Water conservancy project quality detection method and system based on data analysis

Abstract

This invention discloses a method and system for quality inspection of water conservancy projects based on data analysis, belonging to the field of data analysis technology. The method includes the following steps: generating a parameter evolution trajectory array at the current moment; confirming whether dependency drift has occurred; generating adjustment coefficients and updating the next checkpoint trigger time; performing uniform numerical range processing on the components of the difference array; simultaneously correcting the feature weight component set based on the overall magnitude of the difference array; and after verifying the adjusted association rule set and the corrected feature weight component set through tolerance range checks, writing both updated sets into a long-term data storage record. This invention solves the problem that the fixed checkpoint mechanism in the prior art is unable to capture dependency drift caused by the long-term nonlinear evolution of engineering parameters by dynamically adjusting the checkpoint interval through calculation of parameter evolution trajectory similarity.

Description

Technical Field

[0001] This invention relates to the field of data analysis technology, and in particular to a method and system for quality inspection of water conservancy projects based on data analysis. Background Technology

[0002] As large dams, reservoirs, and sluice gates enter their long-term service phase, the need for continuous monitoring and risk assessment of their quality status is becoming increasingly prominent. In actual engineering operation and maintenance, the correlation between key parameters such as seepage pressure, deformation, and stress exhibits significant nonlinear evolution characteristics with changes in service time, geological conditions, and environmental loads. Therefore, long-term incremental analysis of monitoring data is necessary to support quality management and safety early warning.

[0003] Existing data-driven water conservancy project quality inspection systems utilize fixed checkpoint mechanisms to schedule incremental data and update intermediate results, enabling correlation mining and risk assessment of engineering parameters. While these systems can handle a certain scale of time-series data in practical applications, in multi-stage quality inspection scenarios throughout the entire lifecycle of water conservancy projects, the fixed checkpoint mechanism struggles to effectively capture long-term dependency drift due to the non-linear evolution of engineering parameter correlations over service life. This leads to delayed intermediate result updates, error accumulation, and a gradual decrease in the matching degree between early warning outputs and on-site inspection feedback, making it difficult to form a stable iterative optimization loop.

[0004] In the existing technology, there is a lack of dedicated scheduling methods for the long-term time-varying characteristics of water conservancy. The above problems are particularly prominent in the application of quality inspection of complex projects that span multiple years and multiple working conditions, which restricts the in-depth implementation of big data analysis technology in water conservancy project quality management. Summary of the Invention

[0005] This application provides a data analysis-based method and system for quality inspection of water conservancy projects. It solves the problem that the fixed checkpoint mechanism in the prior art is unable to capture the dependency drift caused by the long-term nonlinear evolution of engineering parameters, resulting in delayed updates of intermediate results, error accumulation, and a continuous decline in the matching degree between early warning output and on-site inspection feedback. It realizes the technical effect of dynamically adjusting the checkpoint interval according to the parameter evolution trajectory and performing drift compensation, thereby improving the timeliness and accuracy of intermediate result updates and maintaining a high degree of matching between early warning output and actual engineering status.

[0006] This application provides a data analysis-based method for quality inspection of water conservancy projects, including: extracting the temporal features of the field engineering parameter observation sequence and splicing them together to generate a parameter evolution trajectory array at the current moment;

[0007] Calculate the trajectory similarity value between the current parameter evolution trajectory array and the historical parameter evolution trajectory array, and confirm whether dependency drift has occurred by comparing it with a preset interval;

[0008] An adjustment coefficient is generated by using the ratio of the trajectory similarity value to the historical average similarity value, thereby correcting the default checkpoint interval and updating the next checkpoint trigger time.

[0009] Obtain the difference array between the parameter evolution trajectory array at the current moment and the historical parameter evolution trajectory array, and perform numerical range unification processing on the components of the difference array;

[0010] The support and confidence metrics in the current association rule set are scaled and adjusted according to the processed difference components, while the feature weight component set is corrected according to the overall magnitude of the difference array.

[0011] After the adjusted set of association rules and the corrected set of feature weight components are verified through tolerance range, the updated versions are written into the long-term data storage record.

