A method and system for analyzing the stability of measurement results of a metrological instrument

Abstract

This invention discloses a method and system for stability analysis of metrological instrument measurement results, relating to the field of time series data analysis technology. The method includes measurement data stream alignment, construction of baseline features for measurement result stability, three-state feature decomposition, and stability analysis of metrological instrument measurement results. Repeatability, trend, and structural features of the measurement results are extracted to construct a baseline feature vector for measurement result stability. A three-state feature decomposition method based on baseline constraints is used to decompose the stability changes of the measurement results into drift, noise, and jump features. Based on the baseline features and the three-state feature decomposition results, a comprehensive analysis of the stability state and changing trends of the metrological instrument measurement results is performed. This invention enables refined and causally distinguishable analysis of the stability of metrological instrument measurement results under complex operating conditions and multi-state operation, providing a reliable basis for instrument operating status assessment and measurement result reliability judgment.

Description

Technical Field

[0001] This invention relates to the field of time series data analysis technology, specifically to a method and system for analyzing the stability of measurement results from metrology instruments. Background Technology

[0002] The stability analysis method for measurement results of metrology instruments refers to an analytical method that systematically evaluates the consistency, reliability, and trend of measurement results over time by structuring and analyzing the continuous measurement results output by metrology instruments. This method not only focuses on whether the measurement results fluctuate, but also further distinguishes the different causes of fluctuations to determine the stability and evolution of the measurement results. Through this method, objective and quantifiable analytical basis can be provided for the assessment of the operating status of metrology instruments, the determination of the reliability of measurement results, and subsequent decisions such as calibration and maintenance.

[0003] With the widespread application of electronic measuring instruments in fields such as power metering, flow measurement, and pressure monitoring, the long-term stability of measurement results has become a key factor affecting measurement accuracy and operational reliability. Existing stability analysis methods for measuring instruments are mostly based on simple statistical fluctuation analysis or fixed threshold judgments. These methods typically do not adequately consider issues such as timestamp errors, range switching, calibration status changes, and differences in operating conditions that may occur during data acquisition. This can easily lead to the mixing and analysis of measurement results from different states, resulting in distorted stability assessments.

[0004] Meanwhile, existing methods for analyzing the stability of measurement results of metrology instruments have technical problems. They fail to effectively distinguish between timestamp errors, range switching, calibration status changes, and differences in operating conditions during the measurement data acquisition process. This results in measurement results from different states being mixed and analyzed, making the stability analysis results susceptible to interference from abnormal data and unable to accurately reflect the true stable state.

[0005] In the process of constructing the stability baseline characteristics of existing measurement results, there is a technical problem that the baseline is directly constructed based on the full data or a fixed statistical window without distinguishing between stable and unstable operating intervals. This makes the baseline characteristics easily skewed by transient fluctuations, operating condition changes or abnormal data, thereby reducing the reliability of stability assessment.

[0006] In existing feature decoupling and decomposition processes, there is a technical problem that it is difficult to distinguish whether changes in the stability of measurement results are caused by long-term systematic drift, random noise enhancement, or sudden anomalies, thus failing to provide a clear basis for instrument condition assessment and maintenance decisions. Summary of the Invention

[0007] To address the above issues and overcome the shortcomings of existing technologies, this invention provides a method and system for analyzing the stability of measurement results from metrological instruments. The technical solution adopted by this invention is as follows: This invention provides a method for analyzing the stability of measurement results from metrological instruments, which includes the following steps:

[0008] Step S1: Measure data stream alignment;

[0009] Step S2: Constructing baseline characteristics for the stability of measurement results;

[0010] Step S3: Three-state feature decomposition;

[0011] Step S4: Stability analysis of measurement results from measuring instruments.

[0012] Further, in step S1, the measurement data stream alignment is used to eliminate the sequence inconsistency caused by the measurement results of the measuring instrument under the acquisition time sequence, operating condition switching and operating status change. Specifically, based on the timestamp information corresponding to the measurement results of the measuring instrument, the instrument operating status identifier and the calibration status change information, the continuous measurement results are rearranged in time order, and when the range switching, calibration status change or operating condition switching is detected, the measurement data stream is segmented and aligned to obtain the measurement data stream sequence.

[0013] The measuring instrument specifically refers to an electronic measuring device that can output measurement results in the form of a time series, and has functions such as switching measurement ranges, changing operating conditions, or adjusting calibration status during operation.

[0014] Further, in step S2, the construction of the measurement result stability baseline feature is used to establish a stability reference benchmark that can characterize the inherent statistical structure of the measurement results under stable operating conditions. Specifically, a multidimensional statistical result stability baseline construction method based on stable operating interval constraints is adopted. In the measurement data stream sequence, a measurement interval that meets the stable operating conditions is selected, and multidimensional stability features are extracted from the measurement results within the measurement interval to obtain multidimensional stability features. The multidimensional stability features are then combined to construct the corresponding measurement result stability baseline feature vector.

[0015] The stability features include at least repeatability features for characterizing the repeatability of measurement results, trend features for characterizing the long-term trend of measurement results, and structural features for characterizing the distribution of abnormal structures.

[0016] The method for constructing a stability baseline for multidimensional statistical results based on stable operating interval constraints specifically includes the following steps:

[0017] Step S21: Determining the stable operating interval, which is used to determine the measurement interval that meets the stable operating conditions from the measurement data stream sequence. Specifically, the measurement data stream sequence is divided into intervals according to a preset time window, and the measurement intervals that continuously meet the stability judgment conditions are selected based on the statistical characteristic changes of the measurement results in adjacent intervals as stable operating intervals.

[0018] Step S22: Repeatability feature extraction, used to characterize the repeatability of measurement results within the stable operating range. Specifically, within the stable operating range, a sliding window statistical analysis is performed on the measurement results to calculate the fluctuation amplitude characteristics of the measurement results within each window, and the fluctuation amplitude characteristics are summarized to obtain repeatability features used to characterize the repeatability of the measurement results.

