Large-span assembled detection door data deduction method

Abstract

The invention relates to the technical field of water conservancy project detection, in particular to a data deduction method for a large-span assembled detection door. According to the method, through historical wear degree data, historical water quality data and historical water pressure data of the large-span assemblable detection door, non-linear fitting of wear degree data is carried out based on a Bayesian inference method and a genetic algorithm, and corresponding slope data and use duration are obtained; a detection door influence parameter prediction model is trained through the historical data, simulation data deduction prediction is performed on the use wear degree of the large-span assembly detection door according to historical water pressure data and historical water quality data of a water conservancy detection target in a real environment, and correction is performed in combination with predicted slope data and use duration data; generating a durability curve of the wear degree; by means of the method, visualized data deduction work can be conducted on the abrasion degree under the unused condition, so that the service life of the large-span assembling detection door is measured, and follow-up generation of related use schemes and correction of used schemes are facilitated.

Description

Technical Field

[0001] The invention relates to the technical field of water conservancy engineering detection, and in particular to a data deduction method for a large-span assembleable detection door. Background Art

[0002] The large-span assembled inspection gate is a special gate structure used in large-scale water conservancy projects such as large-span dams and dam spillways. The inspection gate is mainly used to control the water flow to facilitate daily inspection and maintenance work. Due to its modular structural characteristics, it can be used multiple times across scales in different usage scenarios to save usage costs. At the same time, its modular structure also facilitates the transportation of the inspection gate and improves transportation efficiency.

[0003] Since the large-span assembled detection door needs to control the use scenario of water flow, it needs to have good structural strength and water pressure resistance; and in the long-term use process, the large-span assembled detection door is susceptible to water corrosion and impact. Therefore, with the increase of use time, the service life of the large-span assembled detection door will gradually decrease, and its durability and stability will also decrease, which will gradually affect the degree of control over the water flow.

[0004] If a large-span assembled detection door with lower durability is used, the water flow may not be stably controlled, leading to accidents. Therefore, it is necessary to conduct data deduction on the durability of the large-span assembled detection door in order to more accurately measure the service life of the large-span assembled detection door, so that relevant technical personnel can more accurately formulate or correct plans under visual data to ensure that the large-span assembled detection door can be used more safely under the corresponding plan.

[0005] Therefore, a data deduction method for large-span assembleable detection gates is proposed. Summary of the invention

[0006] The object of the present invention is to provide a data deduction method for a large-span assembled detection door, which performs nonlinear curve fitting on the wear data based on the Bayesian inference method and the genetic algorithm through the historical wear data, historical water quality data and historical water pressure data of the large-span assembled detection door, and obtains the corresponding slope data and the first usage time; establishes a detection door influencing parameter prediction model to perform model training according to these historical data, combines the historical water pressure data and the historical water quality data of the corresponding position information of the water conservancy detection target in the real environment to simulate the use wear of the large-span assembled detection door in the real use environment to perform data prediction and deduction, combines the predicted slope data and the first usage time data for correction, and generates a wear durability curve; through this method, the wear in the unused state can be visualized for data deduction to measure the service life of the large-span assembled detection door, so as to facilitate the subsequent generation of relevant usage plans and the correction of the used plans.

