and discrete topology data elements. Specifically, it includes: S301: Based on the density peak data, extract the data index corresponding to the position where the value exceeds the preset load intensity threshold, retrieve the corresponding discrete mileage interval set and interval feature data according to the data index, and splice and align the extracted mileage interval record and feature record to generate the over-threshold feature set. S302: Call the super-threshold feature set, perform graph structure topology mapping on the internal interval parameters and feature vectors, convert the interval center value into topology endpoints, calculate the feature evolution difference between multiple topology endpoints and convert it into connection weights, and at the same time split and encapsulate the associated topology graph structure to establish discrete topology data elements. Extract all topological endpoints and their corresponding connecting edges that pass the truncation comparison into a global graph network, form an independent locally connected subgraph network structure, encapsulate and allocate memory for each point and edge attribute, disconnect the remaining weakly related edges that are less than the truncation lower limit, and then establish a complete discrete topological data element. S303: Based on the discrete topological data elements, extract the internal feature attribute parameters and topological association states of multiple data elements, perform serialization and combination sorting according to the connection weights between data elements, perform continuous curve numerical fitting on the coordinate items of the rearranged data elements, map the discrete distributed coordinate items, and generate a nonlinear path. S4: Based on a preset window length, count the data entries corresponding to the discrete topological data elements in the nonlinear path and calculate the density gradient difference by difference. Combine this with a preset mutation judgment threshold for comparison, and output the path reconstruction sequence. The path reconstruction sequence includes data entry count, density gradient difference, and mutation judgment identifier. Specifically, it includes: S401: Construct a sliding window based on a preset window length and perform step-by-step traversal along the nonlinear path. At each step position, accumulate the total number of discrete topological data element associated data entries within the window. Then, splice the total number of entries obtained from multiple positions according to the window sliding sequence to generate a topological data counting sequence. S402: Call the topological data counting sequence, extract the total number of adjacent entries and perform a subtraction operation on the discrete difference term, divide the discrete difference term with the sliding step to obtain the single-point density change rate, and combine and splice multiple single-point density change rates according to the arrangement order to obtain the density gradient difference set. S403: For the density gradient difference set, extract multiple single-point density change rates and compare them with a preset mutation judgment threshold. Extract the associated index at the location where the mutation judgment threshold is exceeded and remove it. Based on the retained index, associate the original sequence coordinate points with curve reconnection fitting to generate a path reconstruction sequence. S5: Call the path reconstruction sequence to truncate the nonlinear path to construct independent working interval segments, extract vibration amplitude and probability density to analyze the fatigue state throughout the entire life cycle and compare it with the preset fatigue safety value, and output operation and maintenance monitoring instructions; the operation and maintenance monitoring instructions include independent working interval segments, fatigue state throughout the entire life cycle, and fatigue safety comparison results.

2. The tobacco farm machine full life cycle operation and maintenance monitoring method according to claim 1, characterized in that, The specific steps of S1 are as follows: S101: Acquire the analog voltage signal from the accelerometer of the tobacco agricultural machinery and the engine speed pulse signal, perform linear normalization on the analog voltage signal, perform time counting and pulse frequency conversion on the pulse signal, match and align the vibration voltage value and the speed pulse value, and generate a time-aligned dataset. S102: Based on the time-aligned dataset, the amplitude of the analog voltage signal is calculated, the vibration amplitude sequence is extracted from the voltage amplitude at each time point, the engine speed pulse signal is converted by pulse counting to obtain the speed sequence, and then re-aggregated with the vibration amplitude sequence according to the timestamp to obtain the vibration amplitude and speed sequence. S103: Call the vibration amplitude and speed sequence and combine them to obtain the running mileage information of the tobacco agricultural machinery. Integrate each data according to the timestamp, and serialize and combine each record. Analyze the correspondence between vibration amplitude, speed value and running mileage to generate initial running data.

3. The tobacco farm machine full life cycle operation and maintenance monitoring method according to claim 1, characterized in that, The ratio threshold is determined by subtracting the maximum amplitude and distribution mean of the feature vectors included in the collected interval feature data, obtaining the deviation value, counting the total number of feature vectors included in the distribution sequence, and dividing the deviation value by the total number of feature vectors.

4. The method for monitoring and maintaining tobacco agricultural machinery throughout its entire lifecycle according to claim 1, characterized in that, The load strength threshold is obtained by extracting the vibration amplitude and rotation speed parameters from the density peak data, multiplying the two to obtain the strength value, statistically analyzing the median and standard deviation parameters of multiple strength values, multiplying the variance parameter with a preset adjustment coefficient to generate a bias value, and then adding the median and bias value to determine the load strength threshold.

5. The method for monitoring and maintaining tobacco agricultural machinery throughout its entire lifecycle according to claim 1, characterized in that, The specific steps of S5 are as follows: S501: Call the path reconstruction sequence to truncate the nonlinear path to construct independent work interval segments, extract coordinate index items and cut and split the path continuity equation to obtain multiple sets of discrete curve trajectory items, assemble the multiple sets of discrete curve trajectory items with the associated mileage range, and establish an independent work interval set. S502: Based on the set of independent work intervals, extract the vibration amplitude parameter and probability density parameter within the same interval, perform multiplication to obtain the interval fatigue cumulative value, and perform overall superposition and summation of the interval fatigue cumulative values ​​obtained from all interval records to obtain the full life cycle fatigue state. S503: For the fatigue state of the entire life cycle, if the preset fatigue safety value is exceeded, extract the corresponding equipment identification code and the over-limit value item of the tobacco agricultural machinery, perform time sequence arrangement and protocol conversion encoding on the equipment identification code and the over-limit value item, and generate operation and maintenance monitoring instructions.

6. A full life-cycle operation and maintenance monitoring system for tobacco agricultural machinery, characterized in that, The system is used to implement the tobacco agricultural machinery full life cycle operation and maintenance monitoring method according to any one of claims 1-5, and the system includes: The data analysis module collects the analog voltage and engine speed pulse signals from the accelerometer of the tobacco farm machinery and calculates the vibration amplitude and speed respectively. It also obtains the running mileage of the tobacco farm machinery and aligns it with the vibration amplitude and speed over time to generate initial running data and transmit it to the density analysis module. The density analysis module, based on the initial running data, truncates the running mileage according to a preset fixed mileage step size to obtain a set of discrete mileage intervals, calls the vibration amplitude and rotation speed combination as interval feature data and calculates the probability density, generates density peak data and transmits it to the path generation module; The path generation module compares the density peak data with a preset load intensity threshold, converts the discrete mileage interval set and interval feature data that exceed the preset load intensity threshold into discrete topological data elements, generates a nonlinear path, and transmits it to the sequence reconstruction module. The sequence reconstruction module counts the data entries corresponding to the discrete topological data elements in the nonlinear path based on a preset window length and calculates the density gradient difference by difference. It then compares the results with a preset mutation judgment threshold, outputs the path reconstruction sequence, and transmits it to the operation and maintenance monitoring module. The operation and maintenance monitoring module calls the path reconstruction sequence to truncate the nonlinear path and construct independent operation intervals, extracts vibration amplitude and probability density to analyze the fatigue state throughout the entire life cycle and compares it with the preset fatigue safety value, and outputs operation and maintenance monitoring instructions.

Images