[0019] Step S23: Trend feature extraction, used to characterize the long-term trend of measurement results within the stable operating range, specifically involves performing trend fitting processing on the measurement result sequence within the stable operating range and extracting trend features that reflect the strength of the trend of measurement results changing over time.

[0020] Step S24: Structural feature extraction, used to characterize the abnormal distribution structure of measurement results within the stable operating range, specifically, to analyze the statistical distribution pattern of measurement results within the stable operating range and extract structural features that reflect the degree of skewness of the measurement result distribution, the degree of peak concentration, and the proportion of outliers;

[0021] Step S25: Construction of stability baseline feature vector, used to form a unified reference benchmark for subsequent stability analysis. Specifically, the repeatability feature, trend feature and structural feature are normalized, and the normalized multidimensional stability features are combined to construct the measurement result stability baseline feature vector.

[0022] The baseline feature vector of measurement result stability includes baseline statistical features such as repeatability, trend stability ratio, measurement result distribution skewness, peak concentration, and anomaly proportion.

[0023] Further, in step S3, the three-state feature decomposition is used to structurally distinguish the different causes of stability changes in the measurement results. Specifically, a baseline-constrained three-state feature decomposition method is adopted. Based on the baseline feature vector of the measurement results stability, the measurement data stream sequence is subjected to feature decomposition processing. The change features of the measurement results relative to the baseline feature vector of the measurement results stability are divided into drift features used to characterize long-term systematic changes in the measurement results, noise features used to characterize the random fluctuation amplitude of the measurement results, and jump features used to characterize sudden abnormal changes in the measurement results, thus obtaining three types of stability sub-features that are distinguishable from each other and reflect different causes of stability changes.

[0024] The baseline-constrained tri-state eigenvalue decomposition method specifically includes the following steps:

[0025] Step S31: Baseline consistency processing, used to map the measurement result sequence to a comparable feature space relative to the measurement result stability baseline feature vector. Specifically, based on the baseline statistical features in the measurement result stability baseline feature vector, the measurement data stream sequence is normalized to obtain the measurement result feature sequence relative to the stability baseline.

[0026] Step S32: Drift feature extraction, used to characterize the long-term systematic change characteristics of the measurement results, specifically, smoothing trend estimation processing is performed on the normalized measurement result feature sequence, the trend component of the measurement results changing over time is extracted, and a drift feature is constructed based on the trend component to characterize the long-term systematic change of the measurement results, and the trend component is removed from the measurement result feature sequence to obtain the detrended feature sequence.

[0027] Step S33: Jump feature extraction, used to characterize sudden abnormal change features in the measurement results. Specifically, in the detrended feature sequence, based on the baseline fluctuation scale corresponding to the baseline feature vector of the measurement result stability, threshold determination is performed on the adjacent changes in the measurement result feature sequence to identify the sudden change location, and jump features reflecting the frequency and amplitude of the sudden change are extracted. The sudden change component is removed from the detrended feature sequence to obtain the remaining measurement result feature sequence.

[0028] Step S34: Noise feature extraction, used to characterize the random fluctuation amplitude of the measurement results. Specifically, after drift feature extraction and jump feature extraction, statistical analysis is performed on the fluctuation of the remaining measurement result feature sequence to extract noise features that reflect the random fluctuation amplitude of the measurement results.

[0029] Step S35: Construction of three-state stability sub-features, used to form a set of stability sub-features that are distinct from each other and reflect different causes of stability changes. Specifically, the drift feature, noise feature and jump feature are output as independent stability sub-features, which serve as the input basis for subsequent stability analysis of measurement results of metrology instruments.

[0030] Further, in step S4, the stability analysis of the measurement results of the measuring instrument is used to comprehensively determine the stability state of the measurement results of the measuring instrument based on the stability baseline characteristics and the three-state characteristic decomposition results of the measurement results. Specifically, it analyzes the stability degree and stability change trend of the measurement results of the measuring instrument based on the degree of deviation of the measurement results from the stability baseline characteristic vector of the measurement results, and in combination with the changes of the drift characteristics, noise characteristics and jump characteristics, to obtain the stability analysis results.

[0031] The present invention provides a system for analyzing the stability of measurement results of measuring instruments, comprising a data optimization module, a feature extraction module, a three-state feature decomposition module, and a result analysis module;

[0032] The data optimization module is used to measure data stream alignment, obtain a measurement data stream sequence through measurement data stream alignment, and send the measurement data stream sequence to the feature extraction module;

[0033] The feature extraction module is used to construct the stability baseline feature of the measurement result. By constructing the stability baseline feature of the measurement result, the stability baseline feature vector of the measurement result is obtained, and the stability baseline feature vector of the measurement result is sent to the three-state feature decomposition module and the result analysis module.

[0034] The tri-state feature decomposition module is used for tri-state feature decomposition. Through tri-state feature decomposition, three types of stability sub-features are obtained, and the three types of stability sub-features are sent to the result analysis module.

[0035] The result analysis module is used for stability analysis of the measurement results of the measuring instrument. Through the stability analysis of the measurement results of the measuring instrument, the stability analysis results are obtained.

[0036] The beneficial effects achieved by the present invention using the above solution are as follows:

[0037] (1) In view of the technical problem that the existing methods for analyzing the stability of measurement results of metrology instruments do not effectively distinguish the timestamp errors, range switching, calibration status changes and operating condition differences during the measurement data acquisition process, resulting in the mixed analysis of measurement results under different conditions, which makes the stability analysis results susceptible to abnormal data interference and difficult to accurately reflect the true stable state, this solution creatively adopts a comprehensive measurement result stability analysis method that combines data alignment, baseline feature modeling and three-state feature decomposition. By performing unified time reference correction, sequence segmentation and state alignment processing on the measurement data at the front end, and introducing baseline reference and cause differentiation mechanism in the subsequent analysis, the solution achieves accurate characterization of the stable state of measurement results of metrology instruments under complex operating conditions.

