Data management method of water conservancy and hydropower engineering based on BIM information processing

Abstract

The present invention relates to the technical field of data processing, and in particular to a water conservancy and hydropower engineering data management method based on BIM information processing, which comprises obtaining water pressure time series data at multiple positions of a dam and horizontal displacement time series data and vertical displacement time series data of the dam, segmenting the time series data to obtain multiple water pressure time series segments, multiple horizontal displacement time series segments and multiple vertical displacement time series segments, obtaining unstable time series segments in the water pressure time series segments, obtaining non-noise credibility of each unstable time series segment according to the correlation between the unstable time series segments and the associated horizontal displacement time series segments and the vertical displacement time series segments, and obtaining the final noise influence degree of each unstable time series segment based on the relationship between the non-noise credibility of each unstable time series segment, thereby obtaining a target smoothing coefficient of each unstable time series segment, and the accuracy of the obtained target smoothing coefficient is high.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a water conservancy and hydropower engineering data management method based on BIM information processing. Background Art

[0002] In water conservancy and hydropower projects, stability monitoring of dam structures is crucial. Changes in reservoir water levels will cause changes in water pressure, which in turn will generate thrust and buoyancy on the dam, resulting in horizontal and vertical displacements. In order to adjust the water storage strategy more timely to deal with safety hazards, it is necessary to predict these changes by analyzing the water pressure time series data at the dam. However, traditional methods have shortcomings in dealing with noise and outliers. Factors such as equipment failure and measurement errors can cause data fluctuations and affect data authenticity. Exponential smoothing is an effective time series prediction method that balances new and old data through adaptive smoothing coefficients, reduces the impact of noise, and improves model sensitivity and response speed. Selecting a suitable smoothing coefficient is the key. The larger the smoothing coefficient, the more the model relies on recent observations and the faster the response speed; the smaller the smoothing coefficient, the more the model relies on historical data and the smoother the prediction results. However, when using the exponential smoothing method to process the water pressure time series data of the dam, the smoothing coefficient is usually obtained empirically, with poor accuracy, making it impossible to accurately process the water pressure data. Summary of the invention

[0003] In order to solve the technical problem that when using the exponential smoothing method to process the water pressure time series data of the dam, the existing smoothing coefficient is not accurately obtained, so that the water pressure data cannot be accurately processed, the purpose of the present invention is to provide a water conservancy and hydropower engineering data management method based on BIM information processing, and the technical scheme adopted is as follows:

[0004] In a first aspect of the present invention, a method for managing water conservancy and hydropower engineering data based on BIM information processing is provided, comprising:

[0005] Obtain the water pressure time series data at multiple locations of the dam, as well as the horizontal displacement time series data and vertical displacement time series data of the dam;

[0006] Segmenting the water pressure time series data to obtain a plurality of water pressure time series segments, and correspondingly segmenting the horizontal displacement time series data and the vertical displacement time series data to obtain a plurality of horizontal displacement time series segments and a plurality of vertical displacement time series segments;

[0007] Acquire an unstable time series segment from the plurality of water pressure time series segments;

[0008] According to the correlation between the unstable timing segment and the associated horizontal displacement timing segment and vertical displacement timing segment, the non-noise credibility of the unstable timing segment is obtained;

[0009] According to the relationship between the non-noise credibility of each unstable time series segment and the associated water pressure time series segment, the final noise influence degree of each unstable time series segment is obtained;

[0010] According to the final noise influence degree of each unstable time series segment, the target smoothing coefficient of each unstable time series segment is obtained.

[0011] In an exemplary embodiment, the plurality of positions includes a primary detection position and a plurality of other auxiliary detection positions;

[0012] The water pressure time series data is segmented to obtain a plurality of water pressure time series segments, and the horizontal displacement time series data and the vertical displacement time series data are segmented accordingly to obtain a plurality of horizontal displacement time series segments and a plurality of vertical displacement time series segments, including:

[0013] Acquire a first-order difference sequence of main water pressure time series data, where the main water pressure time series data is the water pressure time series data of the main detection position;

[0014] Obtaining a data mutation degree of each data point in the first-order difference sequence, wherein the data mutation degree is related to a first-order difference value of the data point;

[0015] The mutation points are selected according to the data mutation degree of each data point, and the main water pressure time series data and the multiple auxiliary water pressure time series data are segmented according to the mutation points to obtain multiple main water pressure time series segments corresponding to the main water pressure time series data, and multiple auxiliary water pressure time series segments corresponding to each auxiliary water pressure time series data; the auxiliary water pressure time series data is the water pressure time series data of the auxiliary detection position;

[0016] The horizontal displacement time series data and the vertical displacement time series data are segmented according to the mutation points to obtain a plurality of horizontal displacement time series segments and a plurality of vertical displacement time series segments.

[0017] In an exemplary embodiment, obtaining an unstable time series segment from the plurality of water pressure time series segments includes:

[0018] According to the data mutation degree of each data point in each main water pressure time series segment, the fluctuation degree of the data mutation degree of each main water pressure time series segment is obtained;

[0019] The main water pressure time series segments whose fluctuation degree meets the preset conditions are regarded as unstable time series segments.

[0020] In an exemplary embodiment, according to the correlation between the unstable timing segment and the associated horizontal displacement timing segment and the vertical displacement timing segment, the non-noise credibility of the unstable timing segment is obtained, including:

[0021] Obtain each associated water pressure timing segment, each associated horizontal displacement timing segment and each associated vertical displacement timing segment; each associated water pressure timing segment is an auxiliary water pressure timing segment of each auxiliary detection position at the same time as the target unstable timing segment, and the target unstable timing segment is any unstable timing segment of the main detection position; each associated horizontal displacement timing segment is a horizontal displacement timing segment at the same time as each unstable timing segment, and each associated vertical displacement timing segment is a vertical displacement timing segment at the same time as each unstable timing segment;

[0022] Based on the correlation between the first water pressure time series segment and each associated horizontal displacement time series segment and each associated vertical displacement time series segment, the comprehensive initial noise influence degree of the first water pressure time series segment is obtained; the first water pressure time series segment is any one of the associated water pressure time series segments and the target unstable time series segment;

[0023] According to the comprehensive initial noise influence degree of the first water pressure time series segment and the water pressure stability at each position, the non-noise credibility of the first water pressure time series segment at each position is obtained.

