Automatic test method for aging characteristic of oscillator

Abstract

The invention relates to the technical field of electronic testing, in particular to an oscillator aging characteristic automatic testing method, which comprises the following steps: identifying frequency fluctuation direction change, extracting mutation and duration time marking fragments, merging a content recombination distribution diagram with consistent direction, constructing a trend path, extracting an intersection point and a reversal point identification hopping structure, and determining the aging characteristic of an oscillator. And filing continuous offset fragments, eliminating interrupted content, writing into a sequence, and controlling the number to generate a detection scheme. According to the method, high-density sorting of frequency data is realized by extracting a frequency fluctuation direction and a sudden change point, accurately positioning a frequency change trend, identifying continuous offset and turning behaviors, combining trend path extension and intersection point extraction, enhancing the capability of capturing frequency response structure change, retaining fragments with consistent directions, and eliminating interruption interference; and the fragments are filed according to trends, and capacity constraints are set, so that trend identification precision and response structure expression ability in aging behaviors are improved, and analysis depth and reliability judgment support of a performance evolution process are enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of electronic testing, and in particular to an automatic testing method for aging characteristics of an oscillator. Background Art

[0002] The field of electronic testing technology encompasses technical means for quantitatively evaluating the electrical performance, electrical parameters, functional status, and operational stability of various electronic devices or circuits. The core content of this technical field includes test devices and methods for performing electrical signal measurement, voltage and current analysis, frequency detection, and noise measurement on electronic components such as resistors, capacitors, transistors, integrated circuits, and oscillators under different operating conditions. Electronic testing technology can be subdivided into several subcategories, including component-level testing, system-level testing, and automated testing, and is widely used in electronic manufacturing, quality control, aging screening, and failure analysis. Its overall technical system usually consists of a signal source, a measurement interface, a data acquisition unit, a measurement algorithm, and a control program. The testing method can be static testing, dynamic testing, scanning testing, or periodic sampling testing.

[0003] Among them, the automatic test method for the aging characteristics of the oscillator refers to a method for performing fixed-point acquisition and time series measurement of performance parameters such as output frequency stability, phase noise, and frequency drift of devices such as voltage-controlled oscillators or crystal oscillators during continuous operation, by setting different working hours and environmental conditions. The technical matters covered by this patent subject include powering on the oscillator during a preset aging cycle, using a frequency meter and a phase-locked loop measurement device to capture the frequency and track the drift, continuously recording the output frequency at set time intervals, and verifying key frequency change points. The specific method is to periodically power on and off the oscillator through a programmable control device, and to perform sequence comparison of the collected frequency data through a frequency analysis device to form a data matrix for aging laws, which is used to support the technical evaluation of the performance changes of the oscillator during long-term operation.

[0004] Fixed-time sampling mechanisms that rely on frequency recording often focus solely on point-to-point numerical changes in practice, failing to conduct in-depth analysis of fluctuation direction, trend extension paths, or trend relationships between multiple points. This results in a separation of trend reversals and continuous excursions in the frequency response, making it difficult to construct a dynamic structural map of frequency behavior. Directional reversals or trend crossings in frequency variations are easily dismissed as general excursions, lacking a response mechanism to classify and identify their structural properties, making it impossible to construct a database of aging behavior signatures focused on trend continuity. In archiving, most techniques rely on time series superposition or mean comparison, lacking effective filtering and reconstructing logic for trend consistency in frequency data. This results in a redundant accumulation of spurious fragments in the data, impairing the clarity of the archived data and limiting the efficient extraction of key features of the aging path. For example, during the aging process of high-frequency oscillators, frequency abrupt changes are often intertwined with slowly varying sequences. Without directional integration and trend reconstructing, the aging response behavior can be blurred, leading to assessment errors and compromising the accuracy of device lifetime assessment. The inadequate understanding and processing of the structural properties of frequency data by current technologies has become a major bottleneck restricting the effectiveness of high-precision aging testing. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings in the prior art and to propose an automatic testing method for aging characteristics of an oscillator.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: an automatic testing method for oscillator aging characteristics, comprising the following steps:

[0007] S1: Call the frequency and time records in the aging test, identify the fluctuation direction changes of three adjacent points, extract the direction mutation and duration, mark the fluctuation segments, and generate the frequency turning segment identification group;

[0008] S2: calling the frequency turning segment identification group, extracting the fluctuation boundary, merging the continuous segments with the same offset direction, removing the unsustainable fluctuation content, reorganizing and marking the direction, and generating a test integration segment distribution map;

[0009] S3: calling the test normalization segment distribution map, extracting the first and last frequency records of each segment, constructing a frequency extension trend path according to the time series, extracting the intersection and reversal points between any two paths, determining whether a structure jump is formed, performing structure identification on the jump point, and forming a frequency response turning structure marker table;

[0010] S4: calling the frequency response turning structure mark table, determining whether the frequency segment trend between boundaries maintains continuous change, archiving the continuous offset segments, removing the interrupted change content, and generating a trend structure archiving preparation list;

[0011] S5: Call the archiving preparation list, extract the frequency and time value combination of each segment and write it into the archiving sequence, control the upper limit of the archiving area, and remove the tail segment when it exceeds the upper limit, so as to obtain the oscillator aging characteristic detection plan.

[0012] As a further solution of the present invention, the frequency turning paragraph identification group includes the direction change point position, fluctuation direction mark, duration length, fluctuation paragraph number, and paragraph time interval; the test integration segment distribution diagram includes a unified direction segment sequence, time merging order, offset direction mark, continuous segment number, and fluctuation interruption elimination mark; the frequency response turning structure mark table includes the trend path intersection position, direction reversal point position, structural jump mark, trend connection relationship, and structural segment number; the trend structure archiving preparation list includes the integration segment sequence number, trend consistency label, continuous offset mark, interruption segment judgment result, and archiving mark status; the oscillator aging characteristic detection scheme includes archiving frequency sequence, frequency-time combination value, number of archive segments, edge segment elimination rules, and archive content sorting structure.

[0013] As a further solution of the present invention, the specific steps of S1 are:

[0014] S101: Call the frequency record sequence and corresponding time nodes collected during the aging test, construct a continuous three-point frequency triplet, calculate the frequency difference between two adjacent segments in sequence, determine whether the difference direction is consistent, screen out the location of the direction mutation point, and generate a direction mutation index position group;

[0015] S102: Obtaining time nodes between adjacent mutation points based on the directional mutation index position group, calling frequency record values within the segment, dividing the fluctuation segments according to directional consistency, screening continuous directional segments, and calculating segment duration indicators, extracting the start and end nodes and time span of each fluctuation behavior segment, and generating a continuous fluctuation behavior duration group;

[0016] S103: According to the continuous fluctuation behavior duration group and the direction mutation index position group, the start and end time of the fluctuation segment and the mutation point position are matched, the frequency change direction and the segment number are extracted, a number set of the frequency change period is constructed, and a frequency turning segment identification group is generated.