[0012] Furthermore, the steps for generating the parameter evolution trajectory array at the current moment include:

[0013] Collect seepage pressure, deformation, and stress data at the site of the water conservancy project, and perform segmentation within the sliding window;

[0014] Calculate the statistical feature values ​​of each sequence within the window, concatenate multiple feature values ​​in a fixed order to form an initial combination in the form of a fixed-dimensional numerical array, and store it in a memory buffer;

[0015] Each time new data arrives, an append update operation is performed to add the time series feature value corresponding to the new data to the end of the array, while keeping the total number of dimensions unchanged. The combined data is then used as the parameter evolution trajectory array for the current moment.

[0016] Furthermore, the step of confirming whether dependency drift has occurred by comparing with a preset interval includes:

[0017] Before each batch processing task starts, read the parameter evolution trajectory array at the current moment;

[0018] Simultaneously, retrieve the historical parameter evolution trajectory array saved at the previous checkpoint from the checkpoint storage location, calculate the spatial orientation consistency value between the current array and the historical array, and use it as the trajectory similarity value;

[0019] After the calculation is completed, the trajectory similarity value is compared with the pre-set similarity threshold range for judgment.

[0020] When the trajectory similarity value is lower than the lower limit of the threshold interval, it is determined that the current parameter evolution relationship has undergone dependency drift.

[0021] When the value is within the threshold range, it is determined that no dependency drift has occurred.

[0022] Furthermore, the steps to modify the default checkpoint interval and update the next checkpoint trigger time include:

[0023] Retrieve the average similarity value obtained from multiple historical calculations;

[0024] The adjustment coefficient is obtained by calculating the ratio between the current trajectory similarity value and the average similarity value.

[0025] The adjustment coefficient value is multiplied by the default checkpoint interval duration to obtain the new checkpoint interval duration.

[0026] Update the next checkpoint trigger time according to the new checkpoint interval;

[0027] After all current processing tasks are completed, the updated parameter evolution trajectory array is saved to the checkpoint storage location as a historical reference array for the next cycle.

[0028] Furthermore, the steps for performing numerical range unification processing on the components of the difference array include:

[0029] Extract the difference array between the current latest array and the historical array saved at the most recent checkpoint, as the dependency drift amount;

[0030] The difference array is input into the drift compensation process, and the drift compensation operator performs a uniform processing of the numerical range of each dimension component of the difference array to find the highest and lowest points of each independent component.

[0031] Based on the range determined by the highest and lowest points, the numerical range of each component is unified to the standard interval of 0-1, eliminating the numerical differences caused by different parameter measurement scales, and the processed components are determined as standardized difference components.

[0032] Further steps to adjust the support and confidence metrics in the current set of association rules include:

[0033] Read the configuration weight ratios of each dimension obtained through historical data statistical analysis during the initial configuration phase;

[0034] Iterate through the set of currently active association rules, obtain the support and confidence metrics of each rule, and scale and adjust the support and confidence metrics of each rule in the current set of association rules according to the preset weight ratio of each dimension.

[0035] During the adjustment process, the direction and magnitude of the indicator adjustment are kept in a positive relationship with the standardized difference components after processing the corresponding dimension.

[0036] Furthermore, the steps for adjusting the feature weight component set based on the overall magnitude of the difference array include:

[0037] Retrieve the feature weight vector currently used by the system, locate each corresponding component in the feature weight vector, and establish a positive proportional relationship between the correction magnitude and the overall modulus of the difference array.

[0038] Perform a weighted correction operation on each corresponding component in the feature weight vector, so that the correction magnitude is proportional to the overall magnitude of the difference array;

[0039] Immediately after the correction is complete, a numerical overwrite is performed in the computational memory to update the feature weight vector.

[0040] Furthermore, the steps to write the updated data to the long-term data storage record include:

[0041] Retrieve the preset tolerance range value;

[0042] The effectiveness of the correction operation is confirmed by comparing the change in the total number of rule items in the rule set before and after the adjustment, as well as the change in the total weight of the feature weight vector.