[0024] In an exemplary embodiment, based on the correlation between the first water pressure time series segment and each associated horizontal displacement time series segment and each associated vertical displacement time series segment, obtaining the comprehensive initial noise influence degree of the first water pressure time series segment includes:

[0025] Obtaining the correlation between the first water pressure time series segment and each associated horizontal displacement time series segment, and calculating the average to obtain a first correlation degree;

[0026] According to the first correlation degree, obtaining a first initial noise influence degree of the first water pressure time series segment relative to the horizontal displacement;

[0027] Obtaining the correlation between the first water pressure time series segment and each associated vertical displacement time series segment, and calculating the average to obtain a second correlation degree;

[0028] According to the second correlation degree, obtaining a second initial noise influence degree of the first water pressure time series segment relative to the vertical displacement;

[0029] According to the first initial noise influence degree and the second initial noise influence degree, the comprehensive initial noise influence degree of the first water pressure time series segment is obtained.

[0030] In an exemplary embodiment, obtaining the first initial noise influence degree of the first water pressure time series segment relative to the horizontal displacement according to the first correlation degree includes:

[0031] Using the negative correlation normalized value of the first correlation degree as the first initial noise influence degree;

[0032] According to the second correlation degree, obtaining the second initial noise influence degree of the first water pressure time series segment relative to the vertical displacement includes:

[0033] The negative correlation normalized value of the second correlation degree is used as the second initial noise influence degree.

[0034] In an exemplary embodiment, according to the comprehensive initial noise influence degree of the first water pressure time series segment and the water pressure stability at each position, the non-noise credibility of the first water pressure time series segment at each position is obtained, including:

[0035] The product of the comprehensive initial noise influence degree of the first water pressure time series segment and the water pressure stability condition at the corresponding position is taken as the non-noise credibility of the first water pressure time series segment at the corresponding position.

[0036] In an exemplary embodiment, the water pressure stability conditions at each position include the water pressure stability conditions at the main detection position and the water pressure stability conditions at each auxiliary detection position;

[0037] The process of obtaining the water pressure stability of the main detection position includes: calculating the ratio of the number of stable time series segments of the main detection position to the total number of water pressure time series segments of the main detection position as the water pressure stability of the main detection position; the stable time series segments are other water pressure time series segments in the water pressure time series segments of the main detection position except the unstable time series segments;

[0038] The process of acquiring the water pressure stability of the auxiliary detection position includes: acquiring the fluctuation degree of the data mutation degree of each auxiliary water pressure timing segment according to the data mutation degree of each data point in each auxiliary water pressure timing segment of the auxiliary detection position; taking the auxiliary water pressure timing segments whose fluctuation degree does not meet the preset conditions as auxiliary stable timing segments, and calculating the ratio of the number of auxiliary stable timing segments of the auxiliary detection position to the total number of auxiliary water pressure timing segments of the auxiliary detection position as the water pressure stability of the auxiliary detection position.

[0039] In an exemplary embodiment, according to the relationship between each unstable time series segment and the non-noise credibility of the associated water pressure time series segment, the final noise impact degree of each unstable time series segment is obtained, including:

[0040] The absolute value of the difference between the non-noise credibility of the target unstable time series segment and the non-noise credibility of each associated water pressure time series segment is obtained, and then the average value is calculated to obtain the final noise influence degree of the target unstable time series segment.

[0041] In an exemplary embodiment, according to the final noise influence degree of each unstable time series segment, the target smoothing coefficient of each unstable time series segment is obtained, including:

[0042]

[0043] in, is the target smoothing coefficient of the vth unstable time series segment at the main detection position, x is the preset noise impact threshold, t is the preset parameter used to control the function growth rate, It is the final noise influence degree of the vth unstable timing segment at the main detection position, and e is a natural constant.

[0044] The present invention has the following beneficial effects: the water pressure time series data and the horizontal displacement time series data and the vertical displacement time series data are all segmented in the same way, so as to improve the reliability of subsequent data processing; then, the unstable time series segments in the water pressure time series segments are analyzed in detail; and the non-noise credibility of each unstable time series segment is obtained in combination with the correlation between the data of the corresponding time period in the horizontal displacement time series data and the vertical displacement time series data. The non-noise credibility indicates the possibility that the data of the corresponding time period is caused by non-noise; and then, according to the relationship between the non-noise credibility of the unstable time series segment and the non-noise credibility of the associated water pressure time series segment, the final noise influence degree of the unstable time series segment is obtained, so as to judge whether the data fluctuation is caused by noise or by changes in other factors in the actual environment; finally, according to the final noise influence degree of the unstable time series segment, the target smoothing coefficient of the unstable time series segment is obtained, and the target smoothing coefficient is closely related to the actual data. The obtained target smoothing coefficient has high accuracy and can better reflect the actual situation, thereby improving the accuracy and reliability of data denoising. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 A flow chart of a water conservancy and hydropower engineering data management method based on BIM information processing provided by one embodiment of the present invention;

[0046] Figure 2 Flow chart for step 2;

[0047] Figure 3 is the flowchart of step 3;

[0048] Figure 4 is the flowchart for step 4;

[0049] Figure 5 It is a flow chart of the process of obtaining the comprehensive initial noise impact degree. DETAILED DESCRIPTION

[0050] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation methods, structures, features and effects of the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0051] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0052] Building a BIM model based on water conservancy projects is conducive to creating a "digital twin" scene that is close to reality, providing a new information display effect for the construction period and management of the project. This process involves the monitoring and acquisition of multiple types of data. These data not only include traditional design and construction information, but also cover monitoring data from topography, geology to hydrology and meteorology. Among them, process management is the most important concern, and it is particularly critical for dam structure stability monitoring data. By real-time monitoring of data on changes in horizontal and vertical displacements (settlement) of the dam caused by changes in water pressure, the structural integrity and stability of the dam can be effectively evaluated, and potential safety hazards can be discovered in a timely manner. Therefore, in the process of building a BIM model, focus on the integration and analysis of monitoring data that can reflect the stability of the dam structure, to ensure that the constructed model can accurately reflect the actual operating status of the dam, and provide strong guarantees for project safety.