[0017] As a further solution of the present invention, the specific steps of S2 are:

[0018] S201: Calling the frequency turning segment identification group to extract the start and end time nodes of all fluctuation periods, arranging the segments in chronological order according to the start time, calculating the time interval index of the start and end nodes of adjacent segments, and comparing whether the offset directions are consistent. A set of segments with the same consecutive offset directions and a time interval that does not exceed the segment continuation threshold is selected to generate a direction-merged time interval group;

[0019] S202: Based on the directional merged time interval group, positions where the time difference between adjacent segments exceeds the segment continuation threshold are identified, segments with fluctuation interruptions are marked, and segment intervals whose fluctuation duration does not reach the fluctuation continuation threshold are removed and removed from the merged segments. Continuous fluctuation segments that meet the merge condition after removal are obtained to generate a continuous segment integration sequence;

[0020] S203: According to the time boundaries and offset directions of the fragments in the continuous segment integration sequence, the remaining fragments are sequentially assigned direction identifiers and regrouped according to direction categories, the distribution positions of the fragments on the time axis are reconstructed, and a time distribution graphic structure of the offset direction classified fragments is established to generate a test integration segment distribution diagram.

[0021] As a further solution of the present invention, the time duration deviation index calculation formula is specifically:

[0022]

[0023] Where, ΔT i represents the time duration deviation index of the i-th segment, Indicates the start time of the i-th segment, Indicates the end time of the i-th segment, n is the total number of all segments, represents the average of all segment start times, Indicates the average of all segment end times.

[0024] As a further solution of the present invention, the specific steps of S3 are:

[0025] S301: Calling the test normalization segment distribution graph, extracting the frequency record values corresponding to the start time node and the end time node of the segment, connecting the frequency change records of the segment boundaries in chronological order, sequentially constructing the frequency extension trend paths of the segments, and numbering all trend paths according to the segment numbers to generate a trend path number sequence;

[0026] S302: Calculate the coordinates of the intersection of any two trend paths on the time axis and the frequency axis based on the continuity of the frequency paths in the trend path number sequence, select locations where the frequency change direction is reversed before and after the intersection, record the handover time of the reversal point in time sequence, and extract the frequency value at the reversal point to generate a trend handover critical point bit group;

[0027] S303: Based on the trend intersection critical point group, extract the frequency change value and time length before and after the frequency change direction is reversed, calculate the comprehensive index of the frequency reversal ratio and the time domain mutation rate, determine whether it exceeds the structural jump threshold, mark the structural mutation attribute of the intersection that meets the jump condition, assign a structural type code, and establish a frequency response turning structure marking table.

[0028] As a further solution of the present invention, the formula for calculating the comprehensive index of frequency reversal ratio and time domain mutation rate is specifically:

[0029]

[0030] Among them, v Represents the comprehensive index of the frequency reversal ratio and time domain mutation rate of the intersection point v, Indicates the frequency change value of the segment after the intersection point v, Indicates the frequency change value of the segment before the intersection point v, Indicates the time length of the segment after the intersection point v, Indicates the time length of the segment before the intersection point v, where v is the serial number of the current trend reversal intersection point.

[0031] As a further solution of the present invention, the specific steps of S4 are:

[0032] S401: calling the frequency response turning structure marker table, extracting the time node corresponding to each structure boundary, dividing the frequency record segments between the boundaries in sequence, extracting the frequency change values and time series within the segments segment by segment, calculating the average frequency change value after sorting by time axis, and generating a trend direction continuity sequence;

[0033] S402: Based on the trend direction continuity sequence, identify a set of segments with consistent trend directions and the locations of segments where the direction is interrupted, assign a unified filing label to the directional continuous segments, record the numbered positions of the interrupted segments in the trend sequence and mark them with exclusion flags, and generate a trend segment label classification result;

[0034] S403: According to the trend segment label classification results, extract the segment time segment corresponding to the archiving label, summarize the segment frequency change path and the corresponding time boundary, remove the interrupted segments with exclusion marks, integrate the segments that meet the trend continuous change conditions to form a list, and establish a trend structure archiving preparation list.

[0035] As a further solution of the present invention, the calculation formula of the average frequency change value is specifically:

[0036]

[0037] Calculate the average frequency change value of each segment, sort it by time axis, and calculate the average frequency change value to generate a trend direction continuity sequence;

[0038] Where Δf k represents the average frequency change value of the kth segment, f i Represents the frequency value of the i-th measurement, f i-1 represents the frequency value of the i-1th measurement, λ irepresents the weight of the i-th measurement, n k represents the number of measurements in the kth segment, and Σ represents the summation symbol.

[0039] As a further solution of the present invention, the specific steps of S5 are:

[0040] S501: calling the trend structure archiving preparation list, extracting the start and end time nodes and frequency record values corresponding to the organized segments, combining the frequency values and time values into a double sequence format according to the internal time sequence of each segment, and writing them into a preset position in the archiving sequence record area to generate an archiving segment combination sequence group;

[0041] S502: Based on the archived segment combination sequence group, the number of currently written segments is calculated, and the number is compared with the upper limit of the segment space that can be accommodated in the archive area to determine whether the total number of segments exceeds the maximum limit. If so, the last several segments are removed according to the number sorting to generate an archive capacity limit result set;

[0042] S503: Based on the fragment combination content retained in the archive capacity limitation result set, the recording order is rearranged according to the fragment number in chronological order, the frequency value change trajectory within the fragment is integrated and written into the logical position area of the archive sequence, and an oscillator aging characteristic detection scheme is established.

[0043] Compared with the prior art, the advantages and positive effects of the present invention are:

[0044] In the present invention, by extracting the frequency fluctuation direction and mutation points, accurately locating the frequency change trend, identifying continuous offset and turning behaviors, combining trend path extension and intersection extraction, the ability to capture frequency response structure changes is enhanced, retaining fragments with consistent directions, eliminating interruption interference, and achieving high-density consolidation of frequency data. The fragments are archived according to trends and capacity constraints are set, thereby improving the trend recognition accuracy and response structure expression ability in aging behavior, and strengthening the analysis depth and reliability judgment support of the performance evolution process. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0046] Figure 1 Schematic diagram of the steps of the present invention. DETAILED DESCRIPTION

[0047] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0048] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as an "exemplary" in the present invention should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of the word "exemplary" is intended to present concepts in a concrete manner. Furthermore, in the embodiments of the present invention, "and / or" can mean both or either of the two.