[0043] A write command is only allowed after both changed values ​​have passed the consistency check.

[0044] The updated set of association rules and the set of feature weight components are written to persistent storage medium for long-term data preservation. The complete sequence of operations described above is repeated in each subsequent incremental update cycle until an external process termination signal is received.

[0045] This application provides a water conservancy project quality inspection system based on data analysis, which is used to implement a water conservancy project quality inspection method based on data analysis, including: a sequence extraction module, a trajectory similarity calculation module, an adjustment coefficient generation module, a difference array acquisition module, an association rule set adjustment module, and a data writing module;

[0046] The sequence extraction module is used to extract the temporal features of the field engineering parameter observation sequence and splice them to generate the parameter evolution trajectory array at the current moment.

[0047] The trajectory similarity calculation module is used to calculate the trajectory similarity value between the parameter evolution trajectory array at the current moment and the historical parameter evolution trajectory array, and to confirm whether dependency drift has occurred by comparing it with a preset interval.

[0048] The adjustment coefficient generation module is used to generate an adjustment coefficient by using the ratio of the trajectory similarity value to the historical average similarity value, thereby correcting the default checkpoint interval duration and updating the next checkpoint trigger time.

[0049] The difference array acquisition module is used to acquire the difference array between the parameter evolution trajectory array at the current moment and the historical parameter evolution trajectory array, and to perform numerical range unification processing on the components of the difference array;

[0050] The association rule set adjustment module is used to scale and adjust the support and confidence metrics in the current association rule set according to the processed difference components, and at the same time, adjust the feature weight component set according to the overall magnitude of the difference array.

[0051] The data writing module is used to write the updated set of association rules and the corrected set of feature weight components into a long-term data storage record after passing tolerance range verification.

[0052] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0053] Considering that the parameter correlation of water conservancy projects will change nonlinearly over time during long-term service, the processing flow is scheduled by dynamically adjusting the checkpoint interval based on trajectory similarity. This generates intermediate results that are consistent with the current state of the project, improves the timeliness of updates, and helps maintain the consistency between the analysis results and the actual evolution of the project.

[0054] Furthermore, in the intermediate result generation stage, the difference between the current trajectory and the historical trajectory is extracted and processed within a unified range to compensate and adjust the association rules and feature weights. This timely corrects the deviation caused by long-term evolution, reduces the lag and cumulative error of parameter updates, and thus improves the matching degree between the early warning output and the actual inspection feedback.

[0055] Furthermore, throughout the entire incremental update cycle, the effectiveness of the adjustment is confirmed through consistency verification and the complete operation sequence is repeated to ensure that the iteration process can be stably closed, avoiding the problem of continuous decline in matching degree, effectively forming an optimizable feedback loop, and improving the applicability and reliability of the overall detection method in multi-stage quality management. Attached Figure Description

[0056] Figure 1 A flowchart of a data analysis-based water conservancy engineering quality inspection method provided in this application embodiment;

[0057] Figure 2 This is a schematic diagram of the structure of a water conservancy engineering quality inspection system based on data analysis, provided in an embodiment of this application. Detailed Implementation

[0058] This application provides a data analysis-based method and system for quality inspection of water conservancy projects. It solves the problems in the prior art where the fixed checkpoint mechanism is difficult to capture the dependency drift caused by the long-term nonlinear evolution of engineering parameters, resulting in delayed updates of intermediate results, error accumulation, and a continuous decline in the matching degree between early warning output and on-site inspection feedback. By calculating the similarity of parameter evolution trajectory, dynamically adjusting the checkpoint interval, extracting the trajectory difference array, and adjusting the association rules and feature weights through drift compensation, the technical effect of timely and accurate updates of intermediate results and a high degree of matching between early warning output and actual engineering status is achieved.