[0053] In water conservancy and hydropower projects, the stability of the dam structure is a key link in monitoring and management. This stability is usually reflected by the displacement data of the dam. Changes in the water level of the reservoir will cause changes in water pressure, and this pressure change will produce thrust and buoyancy on the dam, resulting in horizontal and vertical displacements. This embodiment analyzes the impact of water pressure changes during the reservoir water storage and release process on horizontal and vertical displacements to obtain the target smoothing coefficient, i.e., the adaptive smoothing coefficient, for each data point in the water pressure time series data. These adaptive smoothing coefficients are then applied to the real-time monitoring module in the BIM information processing system to achieve real-time monitoring of the stability of the reservoir bank.

[0054] This embodiment uses an exponential smoothing method to reduce the impact of data noise. This embodiment provides a water conservancy and hydropower engineering data management method based on BIM information processing, which is used to obtain an accurate smoothing coefficient in the exponential smoothing method to improve the accuracy of data denoising.

[0055] like Figure 1 As shown, the method comprises the following steps:

[0056] Step 1: Obtain the water pressure time series data at multiple locations of the dam, as well as the horizontal displacement time series data and vertical displacement time series data of the dam.

[0057] Pressure sensors are installed at multiple key locations of the dam to detect water pressure data at the corresponding locations. In addition, a dam displacement detection device is installed at a key location of the dam to detect the horizontal and vertical displacements of the dam. Horizontal displacement refers to the movement of the dam in the horizontal direction, and vertical displacement (settlement) refers to the movement of the dam in the vertical direction.

[0058] In an exemplary embodiment, the multiple key positions set at the dam include a main detection position and multiple other auxiliary detection positions. The main detection position may be the most critical detection position, and the other multiple auxiliary detection positions are other detection positions except the most critical position. The selection of these positions is determined by the actual situation of the dam.

[0059] The data at the main detection position is an object for data filtering according to the target smoothing coefficient finally obtained, and each auxiliary detection position is used for data auxiliary processing.

[0060] In this embodiment, the water pressure data obtained is the water pressure data during the water storage process, that is, the water pressure data during the process of the water level gradually rising.

[0061] In the data management of water conservancy and hydropower projects, in order to more accurately predict the changes in water pressure so as to adjust the water storage strategy in time, this embodiment obtains an adaptive smoothing coefficient, i.e., the α value, through the change characteristics of the water pressure time series data itself, thereby reducing the impact of noise and outliers in the data on the prediction results and balancing new and old data, thereby improving the sensitivity and response speed of the prediction model.

[0062] The water pressure data at each sampling time of multiple positions of the dam, as well as the horizontal displacement data and vertical displacement data of the dam, are obtained according to the same sampling period. It should be understood that the sampling times of various data are also the same. Then, for any sampling time, the water pressure data at each position of the dam, as well as the horizontal displacement data and vertical displacement data of the dam at that sampling time can be obtained.

[0063] Thus, the water pressure time series data at multiple locations of the dam, as well as the horizontal displacement time series data and vertical displacement time series data of the dam are obtained. The water pressure time series data is composed of the water pressure data at each sampling moment of the corresponding position, the horizontal displacement time series data is composed of the horizontal displacement data at each sampling moment, and the vertical displacement time series data is composed of the vertical displacement data at each sampling moment. The data collection time periods corresponding to the water pressure time series data, the horizontal displacement time series data, and the vertical displacement time series data are also the same, that is, the lengths of these time series data are the same.

[0064] Step 2: Segment the water pressure time series data to obtain multiple water pressure time series segments, and segment the horizontal displacement time series data and the vertical displacement time series data accordingly to obtain multiple horizontal displacement time series segments and multiple vertical displacement time series segments.

[0065] The size of the smoothing coefficient α depends on the relative stability of the time series data. During the process of water storage and release, as the reservoir water level rises or falls, the water pressure will gradually increase or decrease, showing the characteristics of regular changes. The speed of water pressure change helps to reflect the performance of regularity. The water storage process is usually divided into multiple stages. There will be points with large changes in the rate change characteristics. Identifying these mutation points is conducive to more accurate analysis of the changing trend of water pressure. The time series data is divided into several segments according to the mutation points. Each segment represents a relatively stable water pressure change process. The water pressure time series data is segmented, and the trend information of the relative stability of each segment can be analyzed to obtain the smoothing coefficient required for each segment.

[0066] In an exemplary embodiment, Figure 2 As shown, step 2 includes the following sub-steps:

[0067] Step 2-1: Obtain the first-order difference sequence of the main water pressure time series data.

[0068] The speed of water pressure change helps to reflect the performance of regularity, and obtains a data sequence that can reflect the rate change characteristics of water pressure time series data, that is, obtains the first-order difference sequence of the main water pressure time series data, where the main water pressure time series data is the water pressure time series data of the main detection position. The first-order difference sequence can reflect the speed of increase and decrease of water pressure over time, which helps to identify the speed of water pressure change.

[0069] Step 2-2: Obtain the data mutation degree of each data point in the first-order difference sequence. The data mutation degree is related to the first-order difference value of the data point.

[0070] The water storage process is usually divided into multiple stages. There will be points with large changes in the rate change characteristics. Identifying these mutation points is conducive to more accurate analysis of the changing trend of water pressure. The time series data is divided into several segments according to the mutation points. Each segment represents a relatively stable water pressure change process.

[0071] The data mutation degree of each data point in the first-order difference sequence is obtained, wherein the larger the first-order difference value of the data point is, the greater the degree of data change is. Therefore, the normalized result of the first-order difference value of each data point can be directly used as the data mutation degree of each data point. The normalization method in this embodiment can be: , It represents the object that needs to be normalized, and exp represents the exponential function with the natural constant e as the base.