[0049] In the embodiments of the present invention, the terms "image" and "picture" may be used interchangeably. It should be noted that, when the distinction between them is not emphasized, their intended meanings are the same. The terms "of," "corresponding," and "corresponding" may be used interchangeably. It should be noted that, when the distinction between them is not emphasized, their intended meanings are the same.

[0050] In the embodiments of the present invention, sometimes a subscript such as W1 may be written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0051] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0052] See also Figure 1 , an automatic testing method for oscillator aging characteristics, comprising the following steps:

[0053] S1: Recall the frequency record sequence and time nodes collected during the aging test, identify the rising and falling directions of the frequency fluctuations between three adjacent points, determine whether the direction change points constitute continuous fluctuation behavior, extract the direction mutation period and the duration of the fluctuation behavior, mark the segment position and integrate it into the fluctuation segment record list, and generate the frequency turning segment identification group;

[0054] S2: Call the frequency turning segment identification group, extract all fluctuation period boundaries, merge the consecutive periods with the same offset direction in chronological order, identify the location of fluctuation interruptions and remove the segments that did not continue to form fluctuations, reorganize the positions of the retained segments and assign offset direction identifications, classify them into test segments with the same direction, and generate a test integration segment distribution map;

[0055] S3: Call the test normalization segment distribution map, extract the first and last record contents of the segment, construct an extended trend line based on the time sequence of the frequency change path, extract the intersection points and direction reversal points of any two trend paths, determine whether the trend intersection position constitutes a structural jump, perform structural identification actions on the intersection positions where the jump is established, and generate a frequency response turning structure marker table;

[0056] S4: Call the frequency response turning structure tag table, extract each set structure boundary, perform trend consistency processing on the frequency record segments between the boundaries, determine whether the trend direction maintains continuous change, assign archiving tags to the continuous offset behavior segments, perform retention and exclusion judgment on the change interruption segments, sort out the consolidated segments that meet the continuous offset conditions, and generate a trend structure archiving preparation list;

[0057] S5: Call the trend structure archiving preparation list, extract the listed organized segments, perform combination processing on the frequency record and time corresponding value in each segment, write the combination result into the archiving sequence record area, compare the cumulative number of segments in the archiving area with the upper limit of space, and when the number of segments reaches the upper limit, remove the edge segments with the lower number, maintain the archive content in sequence, and obtain the oscillator aging characteristic detection plan.

[0058] The frequency turning segment identification group includes the direction change point position, fluctuation direction mark, duration length, fluctuation segment number, and segment time interval. The test integration segment distribution diagram includes the unified direction segment sequence, time merging order, offset direction mark, continuous segment number, and fluctuation interruption elimination mark. The frequency response turning structure mark table includes the trend path intersection position, direction reversal point position, structural jump mark, trend connection relationship, and structure segment number. The trend structure archiving preparation list includes the integration segment sequence number, trend consistency label, continuous offset mark, interruption segment judgment result, and archiving mark status. The oscillator aging characteristic detection plan includes the archiving frequency sequence, frequency-time combination value, number of archived segments, edge segment elimination rules, and archiving content sorting structure.

[0059] The specific steps of S1 are:

[0060] S101: Call the frequency record sequence and corresponding time nodes collected during the aging test, construct a continuous three-point frequency triplet, calculate the frequency difference between two adjacent segments in sequence, determine whether the difference direction is consistent, screen out the location of the direction mutation point, and generate a direction mutation index position group;

[0061] The frequency record sequence and corresponding time nodes collected during the aging test are called, that is, the frequency record values are read sequentially from the sampling system. For example, the frequency is recorded every 5 seconds to form a set of frequency sequences arranged in ascending time order, such as [59.92, 59.94, 59.90, 59.87, 59.91, 59.93, 59.89, 59.85]. The time nodes corresponding to each frequency value are 0 seconds, 5 seconds, 10 seconds, 15 seconds, 20 seconds, 25 seconds, 30 seconds, and 35 seconds. After the reading is completed, three-point continuous frequency triplets are constructed one by one starting from the first item, that is, (59.92, 59.94, 59.90), (59.94, 59.90, 59.87), and (59.90, 59.87, 59.91), and the window is sliding backward to construct until the countdown. Up to the third item, calculate the difference between the first two frequency values and the last two frequency values in each group of triples in turn. For example, the difference of the first group of triples is 0.02 and -0.04, the difference of the second group of triples is -0.04 and -0.03, and the third group is -0.03 and 0.04. For each group of differences, judge whether their directions are consistent. The specific method is to compare the two differences with 0 respectively. If one is positive and the other is negative, it is regarded as a directional mutation point, and the sequence position of the middle value of the group is recorded as the mutation index position. For example, the direction of the first group of triples is inconsistent, and the time point 5 seconds corresponding to the middle value 59.94 is recorded as the mutation point. The direction of the second group of triples is consistent and ignored. The direction of the third group of triples is inconsistent and 20 seconds is recorded as the mutation point. After processing all triplets in turn, a directional mutation index position group is formed, such as [5 seconds, 20 seconds, 30 seconds].

[0062] S102: Obtaining the time nodes between adjacent mutation points based on the directional mutation index position group, calling the frequency record value within the segment, dividing the fluctuation segments according to directional consistency, screening the continuous directional segments, and calculating the segment duration index. Extracting the start and end nodes and time span of each fluctuation behavior segment, and generating a continuous fluctuation behavior duration group;

[0063] According to the direction mutation index position group [5 seconds, 20 seconds, 30 seconds], the time period between adjacent mutation points is extracted. For example, the first segment is from 5 seconds to 20 seconds, and the second segment is from 20 seconds to 30 seconds. All frequency record values are called in each time period. The frequency sequence in the first segment is [59.94, 59.90, 59.87, 59.91], and in the second segment it is [59.91, 59.93, 59.89]. The difference between the two adjacent items in the first segment is -0.04, -0.03, and +0.04, respectively. Judging by the direction sign, the first and second difference directions are consistent, and the third direction is different. Therefore, the first segment is divided into two segments, namely [59.94, 59.90, 59.87] with the same direction and [59.91, 59.93, 59.89] with the same direction. For the direction change [59.87, 59.91], the start and end time and duration of each segment are recorded as 5 seconds to 15 seconds, lasting 10 seconds, and 15 seconds to 20 seconds, lasting 5 seconds; the difference of the second segment is +0.02 and -0.04, and the direction changes suddenly. It is divided into two segments [59.91, 59.93] and [59.93, 59.89] from 20 seconds to 30 seconds, lasting 5 seconds respectively. After recording the start and end time nodes and duration of each segment of the fluctuation behavior, a continuous fluctuation behavior duration group [(5 seconds, 15 seconds, 10 seconds), (15 seconds, 20 seconds, 5 seconds), (20 seconds, 25 seconds, 5 seconds), (25 seconds, 30 seconds, 5 seconds)] is generated, and each segment is assigned a segment number such as S1, S2, S3, S4, and recorded accordingly in the table structure.