[0038] (2) In the process of constructing the stability baseline features of existing measurement results, there is a problem that the baseline is directly constructed based on the full data or a fixed statistical window without distinguishing between stable operating intervals and unstable intervals. This makes the baseline features easily skewed by transient fluctuations, operating condition switching or abnormal data, thereby reducing the reliability of stability assessment. This solution creatively adopts a multi-dimensional statistical result stability baseline construction method based on stable operating interval constraints. By first identifying stable operating intervals that continuously meet the stability judgment conditions, repeatability features, trend features and structural features are extracted in the interval and uniformly normalized modeled, a stability baseline that can truly reflect the inherent statistical structure of the stable operating state of the measuring instrument is realized.

[0039] (3) In response to the technical problem that it is difficult to distinguish whether the stability change of the measurement result is caused by long-term systematic drift, random noise enhancement or sudden anomaly in the existing feature decoupling process, thus failing to provide a clear basis for instrument status assessment and maintenance decision-making, this solution creatively adopts a three-state feature decomposition method based on baseline constraints. Under the constraint of the stability baseline feature vector, the relative change of the measurement result is decomposed in a structured manner, and the stability change is clearly divided into three types of stability sub-features: drift feature, noise feature and jump feature, thus realizing a clear distinction between the causes of different stability changes. Attached Figure Description

[0040] Figure 1 A flowchart illustrating a method for analyzing the stability of measurement results from a measuring instrument, provided by this invention.

[0041] Figure 2 A schematic diagram of a stability analysis system for measurement results of a measuring instrument provided by the present invention;

[0042] Figure 3 A flowchart illustrating the construction of the stability baseline characteristics of the measurement results in step S2;

[0043] Figure 4 This is a flowchart illustrating the three-state feature decomposition process in step S3.

[0044] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. Detailed Implementation

[0045] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0046] In the description of this invention, it should be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0047] Example 1, see Figure 1 The present invention provides a method for stability analysis of measurement results of measuring instruments, the method comprising the following steps:

[0048] Step S1: Measure data stream alignment;

[0049] Step S2: Constructing baseline characteristics for the stability of measurement results;

[0050] Step S3: Three-state feature decomposition;

[0051] Step S4: Stability analysis of measurement results from measuring instruments.

[0052] By performing the above operations, this solution addresses the technical problem in existing methods for analyzing the stability of measurement results of metrological instruments. This problem stems from the failure to effectively distinguish between time stamp errors, range switching, calibration status changes, and differences in operating conditions during the data acquisition process. This results in the mixing and analysis of measurement results under different conditions, making the stability analysis results susceptible to interference from abnormal data and unable to accurately reflect the true stable state. This solution creatively adopts a comprehensive method for analyzing the stability of measurement results that combines data alignment, baseline feature modeling, and three-state feature decomposition. By performing unified time reference correction, sequence segmentation, and state alignment on the measurement data at the front end, and introducing baseline reference and cause differentiation mechanisms in subsequent analysis, it achieves an accurate characterization of the stable state of metrological instrument measurement results under complex operating conditions.

[0053] For example, in applications where electronic energy meters require periodic calibration and load condition switching, this solution can avoid misjudgments of stability caused by directly splicing data before and after calibration, making the analysis results more consistent with actual operating conditions.

[0054] Example 2, see Figure 1 and Figure 2This embodiment is based on the above embodiment. In step S1, the measurement data stream alignment is used to eliminate the sequence inconsistency caused by the measurement results of the measuring instrument under the acquisition time sequence, operating condition switching and operating status change. Specifically, based on the timestamp information corresponding to the measurement results of the measuring instrument, the instrument operating status identifier and the calibration status change information, the continuous measurement results are rearranged in time order. When the range switching, calibration status change or operating condition switching is detected, the measurement data stream is segmented and aligned to obtain the measurement data stream sequence.

[0055] Preferably, in step S1, the timestamp information corresponding to the measurement results of the measuring instrument is subjected to unified time reference correction processing. Specifically, the measurement time information from different acquisition modules or different recording units is mapped to a unified system time axis to eliminate the time sequence disorder caused by acquisition delay, difference in cache writing order or data retransmission, thereby ensuring the continuity and consistency of the measurement data stream in the time dimension.

[0056] In this embodiment, when performing unified time reference correction, the timestamp is checked for consistency based on the sampling number, data generation sequence identifier or internal clock identifier in the measurement result record. When the time sequence of adjacent measurement results is found to be inconsistent with the sampling number sequence, the corresponding timestamp is corrected first based on the sampling number or data generation sequence identifier to ensure the consistency of the measurement results on the unified time axis.

[0057] When mapping measurement time information to a unified system time axis, if the time interval between adjacent measurement results is detected to deviate from the preset sampling period range, the corresponding timestamp is corrected or interpolated to ensure that the corrected measurement results meet the requirements of continuous sampling order, thereby avoiding the impact of abnormal time intervals on the continuity determination of subsequent measurement data stream sequences.

[0058] The measuring instrument is specifically an electronic measuring device that can output measurement results in the form of a time series. During operation, it may switch measurement ranges, change operating conditions, or adjust calibration status.

[0059] In one embodiment, the metering instrument may be an electronic energy meter; in another embodiment, the metering instrument may be an electronic flow meter or pressure gauge.

[0060] More preferably, during the process of rearranging the time sequence of continuous measurement results, the sampling number, sampling sequence identifier or data generation sequence identifier in the measurement result record is used to perform sequence verification on measurement results with the same or similar timestamps. When an abnormal jump in timestamp or a sequence conflict is detected, the corresponding measurement results are adjusted according to the preset sequence correction rules to avoid affecting the accuracy of subsequent stability feature extraction due to time sorting errors.

[0061] The preset order correction rules in this embodiment include: when there are multiple measurement results with the same timestamp, they are reordered according to the size order of the sampling numbers; when the sampling number is missing, they are sorted according to the data generation sequence identifier or the record writing order; when there is a conflict between the timestamp order and the sampling number order, the sampling number order is used as the priority sorting basis.

[0062] More preferably, when a range switch, calibration status change, or operating condition switch is detected, the measurement results before and after the corresponding time are divided into different data sequence segments. The data sequence segments are processed independently of each other, thereby avoiding the mixing and analysis of measurement results under different range states, different calibration states, or different operating conditions, and ensuring the consistency of the state within each measurement data stream sequence.