[0072] In an exemplary embodiment, the first-order difference sequence is preliminarily divided into several sub-segments according to a preset length, which are defined as original sub-segments. Since the relative stability of the data determines the size of the smoothing coefficient, the trend change information of each original sub-segment is analyzed to obtain the stability characteristics of each original sub-segment.

[0073] Identify the maximum and minimum values ​​in each original sub-segment, calculate the difference between the maximum and minimum values, and obtain the range of each original sub-segment. The smaller the range, the more stable the data rate change of the original sub-segment. Perform linear fitting on each original sub-segment to obtain the absolute value of the slope of the fitting line. The larger the absolute value of the slope, the faster the data rate change of the original sub-segment and the more unstable it is.

[0074] Finally, for any data point in any original sub-segment, the calculation formula for the data mutation degree of the data point is as follows:

[0075]

[0076] Where i represents the i-th data point, z represents the z-th original sub-segment, is the data mutation degree of the ith data point in the zth original sub-segment, is the value of the ith data point in the zth original sub-segment, is the minimum value in the zth original sub-segment, It represents the deviation of the ith data point in the zth original subsegment relative to the minimum value in the zth original subsegment. The larger the deviation, the greater the mutation degree of the data point relative to the minimum value in the corresponding original subsegment, indicating that the data point is more likely to have data anomaly and the degree of anomaly is greater; k is the absolute value of the slope of the zth original subsegment, a is the range of the zth original subsegment; norm is the normalization function.

[0077] The calculation formula reflects the relative mutation degree of the ith data point in the zth original sub-segment, taking into account the data change amplitude (reflected by the range) and the data change rate (reflected by the absolute value of the slope) of the entire original sub-segment.

[0078] Step 2-3: Filter out mutation points according to the degree of data mutation of each data point, and segment the main water pressure time series data and multiple auxiliary water pressure time series data according to the mutation points to obtain multiple main water pressure time series segments corresponding to the main water pressure time series data, and multiple auxiliary water pressure time series segments corresponding to each auxiliary water pressure time series data.

[0079] A mutation threshold is preset, such as 0.6. Data points corresponding to data mutation levels greater than the preset mutation threshold are taken as mutation points. Thus, the time of each mutation point is obtained.

[0080] The main water pressure time series data are segmented according to the moments of each mutation point. Specifically: the water pressure data between the start moment and the moment of the first mutation point in the main water pressure time series data are taken as the first main water pressure time series segment. Subsequently, the water pressure data between the moments of two adjacent mutation points are taken as each main water pressure time series segment. Finally, the water pressure data between the moment of the last mutation point and the end moment in the main water pressure time series data are taken as the last main water pressure time series segment, thereby obtaining multiple main water pressure time series segments corresponding to the main water pressure time series data.

[0081] The auxiliary water pressure timing data is set as the water pressure timing data of the auxiliary detection position. Using the above process, each auxiliary water pressure timing data is segmented according to the moment of each mutation point, and multiple auxiliary water pressure timing segments corresponding to each auxiliary water pressure timing data are obtained. For the auxiliary water pressure timing data of any auxiliary detection position, specifically: the water pressure data between the start moment and the moment of the first mutation point in the auxiliary water pressure timing data is taken as the first auxiliary water pressure timing segment, and subsequently, the water pressure data between the moments of two adjacent mutation points is taken as each auxiliary water pressure timing segment, and finally the water pressure data between the moment of the last mutation point and the end moment in the auxiliary water pressure timing data is taken as the last auxiliary water pressure timing segment, thereby obtaining multiple auxiliary water pressure timing segments corresponding to the auxiliary water pressure timing data.

[0082] Therefore, each main water pressure timing segment and each auxiliary water pressure timing segment are divided according to the same time, and each main water pressure timing segment and each auxiliary water pressure timing segment have a corresponding relationship in time.

[0083] Step 2-4: Segment the horizontal displacement time series data and the vertical displacement time series data according to the mutation points to obtain multiple horizontal displacement time series segments and multiple vertical displacement time series segments.

[0084] Similarly, the horizontal displacement time series data is segmented according to the mutation points to obtain multiple horizontal displacement time series segments. Specifically: the horizontal displacement data between the start time and the first mutation point in the horizontal displacement time series data is taken as the first horizontal displacement time series segment. Subsequently, the horizontal displacement data between the moments of two adjacent mutation points are taken as each horizontal displacement time series segment. Finally, the horizontal displacement data between the moment of the last mutation point and the end time in the horizontal displacement time series data is taken as the last horizontal displacement time series segment, thereby obtaining multiple horizontal displacement time series segments corresponding to the horizontal displacement time series data.

[0085] The vertical displacement timing data is segmented according to the mutation points to obtain multiple vertical displacement timing segments. Specifically: the vertical displacement data between the start time and the first mutation point in the vertical displacement timing data is taken as the first vertical displacement timing segment. Subsequently, the vertical displacement data between the moments of two adjacent mutation points are taken as each vertical displacement timing segment. Finally, the vertical displacement data between the moment of the last mutation point and the end time in the vertical displacement timing data is taken as the last vertical displacement timing segment, thereby obtaining multiple vertical displacement timing segments corresponding to the vertical displacement timing data.

[0086] Step 3: Obtain unstable time series segments from a plurality of water pressure time series segments.

[0087] During the data collection process, there will be fluctuations caused by noise such as equipment failure and measurement errors, which are not true representations of the data. Therefore, after determining that multiple water pressure time series segments are obtained, in order to more accurately adapt to the realization of data management, it is necessary to further analyze the unstable segments and identify the specific causes of instability.

[0088] In an exemplary embodiment, Figure 3 As shown in the figure, the acquisition process of unstable timing segments is as follows:

[0089] Step 3-1: According to the data mutation degree of each data point in each main water pressure time series segment, the fluctuation degree of the data mutation degree of each main water pressure time series segment is obtained.

[0090] For any main water pressure time series segment, the fluctuation degree of the data mutation degree of the data points in the main water pressure time series segment is calculated according to the data mutation degree of each data point in the main water pressure time series segment. In this embodiment, the fluctuation degree is the standard deviation, that is, the standard deviation of the data mutation degree is calculated to reflect the fluctuation characteristics of the time series segment.