[0064] S103: Match the start and end times of the fluctuation segments with the positions of the mutation points based on the continuous fluctuation behavior duration group and the direction mutation index position group, extract the frequency change direction and segment number, construct a number set of the frequency change period, and generate a frequency turning segment identifier group;

[0065] According to the start and end time of each fluctuation segment in the continuous fluctuation behavior duration group and the direction mutation index position group [5 seconds, 20 seconds, 30 seconds], for each fluctuation segment, determine whether its starting point or end point is adjacent to or coincides with a certain mutation point. If the end time of the fluctuation segment is equal to the time of a certain mutation point, it is considered that the segment and the mutation point form a set of frequency turning behaviors, and the frequency change direction corresponding to the fluctuation segment is extracted, that is, the start value and the end value of the segment are compared. If the end value is greater than the start value, it is in an upward direction, otherwise it is in a downward direction. For example, the starting value of the segment (5 seconds, 15 seconds, 10 seconds) is 59.94, the end value is 59.87, and the direction is downward. The starting value of the segment (15 seconds, 20 seconds, 5 seconds) is 59.87, the end value is 59.91, and the direction is upward. Therefore, these two segments are recorded as (1,↓,10 seconds) and (2,↑,5 seconds) respectively. Since the end time of the first segment is 15 seconds and is not in the mutation point, only the end time of the second segment is 20 seconds and appears in the mutation index group. Therefore, only the second segment is taken as the frequency turning segment, and the record number is 2, the direction is rising, and the duration is 5 seconds. Similarly, the direction of the segment (20 seconds, 25 seconds, 5 seconds) is rising, and the direction of the segment (25 seconds, 30 seconds, 5 seconds) is falling. Only the end time of segment 4 is in the mutation point, and it is extracted as the frequency turning segment number 4, the direction is falling, and the duration is 5 seconds. The final frequency turning segment identification group is [(2,↑,5 seconds),(4,↓,5 seconds)], where the number represents the fluctuation segment number, the arrow is the change direction, and the duration is the duration of the segment. In terms of threshold setting, if it is necessary to avoid misjudgment caused by noise when judging mutation points, the minimum amplitude threshold for difference direction mutation detection can be set to 0.03Hz, that is, only when the difference is reversed and the amplitude exceeds 0.03Hz can it be judged as a valid mutation point. The setting basis is that fluctuations within the range of ±0.01Hz under the stable accuracy of the sampling system are measurement fluctuations. When using this threshold, if a certain difference is 0.02 and -0.01, it is not considered a mutation. Only when it is +0.04 and -0.05 are it identified as a mutation point.

[0066] The specific steps of S2 are:

[0067] S201: Calling the frequency turning segment identification group to extract the start and end time nodes of all fluctuation periods, chronologically arranging the segments by start time, calculating the time interval index of the start and end nodes of adjacent segments, and comparing whether the offset directions are consistent. Selecting a set of segments with the same consecutive offset direction and a time interval that does not exceed the segment continuation threshold, and generating a direction-merged time interval group;

[0068] The calculation formula of the time duration deviation indicator is as follows:

[0069]

[0070] Where, ΔT i represents the time duration deviation index of the i-th segment, Indicates the start time of the i-th segment, Indicates the end time of the i-th segment, n is the total number of all segments, represents the average of all segment start times, represents the average of all segment end times;

[0071] The time boundaries of the five segments collected in the frequency transition segment identification group are: (12 seconds, 20 seconds), (22 seconds, 30 seconds), (31 seconds, 38 seconds), (40 seconds, 50 seconds), and (53 seconds, 60 seconds). These time values are automatically marked by the frequency record timestamp information of the aging test system. The sampling frequency is 1Hz, and the sampling system provides data with an accuracy of 0.01 seconds. The start time and end time fields in the system sampling record are extracted and

[0072] Take segment i=2 as an example, its starting time is seconds, and the end time is Seconds, calculate the duration of the fragment itself:

[0073]

[0074] All segments start at Find the average:

[0075]

[0076] All segments end at Find the average:

[0077]

[0078] Calculate the absolute value of the deviation between the current segment and the average start and end time:

[0079]

[0080] Calculate the sum of the duration of all segments:

[0081]

[0082] Substituting into the formula:

[0083]

[0084] The result shows that the time duration deviation index of the second frequency segment is 22.57 seconds. This value is used to compare with the segment continuation threshold. If the threshold is set to 25 seconds, the segment meets the continuation condition and is marked as a merge candidate segment.

[0085] S202: Based on the direction-based merging of time interval groups, positions where the time difference between adjacent segments exceeds the segment continuation threshold are identified, segments with fluctuation interruptions are marked, and segment intervals whose fluctuation duration does not reach the fluctuation continuation threshold are removed and removed from the merged segments. Continuous fluctuation segments that meet the merging conditions after removal are obtained to generate a continuous segment integration sequence;

[0086] Based on the direction merge time interval group [(25 seconds, 35 seconds, ↑)], identify the time difference between adjacent segments. If there are more than one segment, compare the difference between the end time of the current segment and the start time of the next segment in turn to determine whether the time difference is greater than the segment continuation threshold of 10 seconds. In this example, there is only one segment, so there is no time interval limit problem. If there are multiple direction merge segments, such as (0 seconds, 10 seconds, ↓) and (22 seconds, 35 seconds, ↑), then the time difference of (10 seconds, 22 seconds) is 12 seconds greater than 10 seconds, mark it as a segment with fluctuation interruption, and then calculate the duration of each merged segment, that is, the end time minus the start time, such as ( If the duration is less than the fluctuation duration threshold, the segment will be removed. The fluctuation duration threshold is set to 6 seconds. This value is determined by the sampling system resolution and the length of short-term disturbances in the frequency stable state. If the frequency fluctuation does not last more than 6 seconds, it is regarded as a transient disturbance rather than a trend segment and is not accepted. In the example, the merged segment (25 seconds, 35 seconds) lasts for 10 seconds and is greater than 6 seconds, so it is not removed. If a new segment (40 seconds, 45 seconds, ↓) is added, lasting 5 seconds, which is less than the threshold, the segment is deleted from the merged sequence. The segment finally retained is the merged segment segment that meets the continuity and duration requirements, generating a continuous segment normalization sequence [(25 seconds, 35 seconds, ↑)].