[0063] When dividing the data sequence into segments, if a change is detected in the range indicator, operating status indicator, or calibration status indicator output by the measuring instrument, the corresponding time of change is used as the sequence segmentation boundary to divide the measurement data stream into segments.

[0064] When the range indicator or status indicator is not directly recorded, if a sudden change in the measurement result is detected that exceeds the preset range within an adjacent sampling period, and the change continues for more than the preset duration, the corresponding change time will be used as the sequence splitting boundary to avoid data from different operating states being assigned to the same measurement data stream sequence.

[0065] The sequence identification information includes at least one of sequence start time information, sequence end time information, and corresponding running status identification information, so that each measurement data stream sequence has a unique corresponding sequence identifier.

[0066] As a further optimization of this embodiment, after segmenting the measurement data stream sequence, a corresponding sequence identifier is generated for each measurement data stream sequence, and the sequence identifier is associated with and stored with the measurement data stream sequence. This allows for the differentiation of different measurement data stream sequences based on the sequence identifier during subsequent stability baseline feature construction and tri-state feature decomposition, thereby improving the traceability of the stability analysis process and the consistency of the results.

[0067] As a further optimization of this embodiment, after completing the measurement data stream alignment process, the integrity of the obtained measurement data stream sequence is checked. When measurement data is missing, duplicated, or abnormally interrupted, the corresponding measurement data stream sequence is marked or removed to avoid abnormal data streams interfering with the stability analysis of subsequent measurement results. After removal, linear interpolation or polynomial interpolation based on adjacent valid data can preferably be used for default value compensation to maintain the equidistant continuity of the time series.

[0068] Example 3, see Figure 1 , Figure 2 and Figure 3 This embodiment is based on the above embodiment. In step S2, the measurement result stability baseline feature construction is used to establish a stability reference benchmark that can characterize the inherent statistical structure of the measurement results of the measuring instrument under stable operating conditions. Specifically, a multidimensional statistical result stability baseline construction method based on stable operating interval constraints is adopted. In the measurement data stream sequence, a measurement interval that meets the stable operating conditions is selected, and multidimensional stability features are extracted from the measurement results in the measurement interval to obtain multidimensional stability features. The multidimensional stability features are then combined to construct the corresponding measurement result stability baseline feature vector.

[0069] The stability features include at least repeatability features for characterizing the repeatability of measurement results, trend features for characterizing the long-term trend of measurement results, and structural features for characterizing the distribution of abnormal structures.

[0070] The method for constructing a stability baseline for multidimensional statistical results based on stable operating interval constraints specifically includes the following steps:

[0071] Step S21: Determining the stable operating interval, which is used to determine the measurement interval that meets the stable operating conditions from the measurement data stream sequence. Specifically, the measurement data stream sequence is divided into intervals according to a preset time window, and the measurement intervals that continuously meet the stability judgment conditions are selected based on the statistical characteristic changes of the measurement results in adjacent intervals as stable operating intervals.

[0072] The statistical characteristic changes of measurement results within adjacent intervals are specifically defined as a stability judgment function, which characterizes the stability of measurement results within each preset time window, the degree of fluctuation stability, the degree of continuous change of the mean, and the degree of deviation of the distribution structure. The calculation formula is as follows:

[0073] ;

[0074] In the formula, This is the stability criterion value for the k-th time window, where k is the index of the time window. It is the fluctuation stability weighting coefficient. It is the standard deviation of the measurement results in the k-th window. It is the average value of the measurement results in the k-th window. It is a weighting coefficient that changes continuously with the mean. It is the average value of the measurement results in the (k-1)th window. Here, L is the distribution structure deviation weighting coefficient, L is the window length, and i is the measurement result index. This is the i-th measurement result. It is the median of the measurement results in the k-th window;

[0075] Among them, the weighting coefficients are preferably determined based on historical measurement data statistics or set based on the performance indicators of the measuring instruments;

[0076] By using a preset stability threshold T s When satisfied When the k-th window is determined to be a stable operating window, the stable operating windows that continuously meet the stability conditions are combined to form a stable operating interval.

[0077] Step S22: Repeatability feature extraction, used to characterize the repeatability of measurement results within the stable operating range. Specifically, within the stable operating range, a sliding window statistical analysis is performed on the measurement results to calculate the fluctuation amplitude characteristics of the measurement results within each window, and the fluctuation amplitude characteristics are summarized to obtain repeatability features used to characterize the repeatability of the measurement results.

[0078] The formula for calculating the repeatability feature is as follows:

[0079] ;

[0080] In the formula, R is the repeatability characteristic, and M is the number of windows within the stable operating range;

[0081] Step S23: Trend feature extraction, used to characterize the long-term trend of measurement results within the stable operating range, specifically involves performing trend fitting processing on the measurement result sequence within the stable operating range and extracting trend features that reflect the strength of the trend of measurement results changing over time.

[0082] The formula for calculating the trend feature is:

[0083] ;

[0084] In the formula, T represents the trend characteristic, and N represents the number of measurement results in the stable operating range;

[0085] More preferably, to eliminate the influence of the overall fluctuation level of the measurement results on the trend change, a trend stability ratio is constructed based on the overall fluctuation scale of the stable operating range, and the calculation formula is as follows:

[0086] ;

[0087] In the formula, It is the trend stability ratio. It is the overall standard deviation of the stable operating range;

[0088] Step S24: Structural feature extraction, used to characterize the abnormal distribution structure of measurement results within the stable operating range, specifically, to analyze the statistical distribution pattern of measurement results within the stable operating range and extract structural features that reflect the degree of skewness of the measurement result distribution, the degree of peak concentration, and the proportion of outliers;

[0089] Preferably, to achieve a quantitative characterization of the distribution structure characteristics of the measurement results, a characteristic quantity of the skewness of the measurement result distribution is constructed, and the calculation formula is as follows:

[0090] ;

[0091] In the formula, C is a characteristic quantity representing the degree of skewness in the distribution of the measurement results. It is the average value of the stable operating range. It is the standard deviation of the stable operating range;

[0092] The peak concentration feature is constructed, and the calculation formula is as follows:

[0093] ;

[0094] In the formula, P is a characteristic quantity of peak concentration;

[0095] More preferably, to characterize the distribution ratio of abnormal measurement results within the stable operating range, an abnormality proportion feature is constructed, and the calculation formula is as follows:

[0096] ;

[0097] In the formula, O is the anomaly proportion characteristic quantity, and N a This refers to the number of abnormal data.