[0091] Thus, the fluctuation degree of each main water pressure time series segment can be obtained.

[0092] Step 3-2: The main water pressure time series segments whose fluctuation degree meets the preset conditions are regarded as unstable time series segments.

[0093] A preset condition is set, and the main water pressure time series segment whose fluctuation degree meets the preset condition is regarded as an unstable time series segment. The preset condition can be: presetting a fluctuation degree threshold, and the fluctuation degree greater than the fluctuation degree threshold; or calculating the average of the fluctuation degree of the main water pressure time series segment, and the fluctuation degree greater than the average.

[0094] In this embodiment, the main water pressure time series segments corresponding to the fluctuation degree greater than the fluctuation degree mean are taken as unstable time series segments, thereby forming an unstable time series segment set, and the unstable time series segment set includes all unstable time series segments.

[0095] In addition, the auxiliary water pressure timing segments of each auxiliary detection position in the same period as the unstable timing segment, as well as the horizontal displacement timing segment and the vertical displacement timing segment can also be obtained.

[0096] Step 4: According to the correlation between the unstable timing segment and the associated horizontal displacement timing segment and vertical displacement timing segment, the non-noise credibility of the unstable timing segment is obtained.

[0097] Analyze the unstable time series segments to identify the specific causes of instability. According to the differences in the change process between the unstable time series segments, analyze whether the data changes in each unstable time series segment are fluctuations caused by noise or real changes. Changes in water pressure will cause corresponding fluctuations in the horizontal and vertical displacements of the dam. According to the correlation between the water pressure and the fluctuations in the horizontal and vertical displacements in the same period, determine what factors affect the data changes: If the data fluctuations of the horizontal and vertical displacements are random and irregular, and have no obvious correlation with the water pressure changes, then the data fluctuations may be caused by noise. For example, equipment failures, measurement errors, etc. may cause the appearance of noise data. If the data fluctuations of the horizontal and vertical displacements show a significant correlation with the water pressure changes, then the data fluctuations may be caused by other factors in the actual environment. For example, an increase in rainfall may cause the water level in the reservoir to rise, which in turn causes changes in the horizontal and vertical displacements.

[0098] In an exemplary embodiment, Figure 4 As shown, a specific process of obtaining non-noise credibility is given as follows:

[0099] Step 4-1: Obtain each associated water pressure time series segment, each associated horizontal displacement time series segment, and each associated vertical displacement time series segment.

[0100] For ease of explanation, the target unstable timing segment is defined as any unstable timing segment of the main detection position.

[0101] Since each auxiliary detection position has an auxiliary water pressure timing segment at the same time as the target unstable timing segment, the associated water pressure timing segment is defined as the auxiliary water pressure timing segment at the auxiliary detection position at the same time as the target unstable timing segment. Therefore, each associated water pressure timing segment is obtained, and each associated water pressure timing segment is the auxiliary water pressure timing segment at each auxiliary detection position at the same time as the target unstable timing segment.

[0102] Each unstable timing segment has a horizontal displacement timing segment and a vertical displacement timing segment in the same period. The horizontal displacement timing segment in the same period as each unstable timing segment is defined as an associated horizontal displacement timing segment, and the vertical displacement timing segment in the same period as each unstable timing segment is defined as an associated vertical displacement timing segment. Therefore, each associated horizontal displacement timing segment and each associated vertical displacement timing segment are obtained.

[0103] Step 4-2: Based on the correlation between the first water pressure time series segment and each associated horizontal displacement time series segment and each associated vertical displacement time series segment, the comprehensive initial noise influence degree of the first water pressure time series segment is obtained.

[0104] The target unstable time series segment and each associated water pressure time series segment belong to water pressure data at different locations, and all belong to the same time period. For the convenience of subsequent explanation, the first water pressure time series segment is defined as any one of the associated water pressure time series segments and the target unstable time series segment.

[0105] In an exemplary embodiment, Figure 5 As shown in FIG. 1 , the process of obtaining the comprehensive initial noise impact degree includes:

[0106] Step 4-2-1: Obtain the correlation between the first water pressure time series segment and each associated horizontal displacement time series segment, and calculate the average to obtain the first correlation degree.

[0107] By obtaining the correlation between the first water pressure time series segment and each associated horizontal displacement time series segment, the correlation corresponding to each associated horizontal displacement time series segment can be obtained. Among them, the correlation can be calculated using existing similarity calculation methods, such as: cosine similarity, Pearson correlation coefficient, etc. It should be noted that when using calculation methods such as cosine similarity and Pearson correlation coefficient, the calculation results need to be normalized and classified into the 0-1 numerical range. As another implementation method, the DTW distance calculation method can also be used to first calculate the DTW distance, and then perform negative correlation normalization to obtain the correlation. The smaller the similarity, the more obvious the correlation between the water pressure and the horizontal displacement, which may be data fluctuations caused by noise. The negative correlation normalization method in this embodiment can be: .

[0108] Then, the average value of the correlation between the first water pressure time series segment and each associated horizontal displacement time series segment is calculated, and the average value is used as the first correlation degree. The smaller the first correlation degree is, the more likely the data is random and irregular, and the more likely the data fluctuation is caused by noise.

[0109] Step 4-2-2: According to the first correlation degree, obtain the first initial noise influence degree of the first water pressure time series segment relative to the horizontal displacement.

[0110] The smaller the first correlation degree is, the less obvious the correlation is between water pressure and horizontal displacement, and the more likely the data fluctuation is caused by noise, that is, the greater the impact of noise; the larger the first correlation degree is, the more stable the data is, and the smaller the impact of noise is. Therefore, the first correlation degree is inversely proportional to the first initial noise impact degree. The greater the first initial noise impact degree is, the more likely the data fluctuation is caused by noise.

[0111] In an exemplary embodiment, the negative correlation normalized value of the first correlation degree is used as the first initial noise influence degree of the first water pressure time series segment relative to the horizontal displacement.