[0087] S203: Based on the time boundaries and offset directions of the segments in the continuous segment integration sequence, the remaining segments are sequentially assigned direction identifiers and regrouped by direction category, the distribution positions of the segments on the time axis are reconstructed, a time distribution graph structure of the offset direction classified segments is established, and a test integration segment distribution graph is generated;

[0088] According to the time boundary and offset direction of each segment in the continuous segment sorting sequence [(25 seconds, 35 seconds, ↑)], the segments are assigned direction identifiers in sequence. If the segment direction is ↑, it is marked as an "ascending segment" and ↓ as a "descending segment". By calling the offset direction field in each segment record, the value is directly assigned. Then all segments are grouped according to the direction type. All the segments in the ascending direction are sorted and combined into a group in chronological order. For example, they are grouped into an ascending group [(25 seconds, 35 seconds)]. If there are multiple descending direction segments, a descending group is formed. The time order of the segments is not changed within each group. After all the grouping is completed, the time axis is displayed. Reconstruct according to the start and end time of each group of segments. For example, if the time axis is from 0 seconds to 60 seconds, and the rising group segment is between 25 seconds and 35 seconds, then the interval position is drawn as an upward trend in the graphic structure, and the direction arrow is marked upward. If the direction of another segment (40 seconds, 50 seconds) is downward, it is drawn at the corresponding position on the time axis and marked downward, forming a segment distribution graph with time as the horizontal axis and frequency offset direction as the trend indicator. The graphic structure is a two-dimensional broken line or columnar segment block. The horizontal axis scale is segmented at intervals of 5 seconds. The vertical axis is not marked with a frequency scale, only indicating the offset direction classification, so as to establish a test integration segment distribution graph.

[0089] The specific steps of S3 are:

[0090] S301: Call the test normalization segment distribution map, extract the frequency record values corresponding to the start time node and the end time node of the segment, connect the frequency change records of the segment boundaries in chronological order, sequentially construct the frequency extension trend paths of the segments, and number all trend paths according to the segment number to generate a trend path number sequence;

[0091] Call the test normalization segment distribution graph, read the start time and end time nodes of each normalization segment in turn, for example, the segments marked in the distribution graph are (25 seconds, 35 seconds, ↑), (40 seconds, 50 seconds, ↓), extract the frequency values corresponding to the start time 25 seconds and the end time 35 seconds of the segment from the frequency record sequence, assuming that the corresponding frequency values are 59.85Hz and 60.02Hz, then the frequency change of this segment is rising, and then read the frequency values corresponding to the start and end time 40 seconds and 50 seconds of the next segment as 60.02Hz and 59.74Hz, forming a descending path, and after calling the frequency values corresponding to the segment boundary time, connect the boundary frequency values of each segment into a broken line segment path in chronological order, forming a path composed of frequency value points connected in series. The extended trend path constructed, for the first segment, its path structure is connected from (25 seconds, 59.85Hz) to (35 seconds, 60.02Hz), and the second segment is connected from (40 seconds, 60.02Hz) to (50 seconds, 59.74Hz). A broken line trend is formed by continuously arranging the boundary coordinates of each segment in the graphic data structure. When constructing each path segment, the corresponding integrated segment number of the path segment is recorded, and the paths are numbered in sequence according to the order in which the integrated segments appear in the graph. For example, the first path is numbered T1, the second is T2, and they are successively assigned T3, T4 and other labels. Finally, the constructed frequency extended trend path set is output, and the trend path number sequence is [T1, T2].

[0092] S302: Based on the continuity of the frequency paths in the trend path number sequence, the coordinates of the intersection points of any two trend paths on the time axis and the frequency axis are calculated. Positions where the frequency change direction is reversed before and after the intersection are screened. The handover time of the reversal point is recorded in time sequence, and the frequency value at the reversal point is extracted to generate a trend handover critical point bit group.

[0093] According to the starting and ending points of each two adjacent paths in the trend path number sequence [T1, T2], extract the end time of path one and the starting time of path two, that is, take the end point of T1 (35 seconds, 60.02Hz) and the starting point of T2 (40 seconds, 60.02Hz), and determine whether they form a connection boundary on the time axis. Then continue to extract the starting and end points of T1 and the starting and end points of T2, draw the extension lines of these two paths on the linear coordinate system, and calculate whether there is an intersection between the two paths, that is, determine whether the end frequency of path one is equal to the starting frequency of path two, and the previous one is rising and the next one is falling. If this condition is met, it is Intersection point, record the intersection time and frequency value. Here, the intersection point is (40 seconds, 60.02 Hz). If the trend direction of the path before and after this point reverses, that is, from rising to falling or from falling to rising, the direction change is determined by comparing the direction fields of the previous and next segments. If the previous segment is ↑ and the next segment is ↓, or vice versa, it is considered a reversal. The time point of the reversal point is recorded as the handover time, and the intersection frequency is recorded as the handover frequency value. All adjacent path segments are processed in the order of path numbers. A set of records is generated for each trend intersection point, and finally a trend intersection critical point bit group is formed, such as [(40 seconds, 60.02 Hz)].

[0094] S303: Based on the trend intersection critical point group, the frequency change value and time length before and after the frequency change direction reversal are extracted, the frequency reversal ratio and the time domain mutation rate comprehensive index are calculated, and it is determined whether the structural jump threshold is exceeded. The intersection points that meet the jump condition are marked with structural mutation attributes, and a structural type code is assigned to establish a frequency response turning structure marking table;

[0095] The calculation formula of the comprehensive index of frequency reversal ratio and time domain mutation rate is as follows:

[0096]

[0097] Among them, v Represents the comprehensive index of the frequency reversal ratio and time domain mutation rate of the intersection point v, Indicates the frequency change value of the segment after the intersection point v, Indicates the frequency change value of the segment before the intersection point v, Indicates the time length of the segment after the intersection point v, Indicates the time length of the segment before the intersection point v, where v is the serial number of the current trend reversal intersection point;