[0059] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0060] like Figure 1 The diagram shown is a flowchart of a water conservancy engineering quality inspection method based on data analysis provided in this application embodiment. The method is applied to a water conservancy engineering quality inspection system based on data analysis. The method includes the following steps: extracting the temporal features of the field engineering parameter observation value sequence and splicing them to generate a parameter evolution trajectory array at the current moment;

[0061] Calculate the trajectory similarity value between the current parameter evolution trajectory array and the historical parameter evolution trajectory array, and confirm whether dependency drift has occurred by comparing it with a preset interval;

[0062] An adjustment coefficient is generated by using the ratio of the trajectory similarity value to the historical average similarity value, thereby correcting the default checkpoint interval and updating the next checkpoint trigger time.

[0063] Obtain the difference array between the parameter evolution trajectory array at the current moment and the historical parameter evolution trajectory array, and perform numerical range unification processing on the components of the difference array;

[0064] The support and confidence metrics in the current association rule set are scaled and adjusted according to the processed difference components, while the feature weight component set is corrected according to the overall magnitude of the difference array.

[0065] After the adjusted set of association rules and the corrected set of feature weight components are verified through tolerance range, the updated versions are written into the long-term data storage record.

[0066] Furthermore, the steps for generating the parameter evolution trajectory array at the current moment include:

[0067] Collect seepage pressure, deformation, and stress data at the site of the water conservancy project, and perform segmentation within the sliding window;

[0068] Calculate the statistical feature values ​​of each sequence within the window, concatenate multiple feature values ​​in a fixed order to form an initial combination in the form of a fixed-dimensional numerical array, and store it in a memory buffer;

[0069] Each time new data arrives, an append update operation is performed to add the time series feature value corresponding to the new data to the end of the array, while keeping the total number of dimensions unchanged. The combined data is then used as the parameter evolution trajectory array for the current moment.

[0070] In this embodiment, at the water conservancy project site, sensors deployed on the dam body and surrounding area are used to collect real-time data on seepage pressure, deformation, and stress.

[0071] Data Acquisition and Window Segmentation: A sliding window is created in memory, with a fixed time span of 24 hours and a sliding step of 1 hour. The acquired raw time-series data is sequentially entered into this sliding window for segmentation processing.

[0072] Feature extraction and concatenation logic: For each type of sensor data sequence within the sliding window, its basic statistical feature values ​​are calculated, specifically including: mean, variance, and peak value. Subsequently, following a fixed order of "mean pressure - variance pressure - peak pressure - mean deformation - ...", the above feature values ​​are concatenated into a fixed-dimensional one-dimensional numerical array, which is stored in the memory buffer as the initial combination.

[0073] Append-to-update mechanism: When new incremental data arrives, the entire array is not reconstructed. Instead, the feature values ​​of the new data within a very short time window (e.g., 5 minutes) are extracted and appended to the end of the array. At the same time, the total number of dimensions of the array is maintained by eliminating the data node with the oldest timestamp. The combined data at this point is then identified as the "parameter evolution trajectory array at the current moment".

[0074] Furthermore, the step of confirming whether dependency drift has occurred by comparing with a preset interval includes:

[0075] Before each batch processing task starts, read the parameter evolution trajectory array at the current moment;

[0076] Simultaneously, retrieve the historical parameter evolution trajectory array saved at the previous checkpoint from the checkpoint storage location, calculate the spatial orientation consistency value between the current array and the historical array, and use it as the trajectory similarity value;

[0077] After the calculation is completed, the trajectory similarity value is compared with the pre-set similarity threshold range for judgment.

[0078] When the trajectory similarity value is lower than the lower limit of the threshold interval, it is determined that the current parameter evolution relationship has undergone dependency drift.

[0079] When the value is within the threshold range, it is determined that no dependency drift has occurred.

[0080] In this embodiment, a trajectory similarity evaluation mechanism is introduced to identify whether the state of engineering facilities has undergone abrupt changes or gradual degradation. The steps include:

[0081] Data retrieval: When the batch processing task starts, the parameter evolution trajectory array at the current moment in memory is recorded as... Simultaneously, it retrieves the historical parameter evolution trajectory array from the checkpoint storage table of the persistent database (such as MySQL or PostgreSQL) at the last checkpoint time, denoted as... .