[0098] Step S25: Construction of stability baseline feature vector, used to form a unified reference benchmark for subsequent stability analysis. Specifically, the repeatability feature, trend feature and structural feature are normalized, and the normalized multidimensional stability features are combined to construct the measurement result stability baseline feature vector.

[0099] ;

[0100] In the formula, F n It is a multidimensional stability feature obtained after normalization, where F is the feature combination composed of the repeatability feature, trend feature, and structural feature. min Fmax is the minimum value of the feature, and Fmax is the maximum value of the feature.

[0101] ;

[0102] In the formula, B is the baseline eigenvector of the measurement result stability, and R... nIt is a normalized repeatability feature. It is the normalized trend stable ratio. It is a characteristic quantity of the skewness of the distribution of normalized measurement results. It is a characteristic quantity of the normalized peak concentration. It is the normalized anomaly proportion characteristic.

[0103] By performing the above operations, this solution addresses the technical problem that existing measurement result stability baseline feature construction processes often involve directly constructing baselines based on full data or fixed statistical windows without distinguishing between stable and unstable operating intervals. This leads to baseline features being easily skewed by transient fluctuations, operating condition changes, or abnormal data, thereby reducing the reliability of stability assessment. The solution creatively adopts a multi-dimensional statistical result stability baseline construction method based on stable operating interval constraints. By first identifying stable operating intervals that continuously meet stability judgment conditions, repeatability features, trend features, and structural features are extracted from these intervals and uniformly normalized and modeled. This achieves a stability baseline that can truly reflect the inherent statistical structure of the stable operating state of the measuring instrument.

[0104] For example, during the long-term operation of an electronic flow meter, short-term flow pulsations are not included in the baseline modeling, thereby avoiding frequent changes in baseline characteristics due to instantaneous fluctuations and improving the reference consistency of stability analysis.

[0105] Example 4, see Figure 1 , Figure 2 and Figure 4 This embodiment is based on the above embodiment. In step S3, the three-state feature decomposition is used to structurally distinguish the different causes of the stability change of the measurement result. Specifically, a baseline-constrained three-state feature decomposition method is adopted. Based on the baseline feature vector of the measurement result stability, the measurement data stream sequence is subjected to feature decomposition processing. The change features of the measurement result relative to the baseline feature vector of the measurement result stability are divided into drift features used to characterize the long-term systematic change of the measurement result, noise features used to characterize the random fluctuation amplitude of the measurement result, and jump features used to characterize the sudden abnormal change in the measurement result. Three types of stability sub-features that are distinguishable to each other and reflect different causes of stability change are obtained.

[0106] The baseline-constrained tri-state eigenvalue decomposition method specifically includes the following steps:

[0107] Step S31: Baseline consistency processing, used to map the measurement result sequence to a comparable feature space relative to the measurement result stability baseline feature vector. Specifically, based on the baseline statistical features in the measurement result stability baseline feature vector, the measurement data stream sequence is normalized to obtain the measurement result feature sequence relative to the stability baseline.

[0108] The formula for calculating the characteristic sequence of the measurement results relative to the stability baseline is as follows:

[0109] ;

[0110] In the formula, It is the characteristic sequence of the measurement results, x t This is the measurement result at time t. It is the baseline center value derived from the baseline eigenvector of the measurement results stability, preferably the statistical center value of the measurement results in the stable operating interval. It is the baseline fluctuation scale;

[0111] The formula for calculating the baseline fluctuation scale is as follows:

[0112] ;

[0113] In the formula, It is the benchmark fluctuation scale within the stable operating range. It is the repeatability adjustment coefficient. It is the structural offset adjustment coefficient. It is the abnormal proportion adjustment coefficient;

[0114] Step S32: Drift feature extraction, used to characterize the long-term systematic change characteristics of the measurement results, specifically, performing smoothing trend estimation processing on the normalized measurement result feature sequence, extracting the trend component of the measurement results changing over time, constructing drift features to characterize the long-term systematic change of the measurement results based on the trend component, and removing the trend component from the measurement result feature sequence to obtain a detrended feature sequence;

[0115] The calculation formula for constructing the drift characteristics based on the trend components to characterize the long-term systematic changes in the measurement results is as follows:

[0116] ;

[0117] In the formula, This is the drift feature, and j is the index of the trend estimation window. It is the trend estimation window length, Z j It is the feature sequence of measurement results corresponding to window j. It is the mean of the feature sequences within the window;

[0118] The formula for calculating the detrended feature sequence is:

[0119] ;

[0120] In the formula, It is a detrended feature sequence;

[0121] Step S33: Jump feature extraction, used to characterize sudden abnormal change features in the measurement results. Specifically, in the detrended feature sequence, based on the baseline fluctuation scale corresponding to the baseline feature vector of the measurement result stability, threshold determination is performed on the adjacent changes in the measurement result feature sequence to identify the sudden change location, and jump features reflecting the frequency and amplitude of the sudden change are extracted. The sudden change component is removed from the detrended feature sequence to obtain the remaining measurement result feature sequence.