[0098] The frequency path at the intersection point v=3 is the reversal point. The frequency recording data is obtained through a 1Hz frequency recording system with a frequency recording accuracy of 0.01Hz and a timestamp accuracy of 0.1 seconds. The start and end times of the two consecutive segments before and after the reversal point v are extracted from the sampling log, which are 37.5 seconds to 41.0 seconds and 41.0 seconds to 46.0 seconds respectively. The corresponding frequency start and end values are 60.12Hz to 59.80Hz and 59.80Hz to 60.28Hz respectively. By calculating the frequency change value and They are

[0099]

[0100] Calculate the corresponding time length and The start and end time interval of the segment is

[0101]

[0102] Substitute the above parameters into the formula for calculation:

[0103]

[0104] Calculate the numerator:

[0105] (0.48-0.32) 2 =(0.16) 2 =0.0256;

[0106] Calculate the first part of the denominator:

[0107] (5.0+3.5)=8.5;

[0108] Compute the product difference term:

[0109] 0.32·5.0=1.60;

[0110] 0.48·3.5=1.68;

[0111] |1.60-1.68|=0.08;

[0112] Take the square root after multiplying the product:

[0113] 8.5·0.08=0.68;

[0114]

[0115] Final calculation:

[0116]

[0117] The results show that the frequency jump joint index at the reversal point v=3 is 0.0310, which is used to compare with the structural jump identification threshold. For example, if the system sets the threshold to 0.025, then 0.0310 exceeds the threshold, and the intersection is marked as a structural jump point and included in the frequency response turning structure mark table record, with a structure type code for trend structure identification.

[0118] The specific steps of S4 are:

[0119] S401: Call the frequency response turning structure marker table, extract the time node corresponding to each structure boundary, divide the frequency record segments between the boundaries in sequence, extract the frequency change value and time series within each segment, sort them according to the time axis, calculate the average frequency change value, and generate a trend direction continuity sequence;

[0120] The calculation formula for the average frequency change value is as follows:

[0121]

[0122] Calculate the average frequency change value of each segment, sort it by time axis, and calculate the average frequency change value to generate a trend direction continuity sequence;

[0123] Where Δf k represents the average frequency change value of the kth segment, f i Represents the frequency value of the i-th measurement, f i-1 represents the frequency value of the i-1th measurement, λ i represents the weight of the i-th measurement, n k represents the number of measurements in the kth segment, and Σ represents the summation symbol;

[0124] Parameter details:

[0125] Δf k It is obtained by calculating the weighted average of all frequency changes in each frequency record segment, with weight λ i The allocation may be based on the confidence level of the measurement or the measurement interval. For example, shorter time intervals or higher measurement accuracy may be given greater weight. i and f i-1 is the continuously measured frequency value, extracted directly from the frequency record data. k is the number of data points in each segment, which is determined by the actual number of measurements. i Assumed to be 1 unless specific circumstances require adjustment to reflect the varying importance of data points.

[0126] Calculation example:

[0127] Suppose there are three measurements within a certain segment, with frequency values of 59.95 Hz, 60.00 Hz, and 59.90 Hz. For these values, the difference between the consecutive measurements is |60.00 - 59.95| = 0.05 Hz and |59.90 - 60.00| = 0.10 Hz. If all measurements have a weight of 1, the average frequency change is calculated as:

[0128]

[0129] The results show that the average frequency change in the investigated segment is 0.106 Hz. This result is used to further determine the trend and continuity of the frequency change and to decide whether the change in the segment should be classified into a specific trend category.

[0130] S402: Based on the trend direction continuity sequence, identify the set of segments with consistent trend directions and the locations of segments where the direction is interrupted, assign a unified filing label to the directional continuous segments, record the numbered positions of the interrupted segments in the trend sequence and mark them with exclusion flags, and generate trend segment label classification results;

[0131] Based on the established trend direction continuity sequence, such as [↓,↓,↓], the direction is judged item by item to see if it is consistent. The direction of the first item is compared with the second item. If the same, they are classified into the same continuous segment set. The direction continuity judgment action is completed by comparing the string values item by item. If the two direction tags are equal, they are assigned the same filing label. For example, if the direction of the first and second items are both downward, the filing label is G1. Continue to judge whether the direction of the second and third items is consistent. If they are still downward, they are classified into G1. The same process is applied to all subsequent items. If a direction is different from the previous one, Consistent, for example, if the direction sequence is [↓,↓,↑], the third item is marked as a direction interruption fragment, and its number position in the trend sequence is recorded as serial number 3. At the same time, an exclusion mark is added to the interruption item, and the exclusion mark is set to X mark, indicating that it is not included in the continuous trend analysis set. The final output is the trend fragment label classification result, which includes the fragment sequence number, start and end time, frequency direction, archive label or exclusion mark. The example result is [(1, 40 seconds-45 seconds,↓,G1), (2, 45 seconds-50 seconds,↓,G1), (3, 50 seconds-55 seconds,↑,X)].

[0132] S403: Based on the trend segment label classification results, extract the segment time segments corresponding to the archiving labels, summarize the segment frequency change paths and corresponding time boundaries, remove interrupted segments with exclusion marks, integrate the segments that meet the trend continuous change conditions to form a list, and establish a trend structure archiving preparation list;

[0133] According to the trend segment label classification results, extract all segment information without exclusion marks, that is, filter out the segments with the archive label G1 or other valid labels. In the example, extract the segments (1, 40 seconds-45 seconds, ↓, G1) and (2, 45 seconds-50 seconds, ↓, G1), and obtain the corresponding start time, end time and frequency path from them. For example, the frequency of the first segment drops from 60.02Hz to 59.93Hz, and the frequency of the second segment drops from 59.93Hz to 59.84Hz. Extract the start and end time points for each valid segment respectively. , the start and end frequency values are marked with the corresponding numbers, and the segments with exclusion marks (3, 50 seconds-55 seconds) are eliminated and not included in the structure filing. When integrating the remaining valid segments, they are arranged in chronological order. In the final list, each segment number, start and end time, frequency change amplitude and direction mark are listed in turn to form a trend structure filing preparation list. The example list is recorded as: [(T1, 40 seconds-45 seconds, 60.02Hz→59.93Hz,↓), (T2, 45 seconds-50 seconds, 59.93Hz→59.84Hz,↓)].