[0082] Spatial orientation consistency calculation: The cosine similarity algorithm is used to calculate the trajectory similarity between the two. The calculation formula is as follows:

[0083] ;

[0084] in, This represents the total number of dimensions in the trajectory array; and These represent the current array and the historical array at the [number]th position, respectively. The formula calculates the cosine of the spatial angle, filtering out the interference of absolute numerical magnitudes and reflecting the consistency of the angle in the evolutionary trend.

[0085] Threshold comparison: Set the preset similarity threshold range to 0.85 to 1.00. When the calculated similarity... When the value is less than 0.85 (i.e., below the lower limit), it is determined that the parameter evolution relationship of the water conservancy facility has undergone dependency drift; when When the value is between 0.85 and 1.00, it is determined that no significant drift has occurred and the facility is in a stable condition.

[0086] Furthermore, the steps to modify the default checkpoint interval and update the next checkpoint trigger time include:

[0087] Retrieve the average similarity value obtained from multiple historical calculations;

[0088] The adjustment coefficient is obtained by calculating the ratio between the current trajectory similarity value and the average similarity value.

[0089] The adjustment coefficient value is multiplied by the default checkpoint interval duration to obtain the new checkpoint interval duration.

[0090] Update the next checkpoint trigger time according to the new checkpoint interval;

[0091] After all current processing tasks are completed, the updated parameter evolution trajectory array is saved to the checkpoint storage location as a historical reference array for the next cycle.

[0092] In this embodiment, after confirming the drift state, the saving frequency of the next data snapshot (checkpoint) is adaptively adjusted according to the severity of the evolution to save storage overhead and capture key anomalies. Specific steps include:

[0093] Historical average calculation: Retrieve the trajectory similarity values ​​from the five most recent calculations from the database, and calculate their arithmetic mean, which is recorded as the historical average similarity value. .

[0094] Adjusting coefficient generation and duration update: Calculating the current similarity Compared with historical average The ratio of the two values ​​is used to generate the adjustment coefficient. (Right now The default checkpoint interval is [specified value]. (For example, set to 120 minutes). Calculate the new checkpoint interval duration. :

[0095] ;

[0096] If the current trajectory similarity is lower than the historical average level (indicating rapid facility evolution and potential hidden dangers). Setting the value to less than 1 shortens the interval between the next checkpoint triggers, enabling high-frequency encrypted monitoring; conversely, setting the value to more than 1 lengthens the interval. Update the timer; after the current task completes, Write it to the database as a historical reference for the next cycle.

[0097] Furthermore, the steps for performing numerical range unification processing on the components of the difference array include:

[0098] Extract the difference array between the current latest array and the historical array saved at the most recent checkpoint, as the dependency drift amount;

[0099] The difference array is input into the drift compensation process, and the drift compensation operator performs a uniform processing of the numerical range of each dimension component of the difference array to find the highest and lowest points of each independent component.

[0100] Based on the range determined by the highest and lowest points, the numerical range of each component is unified to the standard interval of 0-1, eliminating numerical differences caused by different parameter measurement scales. The processed components are then identified as standardized difference components.

[0101] In this embodiment, difference analysis and standardization are required to quantify the specific drift amplitude.

[0102] Difference array extraction: Extract the current array With historical array Subtracting each item one by one yields the difference array, where the original difference components of each dimension are denoted as... .

[0103] Drift Compensation and Standardization: Given the significant difference between the measurement scales of osmotic pressure (typically kPa) and deformation (typically millimeters), directly using the absolute difference would cause subsequent weight updates to fail. A drift compensation operator performs a uniformization of the numerical range (i.e., range normalization). The largest component in the difference array is then identified. and minimum component .

[0104] Standardization formula: Standardized difference components of each dimension The calculation is as follows:

[0105] ;

[0106] After processing, the drift amplitude of all indicator dimensions was strictly unified to the standard range of 0 to 1, completely eliminating the barrier between dimensions and measurement scales.

[0107] Further steps to adjust the support and confidence metrics in the current set of association rules include:

[0108] Read the configuration weight ratios of each dimension obtained through historical data statistical analysis during the initial configuration phase;

[0109] Iterate through the set of currently active association rules, obtain the support and confidence metrics of each rule, and scale and adjust the support and confidence metrics of each rule in the current set of association rules according to the preset weight ratio of each dimension.