[0122] More preferably, adjacent changes are constructed, and the calculation formula is as follows:

[0123] ;

[0124] In the formula, These are adjacent changes, representing the changes in residuals at adjacent time points;

[0125] Based on the baseline fluctuation scale corresponding to the stability baseline feature vector of the measurement results, a jump judgment threshold is constructed, and the calculation formula is as follows:

[0126] ;

[0127] In the formula, It is the threshold for determining the transition. It is the jump threshold adjustment coefficient;

[0128] When satisfied When time t is determined to be a sudden change in position, a jump position indicator is generated, calculated using the following formula:

[0129] ;

[0130] In the formula, It is a jump position indicator;

[0131] Furthermore, jump features reflecting the frequency and amplitude of sudden changes are extracted, and jump amplitude and jump frequency features are constructed. The calculation formula is as follows:

[0132] ;

[0133] ;

[0134] In the formula, It is a characteristic quantity of the jump amplitude. It is a jump frequency characteristic quantity. It is a very small constant;

[0135] The transition characteristics include transition timing characteristics, transition amplitude characteristics, and transition frequency characteristics. The calculation formula for the transition timing characteristics is as follows:

[0136] ;

[0137] In the formula, J t It is a jump-time characteristic quantity;

[0138] The formula for calculating the feature sequence of the remaining measurement results is:

[0139] ;

[0140] In the formula, It is the characteristic sequence of the remaining measurement results;

[0141] Step S34: Noise feature extraction, used to characterize the random fluctuation amplitude of the measurement results. Specifically, after drift feature extraction and jump feature extraction, statistical analysis is performed on the fluctuation of the remaining measurement result feature sequence to extract noise features that reflect the random fluctuation amplitude of the measurement results.

[0142] The formula for calculating the noise characteristics is:

[0143] ;

[0144] In the formula, It is a noise characteristic;

[0145] Step S35: Construction of three-state stability sub-features, used to form a set of stability sub-features that are distinct from each other and reflect different causes of stability changes. Specifically, the drift feature, noise feature and jump feature are output as independent stability sub-features, which serve as the input basis for subsequent stability analysis of measurement results of metrology instruments.

[0146] The three-state stability sub-feature satisfies the baseline constraint nonlinear decomposition relationship, and its calculation formula is as follows:

[0147] ;

[0148] In the formula, It is a sequence of measurement results. It is the drift feature in the three-state stability sub-feature. It is a jump feature in the three-state stability sub-feature. It is a noise feature in the three-state stability sub-feature.

[0149] By performing the above operations, this solution addresses the technical problem that in existing feature decoupling and decomposition processes, it is difficult to distinguish whether changes in the stability of measurement results are caused by long-term systematic drift, random noise enhancement, or sudden anomalies, thus failing to provide a clear basis for instrument condition assessment and maintenance decisions. This solution creatively adopts a three-state feature decomposition method based on baseline constraints. Under the constraint of the stability baseline feature vector, the relative changes in measurement results are structurally decomposed, and the stability changes are clearly divided into three types of stability sub-features: drift features, noise features, and jump features, thus achieving a clear distinction between the causes of different stability changes.

[0150] For example, in scenarios where pressure gauge sensors gradually age, this solution can identify long-term systematic shift trends through drift characteristics, rather than misjudging them as random noise or sudden anomalies, thus providing a more targeted analytical basis for instrument calibration or replacement.

[0151] Example 5, see Figure 1 , Figure 2 , Figure 3 and Figure 4 This embodiment is based on the above embodiments. In a specific optional implementation, to further clarify the engineering application of the present invention in actual industrial scenarios, taking the operational stability analysis scenario of a 0.2S-level high-precision electronic energy meter (sampling frequency of 1Hz) as an example, the specific parameter setting strategies and examples of the weight coefficients and judgment thresholds involved in this method are as follows:

[0152] The parameters for determining the stable operating range are set as follows: when determining the stable operating range (step S21), the window length L of the preset time window is preferably set to 300 (i.e., corresponding to 5 minutes of sampling data) to balance statistical significance and real-time calculation.

[0153] In stability criterion function In this context, each weighting coefficient reflects the sensitivity to fluctuations in different dimensions, including:

[0154] Fluctuation stability weighting coefficient The preferred value is 0.5; since the high-frequency random fluctuation of the electricity meter under stable load is the main indicator characterizing the electrical characteristics of the equipment, it is given the maximum weight.

[0155] Weighting coefficient for continuous change of mean The value is preferably set to 0.3; this is used to capture the temperature drift effect caused by the slow drift of ambient temperature.

[0156] Distribution structure deviation from weight coefficient The value is preferably set to 0.2; this is used to limit interference from local non-Gaussian noise.

[0157] Under the above weight configuration, the stability determination threshold T s The preferred setting is 0.05; that is, when the condition is met... At that time, the physical corresponding comprehensive measurement error fluctuation of the electricity meter within the 5-minute window is limited to within 25% of the nominal accuracy, thus being strictly determined as a stable operating range;

[0158] The baseline variability scale adjustment parameter is set specifically during baseline consistency processing (step S31) to adjust the baseline variability scale. To adapt to individual aging differences among different instruments, the optimal values ​​for the reference fluctuation scale and various adjustment coefficients are as follows:

[0159] Benchmark fluctuation scale The standard deviation reference value is preferably set to the value recorded in the factory inspection report of this model of energy meter, which is 0.002 in this embodiment;

[0160] Repeatability adjustment coefficient The preferred setting is 0.15;

[0161] Structural offset adjustment coefficient The preferred setting is 0.20;

[0162] Abnormal proportion adjustment coefficient The preferred setting is 0.30;

[0163] Because an increase in abnormal data has a strong destructive effect on equipment stability, therefore The value is greater than and Under this configuration, when the device is in a sub-healthy state, the baseline fluctuation scale can be adaptively amplified up to 1.65 times the original value, effectively preventing false triggering of subsequent jump alarms.

[0164] The parameter settings for tri-state feature decomposition specifically involve setting isolation boundaries for drift features, jump features, and noise features when extracting tri-state stability sub-features, including...

[0165] Trend estimation window length The preferred setting is 3600 (corresponding to a 1-hour sliding window) to fully smooth short-term random fluctuations and accurately extract long-term systematic drift trends D. t ;

[0166] Jump threshold adjustment coefficient The preferred setting is 3.0; this setting is based on the Laida criterion in statistics, where adjacent changes... More than 3.0 seconds B In such cases, it can be determined with a very high degree of confidence that a sudden jump has occurred, such as a sudden change in load or a transient sensor failure.

[0167] Minimal constant The preferred setting is 1×10 -6 This is used to prevent the denominator from being zero in the formula for calculating the characteristic quantity of the jump amplitude.