[0134] The specific steps of S5 are:

[0135] S501: Call the trend structure archiving preparation list, extract the start and end time nodes and frequency record values corresponding to the consolidated segments, combine the frequency values and time values into a double sequence format according to the internal time sequence of each segment, and write them into a preset location in the archiving sequence record area to generate an archiving segment combination sequence group;

[0136] Call the trend structure archiving preparation list, read the number, start and end time nodes and corresponding frequency change record values of each valid integration segment in turn, and extract all frequency record points in ascending time order for each segment. For example, if the start and end time of segment T1 is 40 seconds to 45 seconds, the corresponding frequency points are [(40 seconds, 60.02Hz), (42 seconds, 59.97Hz), (45 seconds, 59.93Hz)]. Then, after sorting the sequence by time, combine the time value sequence [40, 42, 45] and the frequency value sequence [60.02, 59.97, 59.93] into a set of double sequence structures. The double sequence structure format is "time series + frequency sequence" and is written to the segment position set in the archive sequence record area in a fixed format. For example, if each set of records in the archive area is set to occupy 10 sets of data space, then the T1 data is written to positions 1 to 3. If the start and end time of the subsequent segment T2 is 45 seconds to 50 seconds, the corresponding frequency points are [(45 Seconds, 59.93Hz), (47 seconds, 59.89Hz), (50 seconds, 59.84Hz)], then the double sequences [45, 47, 50] and [59.93, 59.89, 59.84] are formed in the same way and written to the next free position in the archive sequence record area. The writing position order is aligned according to the fragment number order, that is, T1 is in position 1, T2 is in position 2, and they increase in sequence. During the writing operation, it is necessary to check whether there are duplicate values in the time value. If there are duplicate values, the earlier record is used to avoid overwriting. At the same time, it is confirmed that all frequency values are not missing or exceed the physical frequency range setting. For example, if the detection frequency value range is limited to [59.0Hz, 61.0Hz], if the detection frequency points 59.93Hz and 60.02Hz are both within the valid range, then the writing is confirmed. If a frequency record value is 61.25Hz, then the abnormal value exclusion record is removed and marked. After this step is completed, the archive fragment combination sequence group after all fragments are written is generated.

[0137] S502: Based on the archived segment combination sequence group, the number of currently written segments is calculated and compared with the upper limit of the segment space that can be accommodated in the archive area to determine whether the total number of segments exceeds the maximum limit. If so, the last several segments are removed according to the numbering order to generate an archive capacity limit result set;

[0138] Based on the archived segment combination sequence group that has been written, the number of currently written segments N is called. For example, if the number of segments read is 7, this value is compared with the upper limit Nmax of the maximum number of segments that can be accommodated in the archive area. The value of Nmax is set according to the physical storage limit or the storage capacity limited by the sampling system. In this example, the maximum number of segments is set to 5. If the current number N=7 is greater than 5, a culling operation is performed. The numbers and writing sequence table of all written segments are called, and the first 5 segment numbers are retained in ascending order. The excess segment numbers are sorted, marked as "overlimit" and culled. The culling action is to delete the data of the corresponding number records in the archive sequence record area and release the space at the location. For example, if T6 and T7 exceed the number upper limit, the dual sequence data content of T6 and T7 is deleted from the archive sequence and the write location status is updated to unoccupied. The newly formed fragment combination sequence group is the archive capacity limit result set. The limit result set should contain no more than Nmax segment numbers and the corresponding time and frequency dual sequence record content. For example, after limitation, the retained segments are T1 to T5, and the corresponding record content and sequence remain unchanged in the original archive order.

[0139] S503: Based on the fragment combination content retained in the archive capacity limit result set, the recording order is reordered according to the fragment number in chronological order, the frequency value change trajectory within the fragment is integrated and written into the logical location area of the archive sequence, and the oscillator aging characteristic detection scheme is established;

[0140] According to the archive capacity limit, the fragment combination content retained in the result set is extracted, and the time series and frequency series of each fragment are extracted. The fragment number is used as the primary key to sort them in ascending order. The sorting operation is to arrange each data record in the order of T1 to T5. After the sorting is completed, the frequency change trajectory of each fragment is sorted, that is, the changes between the frequency values in each segment are recorded in sequence into a unidirectional trend structure, such as T1 is [60.02→59.93], T2 is [59.93→59.84]. The recorded content after sorting is a trajectory set of continuous frequency change paths. Subsequently, the archive sequence logical position allocation table is called. The table confirms the starting position and write index of the logical storage block corresponding to each number, and writes the trajectory record to the corresponding logical position. If the T1 logical position is L1, the T1 trajectory data is written to the L1 position according to the format. After all fragments are written in numerical order, a complete frequency trend structure list and logical position mapping table are generated, and finally a structured archiving content for oscillator aging characteristic detection is formed. The established detection plan includes four contents: fragment number, logical position identifier, time boundary, and frequency change trajectory. Each content has a clear value range and record format for subsequent aging trend analysis and retrieval processing.

[0141] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. An automatic test method for oscillator aging characteristics, characterized in that: The following steps are involved: S1: Call the frequency and time records in the aging test, identify the fluctuation direction changes of three adjacent points, extract the direction mutation and duration, mark the fluctuation segments, and generate the frequency turning segment identification group; S2: calling the frequency turning segment identification group, extracting the fluctuation boundary, merging the continuous segments with the same offset direction, removing the unsustainable fluctuation content, reorganizing and marking the direction, and generating a test integration segment distribution map; S3: calling the test normalization segment distribution map, extracting the first and last frequency records of each segment, constructing a frequency extension trend path according to the time series, extracting the intersection and reversal points between any two paths, determining whether a structure jump is formed, performing structure identification on the jump point, and forming a frequency response turning structure marker table; S4: calling the frequency response turning structure mark table, determining whether the frequency segment trend between boundaries maintains continuous change, archiving the continuous offset segments, removing the interrupted change content, and generating a trend structure archiving preparation list; S5: Call the archiving preparation list, extract the frequency and time value combination of each segment and write it into the archiving sequence, control the upper limit of the archiving area, and remove the tail segment when it exceeds the upper limit, so as to obtain the oscillator aging characteristic detection plan.

2. The automatic testing method for oscillator aging characteristics according to claim 1, characterized in that: The frequency turning paragraph identification group includes the direction change point position, fluctuation direction mark, duration length, fluctuation paragraph number, and paragraph time interval; the test integration segment distribution diagram includes a unified direction segment sequence, time merging order, offset direction mark, continuous segment number, and fluctuation interruption elimination mark; the frequency response turning structure mark table includes the trend path intersection position, direction reversal point position, structural jump mark, trend connection relationship, and structure segment number; the trend structure archiving preparation list includes the integration segment sequence number, trend consistency label, continuous offset mark, interruption segment judgment result, and archiving mark status; the oscillator aging characteristic detection scheme includes the archiving frequency sequence, frequency-time combination value, number of archive segments, edge segment elimination rules, and archiving content sorting structure.