[0112] Step 4-2-3: Obtain the correlation between the first water pressure time series segment and each associated vertical displacement time series segment, and calculate the average to obtain the second correlation degree.

[0113] Similarly, by obtaining the correlation between the first water pressure time series segment and each associated vertical displacement time series segment, the correlation corresponding to each associated vertical displacement time series segment can be obtained. The correlation is specifically described in step 4-2-1 above and will not be repeated here. The smaller the similarity, the less obvious the correlation between the water pressure and the vertical displacement, which may be caused by noise.

[0114] Then, the average value of the correlation between the first water pressure time series segment and each associated vertical displacement time series segment is calculated, and the average value is used as the second correlation degree. The smaller the second correlation degree is, the more likely the data is random and irregular, and the more likely the data fluctuation is caused by noise.

[0115] Step 4-2-4: According to the second correlation degree, obtain the second initial noise influence degree of the first water pressure time series segment relative to the vertical displacement.

[0116] The smaller the second correlation degree is, the less obvious the correlation is between water pressure and vertical displacement, and the more likely the data fluctuation is caused by noise, that is, the greater the impact of noise; the larger the second correlation degree is, the more stable the data is, and the smaller the impact of noise is. Therefore, the second correlation degree is inversely proportional to the second initial noise impact degree. The greater the second initial noise impact degree is, the more likely the data fluctuation is caused by noise.

[0117] In an exemplary embodiment, the negative correlation normalized value of the second correlation degree is used as the second initial noise influence degree of the first water pressure time series segment relative to the vertical displacement.

[0118] Step 4-2-5: According to the first initial noise impact degree and the second initial noise impact degree, obtain the comprehensive initial noise impact degree of the first water pressure time series segment.

[0119] In an exemplary embodiment, the sum of the squares of the first initial noise influence degree and the second initial noise influence degree of the first water pressure time series segment is calculated, and then the square root is taken to obtain the comprehensive initial noise influence degree of the first water pressure time series segment. The square root is used here as a method for integrating the first initial noise influence degree and the second initial noise influence degree because the square root can effectively balance the weights in the horizontal and vertical directions to avoid data in any one direction from dominating the final result too much.

[0120] By adopting the above process, the comprehensive initial noise influence degree of each time series segment in each associated water pressure time series segment and the target unstable time series segment can be obtained.

[0121] Step 4-3: According to the comprehensive initial noise influence degree of the first water pressure time series segment and the water pressure stability at each position, the non-noise credibility of the first water pressure time series segment at each position is obtained.

[0122] Step 4-2 uses the correlation between the horizontal displacement and the vertical displacement caused by water pressure, and obtains the influence of noise based on the correlation. The smoothing coefficient can be adjusted based on the comprehensive initial noise influence. However, the change characteristics of the water pressure data itself can more directly reflect its own fluctuations. There will be certain deviations when only analyzing the water pressure data at a single location. Therefore, it is necessary to further use the water pressure data at each location to analyze whether the water pressure change is caused by random noise or sudden environmental changes.

[0123] In an exemplary embodiment, the water pressure stability status of each position is first obtained. Since each position includes a main detection position and multiple other auxiliary detection positions, the water pressure stability status of each position includes the water pressure stability status of the main detection position and the water pressure stability status of each auxiliary detection position.

[0124] The process of obtaining the stability of the water pressure at the main detection position includes: the water pressure time segment at the main detection position and the unstable time segment in the water pressure time segment have been obtained in the above step 3, then the stable time segment in the water pressure time segment is obtained, and the stable time segment is the other water pressure time segment in the water pressure time segment at the main detection position except the unstable time segment. In this way, the total number of water pressure time segments at the main detection position and the number of stable time segments are obtained, and the ratio of the number of stable time segments at the main detection position to the total number of water pressure time segments at the main detection position is calculated as the stability of the water pressure at the main detection position. The larger the ratio, the more stable the water pressure.

[0125] The process of obtaining the water pressure stability of the auxiliary detection position includes: for any auxiliary detection position, in the multiple auxiliary water pressure time series segments corresponding to the auxiliary detection position, according to the relevant processes in step 2 and step 3, first obtain the data mutation degree of each data point in each auxiliary water pressure time series segment, then obtain the fluctuation degree of the data mutation degree of each auxiliary water pressure time series segment, and take the auxiliary water pressure time series segment whose fluctuation degree meets the preset conditions as the auxiliary unstable time series segment, that is, take the auxiliary water pressure time series segment whose fluctuation degree does not meet the preset conditions as the auxiliary stable time series segment. In this way, the total number of auxiliary water pressure time series segments of the auxiliary detection position and the number of auxiliary stable time series segments of the auxiliary detection position are obtained, and the ratio of the number of auxiliary stable time series segments of the auxiliary detection position to the total number of auxiliary water pressure time series segments of the auxiliary detection position is calculated as the water pressure stability of the auxiliary detection position. Using the above process, the water pressure stability of each auxiliary detection position is obtained.

[0126] Since the first water pressure timing segment is any one of the associated water pressure timing segments and the target unstable timing segment, and the target unstable timing segment belongs to one of the unstable timing segments of the main detection position, each associated water pressure timing segment is a water pressure timing segment of each auxiliary detection position that is in the same period as the target unstable timing segment. Then, when the first water pressure timing segment is the target unstable timing segment, the product of the comprehensive initial noise influence of the first water pressure timing segment and the water pressure stability of the main detection position is calculated as the non-noise credibility of the target unstable timing segment of the main detection position; when the first water pressure timing segment is any associated water pressure timing segment, the product of the comprehensive initial noise influence of the first water pressure timing segment and the water pressure stability of the auxiliary detection position to which the associated water pressure timing segment belongs is calculated as the non-noise credibility of the associated water pressure timing segment of the auxiliary detection position. Thus, the non-noise credibility of the associated water pressure timing segments of each auxiliary detection position is obtained.

[0127] If the overall stability of the data is greater, but the corresponding segmented data has a greater degree of impact from the comprehensive initial noise, it means that the data change of the segment is unstable, which may be caused by external environmental interference, and the possibility of non-noise is high. The greater the non-noise credibility, the greater the possibility that the corresponding water pressure time series segment is caused by non-noise.