[0168] Example 6, see Figure 1 , Figure 2 This embodiment is based on the above embodiment. In step S4, the stability analysis of the measurement results of the measuring instrument is used to comprehensively determine the stability state of the measurement results of the measuring instrument based on the stability baseline characteristics and the three-state characteristic decomposition results of the measurement results. Specifically, it analyzes the stability degree and stability change trend of the measurement results of the measuring instrument based on the degree of deviation of the measurement results from the stability baseline characteristic vector of the measurement results, and in combination with the changes of the drift characteristics, noise characteristics and jump characteristics, to obtain the stability analysis results.

[0169] Preferably, in step S4, the stability of the measurement result of the measuring instrument is initially evaluated based on the degree of deviation of the measurement result from the stability baseline feature vector of the measurement result. Specifically, the multidimensional stability feature corresponding to the current measurement result is compared with the stability baseline feature vector, the deviation of each dimension of stability feature is calculated, and the measurement result is judged to deviate from the stable operating state based on the deviation.

[0170] More preferably, during the evaluation of the stable state of the measurement results, the drift characteristics, noise characteristics, and jump characteristics obtained in step S3 are combined to comprehensively analyze the causes of the changes in the stability of the measurement results. Specifically, it is determined whether the drift characteristics have a continuous increasing trend, whether the noise characteristics exceed the preset fluctuation range, and whether the jump characteristics show abnormal frequency or amplitude, so as to distinguish whether the changes in the stability of the measurement results are caused by long-term systematic changes, enhanced random fluctuations, or sudden abnormal changes.

[0171] Based on the changes in the drift characteristics, noise characteristics, and jump characteristics, the stability of the measurement results of the measuring instrument is classified and determined. Specifically, when all stability characteristics of the measurement results are within the stability baseline range, the measurement results are determined to be in a stable state; when some stability characteristics deviate but do not reach the abnormal threshold, the measurement results are determined to be in a metastable state; when the stability characteristics deviate significantly from the stability baseline feature vector, the measurement results are determined to be in an unstable state.

[0172] As a further optimization of this embodiment, after the stable state determination is completed, the change trends of the drift characteristics, noise characteristics and jump characteristics in the time dimension are analyzed. By comparing the change direction and change magnitude of the stability characteristics at different times, it is determined whether the stability of the measurement result shows a trend of gradual deterioration, intermittent fluctuation or short-term abnormal recovery, thereby obtaining the stability change trend analysis result.

[0173] The stability state determination result is combined with the stability change trend analysis result to generate the stability analysis result of the measuring instrument measurement result. The stability analysis result is used to characterize the stability of the current measurement result and its change over time, and can be used as the basis for evaluating the operating status of the measuring instrument, judging the reliability of the measurement result, or making subsequent processing decisions.

[0174] Example 7, see Figure 1 and Figure 2 Based on the above embodiments, this embodiment provides a stability analysis system for measurement results of measuring instruments, including a data optimization module, a feature extraction module, a three-state feature decomposition module, and a result analysis module.

[0175] The data optimization module is used to measure data stream alignment, obtain a measurement data stream sequence through measurement data stream alignment, and send the measurement data stream sequence to the feature extraction module;

[0176] The feature extraction module is used to construct the stability baseline feature of the measurement result. By constructing the stability baseline feature of the measurement result, the stability baseline feature vector of the measurement result is obtained, and the stability baseline feature vector of the measurement result is sent to the three-state feature decomposition module and the result analysis module.

[0177] The tri-state feature decomposition module is used for tri-state feature decomposition. Through tri-state feature decomposition, three types of stability sub-features are obtained, and the three types of stability sub-features are sent to the result analysis module.

[0178] The result analysis module is used for stability analysis of the measurement results of the measuring instrument. Through the stability analysis of the measurement results of the measuring instrument, the stability analysis results are obtained.

[0179] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0180] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.