3. The automatic testing method for oscillator aging characteristics according to claim 1, characterized in that: The specific steps of S1 are: S101: Call the frequency record sequence and corresponding time nodes collected during the aging test, construct a continuous three-point frequency triplet, calculate the frequency difference between two adjacent segments in sequence, determine whether the difference direction is consistent, screen out the location of the direction mutation point, and generate a direction mutation index position group; S102: Obtaining time nodes between adjacent mutation points based on the directional mutation index position group, calling frequency record values within the segment, dividing the fluctuation segments according to directional consistency, screening continuous directional segments, and calculating segment duration indicators, extracting the start and end nodes and time span of each fluctuation behavior segment, and generating a continuous fluctuation behavior duration group; S103: According to the continuous fluctuation behavior duration group and the direction mutation index position group, the start and end time of the fluctuation segment and the mutation point position are matched, the frequency change direction and the segment number are extracted, a number set of the frequency change period is constructed, and a frequency turning segment identification group is generated.

4. The automatic testing method for oscillator aging characteristics according to claim 1, characterized in that: The specific steps of S2 are: S201: Calling the frequency turning segment identification group to extract the start and end time nodes of all fluctuation periods, arranging the segments in chronological order according to the start time, calculating the time interval index of the start and end nodes of adjacent segments, and comparing whether the offset directions are consistent. A set of segments with the same consecutive offset directions and a time interval that does not exceed the segment continuation threshold is selected to generate a direction-merged time interval group; S202: Based on the directional merged time interval group, positions where the time difference between adjacent segments exceeds the segment continuation threshold are identified, segments with fluctuation interruptions are marked, and segment intervals whose fluctuation duration does not reach the fluctuation continuation threshold are removed and removed from the merged segments. Continuous fluctuation segments that meet the merge condition after removal are obtained to generate a continuous segment integration sequence; S203: According to the time boundaries and offset directions of the fragments in the continuous segment integration sequence, the remaining fragments are sequentially assigned direction identifiers and regrouped according to direction categories, the distribution positions of the fragments on the time axis are reconstructed, and a time distribution graphic structure of the offset direction classified fragments is established to generate a test integration segment distribution diagram.

5. The automatic testing method for oscillator aging characteristics according to claim 1, characterized in that: The calculation formula of the time duration deviation index is specifically as follows: Where, ΔT i represents the time duration deviation index of the i-th segment, Indicates the start time of the i-th segment, Indicates the end time of the i-th segment, n is the total number of all segments, represents the average of all segment start times, Indicates the average of all segment end times.

6. The automatic testing method for oscillator aging characteristics according to claim 1, characterized in that: The specific steps of S3 are: S301: Calling the test normalization segment distribution graph, extracting the frequency record values corresponding to the start time node and the end time node of the segment, connecting the frequency change records of the segment boundaries in chronological order, sequentially constructing the frequency extension trend paths of the segments, and numbering all trend paths according to the segment numbers to generate a trend path number sequence; S302: Calculate the coordinates of the intersection of any two trend paths on the time axis and the frequency axis based on the continuity of the frequency paths in the trend path number sequence, select locations where the frequency change direction is reversed before and after the intersection, record the handover time of the reversal point in time sequence, and extract the frequency value at the reversal point to generate a trend handover critical point bit group; S303: Based on the trend intersection critical point group, extract the frequency change value and time length before and after the frequency change direction is reversed, calculate the comprehensive index of the frequency reversal ratio and the time domain mutation rate, determine whether it exceeds the structural jump threshold, mark the structural mutation attribute of the intersection that meets the jump condition, assign a structural type code, and establish a frequency response turning structure marking table.

7. The automatic testing method for oscillator aging characteristics according to claim 1, characterized in that: The calculation formula of the frequency reversal ratio and time domain mutation rate comprehensive index is specifically as follows: Among them, v Represents the comprehensive index of the frequency reversal ratio and time domain mutation rate of the intersection point v, Indicates the frequency change value of the segment after the intersection point v, Indicates the frequency change value of the segment before the intersection point v, Indicates the time length of the segment after the intersection point v, Indicates the time length of the segment before the intersection point v, where v is the serial number of the current trend reversal intersection point.

8. The automatic testing method for oscillator aging characteristics according to claim 1, characterized in that: The specific steps of S4 are: S401: calling the frequency response turning structure marker table, extracting the time node corresponding to each structure boundary, dividing the frequency record segments between the boundaries in sequence, extracting the frequency change values and time series within the segments segment by segment, calculating the average frequency change value after sorting by time axis, and generating a trend direction continuity sequence; S402: Based on the trend direction continuity sequence, identify a set of segments with consistent trend directions and the locations of segments where the direction is interrupted, assign a unified filing label to the directional continuous segments, record the numbered positions of the interrupted segments in the trend sequence and mark them with exclusion flags, and generate a trend segment label classification result; S403: According to the trend segment label classification results, extract the segment time segment corresponding to the archiving label, summarize the segment frequency change path and the corresponding time boundary, remove the interrupted segments with exclusion marks, integrate the segments that meet the trend continuous change conditions to form a list, and establish a trend structure archiving preparation list.

9. The automatic testing method for oscillator aging characteristics according to claim 1, characterized in that: The calculation formula of the average frequency change value is specifically: Calculate the average frequency change value of each segment, sort it by time axis, and calculate the average frequency change value to generate a trend direction continuity sequence; Where Δf k represents the average frequency change value of the kth segment, f i Represents the frequency value of the i-th measurement, f i-1 represents the frequency value of the i-1th measurement, λ i represents the weight of the i-th measurement, n k represents the number of measurements in the kth segment, and Σ represents the summation symbol.

10. The automatic testing method for oscillator aging characteristics according to claim 1, characterized in that: The specific steps of S5 are: S501: calling the trend structure archiving preparation list, extracting the start and end time nodes and frequency record values corresponding to the organized segments, combining the frequency values and time values into a double sequence format according to the internal time sequence of each segment, and writing them into a preset position in the archiving sequence record area to generate an archiving segment combination sequence group; S502: Based on the archived segment combination sequence group, the number of currently written segments is calculated, and the number is compared with the upper limit of the segment space that can be accommodated in the archive area to determine whether the total number of segments exceeds the maximum limit. If so, the last several segments are removed according to the number sorting to generate an archive capacity limit result set; S503: Based on the fragment combination content retained in the archive capacity limitation result set, the recording order is rearranged according to the fragment number in chronological order, the frequency value change trajectory within the fragment is integrated and written into the logical position area of the archive sequence, and an oscillator aging characteristic detection scheme is established.

Images