[0128] The water pressure stability represents the overall stability of the data at the corresponding detection location. The greater the water pressure stability, the more stable the data at the detection location. The greater the impact of the comprehensive initial noise, the greater the impact. The comprehensive initial noise impact is multiplied by the water pressure stability. The water pressure stability is equivalent to adjusting the comprehensive initial noise impact with the data stability characteristics, so that the subsequent related analysis is more accurate.

[0129] Step 5: According to the relationship between the non-noise credibility of each unstable time series segment and the associated water pressure time series segment, the final noise impact degree of each unstable time series segment is obtained.

[0130] Through step 4, the non-noise credibility of the target unstable timing segment of the main detection position and the non-noise credibility of the associated water pressure timing segment of each auxiliary detection position are obtained. Then, the absolute value of the difference between the non-noise credibility of the target unstable timing segment of the main detection position and the non-noise credibility of the associated water pressure timing segment of each auxiliary detection position is calculated, and then the average value of these absolute values ​​of the difference is calculated, and the obtained average value is used as the final noise influence degree of the target unstable timing segment of the main detection position.

[0131] The absolute values ​​of these differences represent the difference in stability between the target unstable timing segments at the main detection position and the associated water pressure timing segments at each auxiliary detection position. The smaller the difference in stability, the more consistent the overall performance, so the less likely it is to be affected by noise, and the smaller the final degree of noise impact.

[0132] Since the target unstable timing segment is any unstable timing segment of the main detection position, the above process is used to obtain the final noise impact degree of each unstable timing segment of the main detection position. It should be understood that after obtaining the final noise impact degree of each unstable timing segment, it can be normalized and limited to a value between 0 and 1.

[0133] Step 6: According to the final noise influence degree of each unstable time series segment, obtain the target smoothing coefficient of each unstable time series segment.

[0134] According to the final noise impact degree of each unstable time series segment at the main detection position, the target smoothing coefficient adapted to the actual situation of each unstable time series segment is obtained. If the final noise impact degree is small, that is, it is more likely that the corresponding unstable time series segment is not caused by noise, a larger smoothing coefficient is required to respond to changes faster; conversely, if the final noise impact degree is large, a smaller smoothing coefficient can be selected to improve the stability of the prediction.

[0135] In an exemplary embodiment, the smoothing coefficient is obtained according to the logistic function model, as follows:

[0136]

[0137] in, is the target smoothing coefficient of the vth unstable time series segment at the main detection position, x is the preset noise influence threshold, t is the preset parameter, It is the final noise influence degree of the vth unstable timing segment at the main detection position, and e is a natural constant.

[0138] The preset noise impact degree threshold x is set according to the actual situation (for example, the value of the noise impact degree threshold x is reasonably determined according to the numerical range of the final noise impact degree of each unstable time series segment), and is set to 0.5 in this embodiment. Then, if the final noise impact degree is greater than 0.5, the target smoothing coefficient is larger.

[0139] The preset parameter t is a value greater than 1, which is used to control the growth rate of the function and is set according to actual conditions. In this embodiment, it is set to 5, which can respond to changes quickly while maintaining a certain degree of smoothness.

[0140] Therefore, the above calculation formula is used to obtain the target smoothing coefficient corresponding to each unstable time series segment of the main detection position. The target smoothing coefficient of the vth unstable time series segment of the main detection position is assigned to each data point in the vth unstable time series segment of the main detection position.

[0141] In addition, for each stable time series segment of the main detection position, the initial smoothing coefficient can be used, or a smoothing coefficient with a smaller value can be set to distinguish it from the target smoothing coefficient corresponding to each unstable time series segment mentioned above. This part is not part of the present invention and will not be described in detail.

[0142] In subsequent applications, the exponential smoothing method can be used to process the unstable time series segments of the main detection position according to the target smoothing coefficient of each data point. In addition, the target smoothing coefficient is applied to the real-time monitoring module in the BIM information processing system to monitor the stability of the reservoir bank in real time. In the BIM information processing system, the target smoothing coefficient is used to visualize the processed data. The changing trends of key indicators such as water pressure changes and horizontal displacement are intuitively presented in the form of charts, curves, etc., providing important reference information for decision makers, helping them to more accurately assess the safety status of the reservoir and formulate corresponding countermeasures.

[0143] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0144] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A water conservancy and hydropower engineering data management method based on BIM information processing, characterized in that: include: Obtain the water pressure time series data at multiple locations of the dam, as well as the horizontal displacement time series data and vertical displacement time series data of the dam; Segmenting the water pressure time series data to obtain a plurality of water pressure time series segments, and correspondingly segmenting the horizontal displacement time series data and the vertical displacement time series data to obtain a plurality of horizontal displacement time series segments and a plurality of vertical displacement time series segments; Acquire an unstable time series segment from the plurality of water pressure time series segments; According to the correlation between the unstable timing segment and the associated horizontal displacement timing segment and vertical displacement timing segment, the non-noise credibility of the unstable timing segment is obtained; According to the relationship between the non-noise credibility of each unstable time series segment and the associated water pressure time series segment, the final noise influence degree of each unstable time series segment is obtained; According to the final noise influence degree of each unstable time series segment, a target smoothing coefficient of each unstable time series segment is obtained; The plurality of positions include a main detection position and a plurality of other auxiliary detection positions; The step of obtaining a target smoothing coefficient for each unstable time series segment comprises: in, is the target smoothing coefficient of the vth unstable time series segment at the main detection position, x is the preset noise impact threshold, t is the preset parameter used to control the function growth rate, It is the final noise influence degree of the vth unstable timing segment at the main detection position, and e is a natural constant.