[0181] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A method for analyzing the stability of measurement results from a measuring instrument, characterized in that: The method includes the following steps: Step S1: Measurement data stream alignment. Based on the measurement results of the measuring instrument, the measurement data stream is segmented and aligned to obtain the measurement data stream sequence. Step S2: Construction of measurement result stability baseline features. A multidimensional statistical method for constructing a stability baseline based on stable operating interval constraints is adopted. In the measurement data stream sequence, measurement intervals that meet stable operating conditions are selected, and multidimensional stability features are extracted from the measurement results within these intervals. These multidimensional stability features are then combined to construct the corresponding measurement result stability baseline feature vector. Specifically, this includes the following steps: Step S21: Determining the stable operating interval, which is used to determine the measurement interval that meets the stable operating conditions from the measurement data stream sequence. Specifically, the measurement data stream sequence is divided into intervals according to a preset time window, and the measurement intervals that continuously meet the stability judgment conditions are selected based on the statistical characteristic changes of the measurement results in adjacent intervals as stable operating intervals. The statistical characteristic changes of measurement results within adjacent intervals are specifically defined as a stability judgment function, which characterizes the stability of measurement results within each preset time window, the degree of fluctuation stability, the degree of continuous change of the mean, and the degree of deviation of the distribution structure. The calculation formula is as follows: ； In the formula, This is the stability criterion value for the k-th time window, where k is the index of the time window. It is the fluctuation stability weighting coefficient. It is the standard deviation of the measurement results in the k-th window. It is the average value of the measurement results in the k-th window. It is a weighting coefficient that changes continuously with the mean. It is the average value of the measurement results in the (k-1)th window. Here, L is the distribution structure deviation weighting coefficient, L is the window length, and i is the measurement result index. This is the i-th measurement result. It is the median of the measurement results in the k-th window; among which, each weight coefficient is determined based on the statistical analysis of historical measurement data or set based on the performance indicators of the measuring instrument; By using a preset stability threshold T s When satisfied When the k-th window is determined to be a stable operating window, stable operating windows that continuously meet the stability conditions are combined to form a stable operating interval. Step S22: Repeatability feature extraction, used to characterize the repeatability of measurement results within the stable operating range. Specifically, within the stable operating range, a sliding window statistical analysis is performed on the measurement results to calculate the fluctuation amplitude characteristics of the measurement results within each window, and the fluctuation amplitude characteristics are summarized to obtain repeatability features used to characterize the repeatability of the measurement results. The formula for calculating the repeatability feature is as follows: ； In the formula, R is the repeatability characteristic, and M is the number of windows within the stable operating range; Step S23: Trend feature extraction, used to characterize the long-term trend of measurement results within the stable operating range, specifically, performing trend fitting processing on the measurement result sequence within the stable operating range, and extracting trend features that reflect the strength of the trend of measurement results changing over time, i.e., the trend stability ratio. Step S24: Structural feature extraction, used to characterize the abnormal distribution structure of measurement results within the stable operating range. Specifically, it involves analyzing the statistical distribution pattern of measurement results within the stable operating range and extracting structural features that reflect the degree of skewness in the distribution of measurement results, the degree of peak concentration, and the proportion of outliers. The structural features include the characteristic quantity of the degree of skewness in the distribution of measurement results, the characteristic quantity of peak concentration, and the characteristic quantity of outliers. Step S25: Construction of stability baseline feature vector, used to form a unified reference benchmark for subsequent stability analysis. Specifically, the repeatability feature, trend feature and structural feature are normalized, and the normalized multidimensional stability features are combined to construct the measurement result stability baseline feature vector. The baseline feature vector for measurement result stability includes baseline statistical features such as repeatability characteristics, trend stability ratio, measurement result distribution skewness characteristics, peak concentration characteristics, and anomaly proportion characteristics. Step S3: Tri-state feature decomposition. Using a baseline-constrained tri-state feature decomposition method, based on the stability baseline feature vector of the measurement results, the measurement data stream sequence is subjected to feature decomposition processing. The change features of the measurement results relative to the stability baseline feature vector of the measurement results are divided into drift features for characterizing long-term systematic changes in the measurement results, noise features for characterizing the random fluctuation amplitude of the measurement results, and jump features for characterizing sudden abnormal changes in the measurement results. Three types of stability sub-features are obtained that are distinguishable from each other and reflect different causes of stability changes. Step S4: Stability analysis of measurement results of measuring instruments. Based on the degree of deviation of the measurement results from the stability baseline feature vector of the measurement results, and in combination with the changes in the drift feature, noise feature and jump feature, the stability degree and stability change trend of the measurement results of the measuring instruments are analyzed to obtain the stability analysis results.

2. The method for stability analysis of measurement results of a measuring instrument according to claim 1, characterized in that: In step S1, the measuring instrument specifically refers to an electronic measuring device that can output measurement results in the form of a time series, and has functions such as measurement range switching, operating condition changes, or calibration status adjustment during operation.

3. The method for stability analysis of measurement results of a measuring instrument according to claim 2, characterized in that: In step S3, the baseline-constrained tri-state feature decomposition method specifically includes the following steps: Step S31: Baseline consistency processing; Step S32: Drift feature extraction; Step S33: Jump feature extraction; Step S34: Noise feature extraction; Step S35: Construction of tri-state stability sub-features; In step S31, the baseline consistency processing is used to map the measurement result sequence to a comparable feature space relative to the measurement result stability baseline feature vector. Specifically, based on the baseline statistical features in the measurement result stability baseline feature vector, the measurement data stream sequence is normalized to obtain the measurement result feature sequence relative to the stability baseline.

4. The method for stability analysis of measurement results of a measuring instrument according to claim 3, characterized in that: In step S32, the drift feature extraction is used to characterize the long-term systematic change characteristics of the measurement results. Specifically, it involves performing smoothing trend estimation processing on the normalized measurement result feature sequence, extracting the trend component of the measurement results changing over time, constructing drift features to characterize the long-term systematic change of the measurement results based on the trend component, and removing the trend component from the measurement result feature sequence to obtain a detrended feature sequence. In step S33, the jump feature extraction is used to characterize the sudden abnormal change features in the measurement results. Specifically, in the detrended feature sequence, based on the baseline fluctuation scale corresponding to the baseline feature vector of the measurement result stability, the adjacent change amounts of the measurement result feature sequence are thresholded to identify the sudden change location, and jump features reflecting the frequency and amplitude of the sudden change are extracted. The sudden change component is removed from the detrended feature sequence to obtain the remaining measurement result feature sequence.

5. The method for stability analysis of measurement results of a measuring instrument according to claim 4, characterized in that: In step S34, the noise feature extraction is used to characterize the random fluctuation amplitude of the measurement results. Specifically, after the drift feature extraction and jump feature extraction, the fluctuation of the remaining measurement result feature sequence is statistically analyzed to extract noise features that reflect the random fluctuation amplitude of the measurement results. In step S35, the construction of the three-state stability sub-features is used to form a set of stability sub-features that are distinct from each other and reflect different causes of stability changes. Specifically, the drift feature, noise feature and jump feature are output as independent stability sub-features as input for subsequent stability analysis of measurement results of metrology instruments.

6. A system for analyzing the stability of measurement results of a measuring instrument, used to implement the method for analyzing the stability of measurement results of a measuring instrument as described in any one of claims 1-5, characterized in that: It includes a data optimization module, a feature extraction module, a three-state feature decomposition module, and a result analysis module.

7. The system for analyzing the stability of measurement results of a measuring instrument according to claim 6, characterized in that: The data optimization module is used to measure data stream alignment, obtain a measurement data stream sequence through measurement data stream alignment, and send the measurement data stream sequence to the feature extraction module; The feature extraction module is used to construct the stability baseline feature of the measurement result. By constructing the stability baseline feature of the measurement result, the stability baseline feature vector of the measurement result is obtained, and the stability baseline feature vector of the measurement result is sent to the three-state feature decomposition module and the result analysis module. The tri-state feature decomposition module is used for tri-state feature decomposition. Through tri-state feature decomposition, three types of stability sub-features are obtained, and the three types of stability sub-features are sent to the result analysis module. The result analysis module is used for stability analysis of the measurement results of the measuring instrument. Through the stability analysis of the measurement results of the measuring instrument, the stability analysis results are obtained.

Images