[0110] During the adjustment process, the direction and magnitude of the indicator adjustment are kept in a positive relationship with the standardized difference component after the corresponding dimension processing, so that the changes in the rule indicators can reflect the nonlinear evolution state of the long-term accumulation of field parameters in real time.

[0111] Furthermore, the steps for adjusting the feature weight component set based on the overall magnitude of the difference array include:

[0112] Retrieve the feature weight vector currently used by the system, locate each corresponding component in the feature weight vector, and establish a positive proportional relationship between the correction magnitude and the overall modulus of the difference array.

[0113] Perform a weighted correction operation on each corresponding component in the feature weight vector, so that the correction magnitude is proportional to the overall magnitude of the difference array;

[0114] Immediately after the correction is completed, a numerical overwrite is performed in the computational memory to update the feature weight vector, ensuring that the weight allocation can track the structural evolution of the facility in real time.

[0115] In this embodiment, an internal set of association rules (indicating concurrency relationships between parameters) and a set of feature weight components (indicating the evaluation importance of each parameter) are maintained for quality assessment. Specific steps include:

[0116] Association rule metric scaling: Reads the dimension weight ratios set during the initial configuration phase. Iterate through the set of currently active association rules, extracting the support and confidence metrics for each rule. Then, apply positive scaling to these metrics using the corresponding standardized difference components.

[0117] ;

[0118] The above adjustments allow for a moderate amplification of the rule confidence and support of feature dimensions with more pronounced drift in the evaluation model, thus reflecting the long-term cumulative nonlinear evolution of field parameters in real time.

[0119] Overall Modulus Calculation and Weight Correction: For the correction of the feature weight vector, the overall magnitude of the difference array is used. The square root of the sum of squares of all elements within the difference array is calculated as the overall modulus. :

[0120] ;

[0121] Weighted correction execution: Retrieve the corresponding component from the current feature weight vector. Based on the overall module length Perform proportional weighted adjustment:

[0122] ;

[0123] in This is the smoothing coefficient (usually set to 0.05). After calculation, a numerical overwrite operation is immediately performed in memory to update the feature weight vector.

[0124] Furthermore, the steps to write the updated data to the long-term data storage record include:

[0125] Retrieve the preset tolerance range value;

[0126] The effectiveness of the correction operation is confirmed by comparing the change in the total number of rule items in the rule set before and after the adjustment, as well as the change in the total weight of the feature weight vector.

[0127] A write command is only allowed after both changed values ​​have passed the consistency check.

[0128] The updated set of association rules and the set of feature weight components are written to persistent storage medium for long-term data preservation. The complete sequence of operations described above is repeated in each subsequent incremental update cycle until an external process termination signal is received.

[0129] In this embodiment, to prevent significant model fluctuations caused by transient sensor noise, a strict consistency verification mechanism is added before data is written to disk. The steps are as follows:

[0130] Calculation of validation metrics: 1. Initial number of original association rule items before statistical adjustment Compared with the adjusted final quantity The absolute value of the difference between the two is used to obtain the total change in the number of rule items. .

[0131] 2. Calculate the sum of the feature weight vectors before and after the weighted adjustment, take the absolute value of the difference, and obtain the change in the total weight. .

[0132] Tolerance comparison and disk execution: Retrieve preset tolerance range values ​​(e.g., rule quantity change tolerance set to 20, weight sum change tolerance set to 10% of the original sum). Only when... and The verification program only determines that the drift update is valid and free from malicious noise interference when all values ​​are strictly within their respective tolerance ranges.

[0133] Long-term data storage record: After the verification is passed, the main control program issues a write command to serialize the updated set of association rules and the set of feature weight components and then write them to the persistent storage medium (disk database).

[0134] like Figure 2 The diagram shown is a structural schematic of a water conservancy engineering quality inspection system based on data analysis provided in this application embodiment. The water conservancy engineering quality inspection system based on data analysis provided in this application embodiment includes: a sequence extraction module, a trajectory similarity calculation module, an adjustment coefficient generation module, a difference array acquisition module, an association rule set adjustment module, and a data writing module.