2. A method for managing water conservancy and hydropower engineering data based on BIM information processing as claimed in claim 1, characterized in that: The water pressure time series data is segmented to obtain a plurality of water pressure time series segments, and the horizontal displacement time series data and the vertical displacement time series data are segmented accordingly to obtain a plurality of horizontal displacement time series segments and a plurality of vertical displacement time series segments, including: Acquire a first-order difference sequence of main water pressure time series data, where the main water pressure time series data is the water pressure time series data of the main detection position; Obtaining a data mutation degree of each data point in the first-order difference sequence, wherein the data mutation degree is related to a first-order difference value of the data point; The mutation points are selected according to the data mutation degree of each data point, and the main water pressure time series data and the multiple auxiliary water pressure time series data are segmented according to the mutation points to obtain multiple main water pressure time series segments corresponding to the main water pressure time series data, and multiple auxiliary water pressure time series segments corresponding to each auxiliary water pressure time series data; the auxiliary water pressure time series data is the water pressure time series data of the auxiliary detection position; The horizontal displacement time series data and the vertical displacement time series data are segmented according to the mutation points to obtain a plurality of horizontal displacement time series segments and a plurality of vertical displacement time series segments.

3. A method for managing water conservancy and hydropower engineering data based on BIM information processing as claimed in claim 2, characterized in that: Acquiring an unstable time series segment from the plurality of water pressure time series segments, comprising: According to the data mutation degree of each data point in each main water pressure time series segment, the fluctuation degree of the data mutation degree of each main water pressure time series segment is obtained; The main water pressure time series segments whose fluctuation degree meets the preset conditions are regarded as unstable time series segments.

4. A method for managing water conservancy and hydropower engineering data based on BIM information processing as claimed in claim 3, characterized in that: According to the correlation between the unstable timing segment and the associated horizontal displacement timing segment and vertical displacement timing segment, the non-noise credibility of the unstable timing segment is obtained, including: Obtain each associated water pressure timing segment, each associated horizontal displacement timing segment and each associated vertical displacement timing segment; each associated water pressure timing segment is an auxiliary water pressure timing segment of each auxiliary detection position at the same time as the target unstable timing segment, and the target unstable timing segment is any unstable timing segment of the main detection position; each associated horizontal displacement timing segment is a horizontal displacement timing segment at the same time as each unstable timing segment, and each associated vertical displacement timing segment is a vertical displacement timing segment at the same time as each unstable timing segment; Based on the correlation between the first water pressure time series segment and each associated horizontal displacement time series segment and each associated vertical displacement time series segment, the comprehensive initial noise influence degree of the first water pressure time series segment is obtained; the first water pressure time series segment is any one of the associated water pressure time series segments and the target unstable time series segment; According to the comprehensive initial noise influence degree of the first water pressure time series segment and the water pressure stability at each position, the non-noise credibility of the first water pressure time series segment at each position is obtained.

5. A method for managing water conservancy and hydropower engineering data based on BIM information processing as claimed in claim 4, characterized in that: Based on the correlation between the first water pressure time series segment and each associated horizontal displacement time series segment and each associated vertical displacement time series segment, the comprehensive initial noise influence degree of the first water pressure time series segment is obtained, including: Obtaining the correlation between the first water pressure time series segment and each associated horizontal displacement time series segment, and calculating the average to obtain a first correlation degree; According to the first correlation degree, obtaining a first initial noise influence degree of the first water pressure time series segment relative to the horizontal displacement; Obtaining the correlation between the first water pressure time series segment and each associated vertical displacement time series segment, and calculating the average to obtain a second correlation degree; According to the second correlation degree, obtaining a second initial noise influence degree of the first water pressure time series segment relative to the vertical displacement; According to the first initial noise influence degree and the second initial noise influence degree, the comprehensive initial noise influence degree of the first water pressure time series segment is obtained.

6. A method for managing water conservancy and hydropower engineering data based on BIM information processing as claimed in claim 5, characterized in that: According to the first correlation degree, obtaining a first initial noise influence degree of the first water pressure time series segment relative to the horizontal displacement includes: Using the negative correlation normalized value of the first correlation degree as the first initial noise influence degree; According to the second correlation degree, obtaining the second initial noise influence degree of the first water pressure time series segment relative to the vertical displacement includes: The negative correlation normalized value of the second correlation degree is used as the second initial noise influence degree.

7. A method for managing water conservancy and hydropower engineering data based on BIM information processing as claimed in claim 4, characterized in that: According to the comprehensive initial noise influence degree of the first water pressure time series segment and the water pressure stability at each position, the non-noise credibility of the first water pressure time series segment at each position is obtained, including: The product of the comprehensive initial noise influence degree of the first water pressure time series segment and the water pressure stability condition at the corresponding position is taken as the non-noise credibility of the first water pressure time series segment at the corresponding position.

8. A method for managing water conservancy and hydropower engineering data based on BIM information processing as claimed in claim 7, characterized in that: The water pressure stability conditions at each position include the water pressure stability conditions at the main detection position and the water pressure stability conditions at each auxiliary detection position; The process of obtaining the water pressure stability of the main detection position includes: calculating the ratio of the number of stable time series segments of the main detection position to the total number of water pressure time series segments of the main detection position as the water pressure stability of the main detection position; the stable time series segments are other water pressure time series segments in the water pressure time series segments of the main detection position except the unstable time series segments; The process of acquiring the water pressure stability of the auxiliary detection position includes: acquiring the fluctuation degree of the data mutation degree of each auxiliary water pressure timing segment according to the data mutation degree of each data point in each auxiliary water pressure timing segment of the auxiliary detection position; taking the auxiliary water pressure timing segments whose fluctuation degree does not meet the preset conditions as auxiliary stable timing segments, and calculating the ratio of the number of auxiliary stable timing segments of the auxiliary detection position to the total number of auxiliary water pressure timing segments of the auxiliary detection position as the water pressure stability of the auxiliary detection position.

9. A method for managing water conservancy and hydropower engineering data based on BIM information processing as claimed in claim 4, characterized in that: According to the relationship between the non-noise credibility of each unstable time series segment and the associated water pressure time series segment, the final noise impact degree of each unstable time series segment is obtained, including: The absolute value of the difference between the non-noise credibility of the target unstable time series segment and the non-noise credibility of each associated water pressure time series segment is obtained, and then the average value is calculated to obtain the final noise influence degree of the target unstable time series segment.

Images