[0135] The sequence extraction module is used to extract the temporal features of the field engineering parameter observation sequence and splice them to generate the parameter evolution trajectory array at the current moment.

[0136] The trajectory similarity calculation module is used to calculate the trajectory similarity value between the parameter evolution trajectory array at the current moment and the historical parameter evolution trajectory array, and to confirm whether dependency drift has occurred by comparing it with a preset interval.

[0137] The adjustment coefficient generation module is used to generate an adjustment coefficient by using the ratio of the trajectory similarity value to the historical average similarity value, thereby correcting the default checkpoint interval duration and updating the next checkpoint trigger time.

[0138] The difference array acquisition module is used to acquire the difference array between the parameter evolution trajectory array at the current moment and the historical parameter evolution trajectory array, and to perform numerical range unification processing on the components of the difference array;

[0139] The association rule set adjustment module is used to scale and adjust the support and confidence metrics in the current association rule set according to the processed difference components, and at the same time, adjust the feature weight component set according to the overall magnitude of the difference array.

[0140] The data writing module is used to write the updated set of association rules and the corrected set of feature weight components into a long-term data storage record after passing tolerance range verification.

[0141] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0142] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0143] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0144] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0145] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0146] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A data analysis-based method for quality inspection of water conservancy projects, characterized in that: Includes the following steps: Extract the temporal features of the field engineering parameter observation sequence and splice them together to generate the parameter evolution trajectory array at the current moment; Calculate the trajectory similarity value between the current parameter evolution trajectory array and the historical parameter evolution trajectory array, and confirm whether dependency drift has occurred by comparing it with a preset interval; An adjustment coefficient is generated by using the ratio of the trajectory similarity value to the historical average similarity value, thereby correcting the default checkpoint interval and updating the next checkpoint trigger time. Obtain the difference array between the parameter evolution trajectory array at the current moment and the historical parameter evolution trajectory array, and perform numerical range unification processing on the components of the difference array; The support and confidence metrics in the current association rule set are scaled and adjusted according to the processed difference components, while the feature weight component set is corrected according to the overall magnitude of the difference array. After the adjusted set of association rules and the corrected set of feature weight components are verified through tolerance range, the updated versions are written into the long-term data storage record.

2. The water conservancy project quality inspection method based on data analysis as described in claim 1, characterized in that, The steps to generate the parameter evolution trajectory array at the current moment include: Collect seepage pressure, deformation, and stress data at the site of the water conservancy project, and perform segmentation within the sliding window; Calculate the statistical feature values ​​of each sequence within the window, concatenate multiple feature values ​​in a fixed order to form an initial combination in the form of a fixed-dimensional numerical array, and store it in a memory buffer; Each time new data arrives, an append update operation is performed to add the time series feature value corresponding to the new data to the end of the array, while keeping the total number of dimensions unchanged. The combined data is then used as the parameter evolution trajectory array for the current moment.

3. The water conservancy project quality inspection method based on data analysis as described in claim 1, characterized in that, The steps to confirm whether dependency drift has occurred by comparing with a preset interval include: Before each batch processing task starts, read the parameter evolution trajectory array at the current moment; Simultaneously, retrieve the historical parameter evolution trajectory array saved at the previous checkpoint from the checkpoint storage location, calculate the spatial orientation consistency value between the current array and the historical array, and use it as the trajectory similarity value; After the calculation is completed, the trajectory similarity value is compared with the pre-set similarity threshold range for judgment. When the trajectory similarity value is lower than the lower limit of the threshold interval, it is determined that the current parameter evolution relationship has undergone dependency drift. When the value is within the threshold range, it is determined that no dependency drift has occurred.

4. The water conservancy project quality inspection method based on data analysis as described in claim 1, characterized in that, The steps to correct the default checkpoint interval and update the next checkpoint trigger time include: Retrieve the average similarity value obtained from multiple historical calculations; The adjustment coefficient is obtained by calculating the ratio between the current trajectory similarity value and the average similarity value. The adjustment coefficient value is multiplied by the default checkpoint interval duration to obtain the new checkpoint interval duration. Update the next checkpoint trigger time according to the new checkpoint interval; After all current processing tasks are completed, the updated parameter evolution trajectory array is saved to the checkpoint storage location as a historical reference array for the next cycle.

5. The water conservancy project quality inspection method based on data analysis as described in claim 1, characterized in that, The steps for performing numerical range unification on the components of the difference array include: Extract the difference array between the current latest array and the historical array saved at the most recent checkpoint, as the dependency drift amount; The difference array is input into the drift compensation process, and the drift compensation operator performs uniform processing on the numerical range of each dimension component of the difference array to find the highest and lowest values ​​of each independent component. Based on the range determined by the highest and lowest points, the numerical range of each component is unified to the standard interval of 0-1, eliminating the numerical differences caused by different parameter measurement scales, and the processed components are determined as standardized difference components.

6. The water conservancy project quality inspection method based on data analysis as described in claim 1, characterized in that, The steps for adjusting the support and confidence metrics in the current set of association rules include: Read the configuration weight ratios of each dimension obtained through historical data statistical analysis during the initial configuration phase; Iterate through the set of currently active association rules, obtain the support and confidence metrics of each rule, and scale and adjust the support and confidence metrics of each rule in the current set of association rules according to the preset weight ratio of each dimension. During the adjustment process, the direction and magnitude of the indicator adjustment are kept in a positive relationship with the standardized difference components after processing the corresponding dimension.

7. The water conservancy project quality inspection method based on data analysis as described in claim 1, characterized in that, The steps for adjusting the feature weight component set based on the overall magnitude of the difference array include: Retrieve the feature weight vector currently used by the system, locate each corresponding component in the feature weight vector, and establish a positive proportional relationship between the correction magnitude and the overall modulus of the difference array. Perform a weighted correction operation on each corresponding component in the feature weight vector, so that the correction magnitude is proportional to the overall magnitude of the difference array; Immediately after the correction is complete, a numerical overwrite is performed in the computational memory to update the feature weight vector.

8. The water conservancy project quality inspection method based on data analysis as described in claim 1, characterized in that, The steps to write the updated data to a long-term data storage record include: Retrieve the preset tolerance range value; The effectiveness of the correction operation is confirmed by comparing the change in the total number of rule items in the rule set before and after the adjustment, as well as the change in the total weight of the feature weight vector. A write command is only allowed after both changed values ​​have passed the consistency check. The updated set of association rules and the set of feature weight components are written to persistent storage medium for long-term data preservation. The complete sequence of operations described above is repeated in each subsequent incremental update cycle until an external process termination signal is received.

9. A data analysis-based water conservancy project quality inspection system, used to implement the data analysis-based water conservancy project quality inspection method according to any one of claims 1-8, characterized in that, include: The module includes: sequence extraction module, trajectory similarity calculation module, adjustment coefficient generation module, difference array acquisition module, association rule set adjustment module, and data writing module. The sequence extraction module is used to extract the temporal features of the field engineering parameter observation sequence and splice them to generate the parameter evolution trajectory array at the current moment. The trajectory similarity calculation module is used to calculate the trajectory similarity value between the parameter evolution trajectory array at the current moment and the historical parameter evolution trajectory array, and to confirm whether dependency drift has occurred by comparing it with a preset interval. The adjustment coefficient generation module is used to generate an adjustment coefficient by using the ratio of the trajectory similarity value to the historical average similarity value, thereby correcting the default checkpoint interval duration and updating the next checkpoint trigger time. The difference array acquisition module is used to acquire the difference array between the parameter evolution trajectory array at the current moment and the historical parameter evolution trajectory array, and to perform numerical range unification processing on the components of the difference array; The association rule set adjustment module is used to scale and adjust the support and confidence metrics in the current association rule set according to the processed difference components, and at the same time, adjust the feature weight component set according to the overall magnitude of the difference array. The data writing module is used to write the updated set of association rules and the corrected set of feature weight components into a long-term data storage record after passing tolerance range verification.

Images