[0007] To achieve the above object, the present invention provides the following technical solutions: A method for deducing data of a large-span assembleable detection gate, comprising: Obtain material data of the large-span assemblable inspection door to be simulated, and obtain historical wear data, historical water quality data, and historical water pressure data based on the large-span assemblable inspection door with the same material data; generate a historical change curve based on the historical wear data, and obtain historical slope data and a first usage duration of the historical change curve; obtain corresponding data of the historical wear data, the historical water quality data, and the historical water pressure data based on a fixed time recording point T, and the data collection point of the historical slope data is the same as the fixed time recording point T; Furthermore, the wear change coordinates are generated with the historical wear data as the vertical axis coordinate and the time as the horizontal axis coordinate, and the wear change coordinates are nonlinearly fitted to generate the historical change curve; the historical change curve is nonlinearly fitted to the wear change coordinates by using Bayesian inference method and genetic algorithm; Establishing a detection door influencing parameter prediction model, and performing model training on the detection door influencing parameter prediction model according to the first usage time, the historical slope data, the historical water quality data, the historical water pressure data, and the historical wear data; Furthermore, the detection gate influence parameter prediction model includes a detection gate data encoding module and a detection gate data decoding module, and the model training process is performed in the detection gate data encoding module; The detection gate data encoding module includes a detection gate influence data processing unit, a detection gate influence data feature unit and a detection gate influence data training output unit; The detection door influence data processing unit performs data preprocessing on the historical slope data, the first usage duration, the historical water quality data, the historical water pressure data and the historical wear data to generate a plurality of detection door influence preprocessing data; The detection gate influence data feature unit extracts the feature of each detection gate influence preprocessed data to generate detection gate influence feature data; the detection gate influence data training output unit predicts slope data and wear data based on the detection gate influence feature data, and stores the detection gate influence feature data; Furthermore, the detection gate impact data feature unit includes: a Transformer network, a residual convolution network, an LSTM network and a feature fusion network; Inputting a plurality of the detection gate impact preprocessed data into the Transformer network for feature enhancement encoding; The number of the residual convolutional networks is 8, each of which includes 3 1*1 convolutional networks and 1 1*3 convolutional network; the number of the LSTM networks is 12; the number of the feature fusion networks is 2; The three residual convolutional networks and the six LSTM units are used to extract the first correlation feature of the detection gate influence preprocessing data of the first usage duration, the historical slope data, the historical water quality data, the historical water pressure data and the historical wear data after feature encoding; The three residual convolution networks and the four LSTM unit networks are used to extract the second associated features of the detection gate impact preprocessing data corresponding to the first usage duration, the historical water quality data, the historical water pressure data and the historical wear data after feature encoding; The two residual convolutional networks, the two LSTM units and the one feature fusion network are used to extract the third correlation feature of the detection gate influence preprocessing data corresponding to the first usage duration, the historical slope data and the historical wear data; The first correlation feature and the second correlation feature are fused through one of the feature fusion networks; The first correlation feature and the third correlation feature are fused through one of the feature fusion networks; Acquire the first usage time, water quality data and water pressure data corresponding to the location information of the water conservancy detection target; Furthermore, the water conservancy inspection target is simulated by a digital twin model, and the position information is all installable positions of the large-span assemblable inspection door; and obtaining wear prediction data and slope prediction data according to the detection door influencing parameter prediction model, fitting a first durability change curve according to the wear prediction data; and correcting the first durability change curve according to the slope prediction data; Further, the wear degree prediction data and the slope prediction data are predicted according to the detection gate data decoding module of the detection gate influencing parameter prediction model; the first durability change curve is nonlinearly fitted by Bayesian inference method and genetic algorithm; The detection gate data decoding module includes a detection gate durability prediction processing unit, a detection gate durability prediction coding connection unit and a detection gate durability prediction feature extraction and output unit; The durability prediction processing unit performs data preprocessing on the first usage duration, the historical water quality data and the historical water pressure data to generate durability prediction preprocessing data; The detection gate durability prediction coding connection unit is connected to the detection gate data coding module; the detection gate durability prediction feature extraction and output unit extracts the durability prediction feature of the durability prediction preprocessing data in combination with the detection gate data coding module; and predicts the wear degree prediction data and the slope prediction data based on the durability prediction feature; The detection gate durability prediction feature extraction and output unit includes: 1 Transformer network, 5 residual convolution networks and 6 LSTM networks; 1 Transformer network extracts the feature enhancement coding of the durability prediction preprocessing data, and performs feature association with the features of the detection gate durability prediction coding link unit; 5 residual convolution networks and 6 LSTM networks extract the features of the durability prediction preprocessing data after feature association, and perform prediction output; Furthermore, the detection gate data decoding module generates a plurality of prediction intervals based on the first usage duration, and each prediction interval is associated with the corresponding historical water quality data and the historical water pressure data; The nth prediction will use the wear prediction data of the n-1th time as prediction input data; the nth prediction input data includes the wear prediction data, the first usage time, the historical water quality data and the historical water pressure data, and the total number of the prediction intervals is N. , ; According to the wear data of the large-span assemblable detection door with the same material data and the second use time, a second durability change curve is generated by fitting; and the first durability change curve is corrected twice to generate a third durability change curve; Furthermore, the second use time is used as the horizontal axis, and the wear data is used as the vertical axis, and the second durability change curve is fitted by using Bayesian inference method and genetic algorithm; The first durability change curve is corrected twice to generate a third durability change curve as follows: ; in, The third durability change curve is the first use time The degree of wear, The first durability change curve is the first usage time The degree of wear, The second usage time on the second durability change curve Corresponding wear degree; The durability of the large-span assemblable inspection door is simulated and deduced based on the third durability change curve and the second durability change curve.

[0008] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention collects historical wear data, historical water quality data and historical water pressure data of large-span assembled detection doors, and performs nonlinear curve fitting on the wear data based on Bayesian inference method and genetic algorithm. This method can understand the wear consumption data of large-span assembled detection doors of various materials under the influence of different water pressures and different water qualities based on existing data, and obtain corresponding slope data. On the one hand, these historical data can provide a large amount of training data for subsequent service life simulation data deduction and prediction. At the same time, according to the slope of the curve fitted by the existing data, data correction can also be performed on the prediction results to prevent the error of the model deduction from being too large, resulting in a decrease in the accuracy of the result, thereby improving the reliability of data deduction.

[0009] 2. The present invention uses an encoder and decoder structure to perform feature training on the prediction data in the encoder. In order to improve the accuracy of the prediction, a combined model of LSTM and residual convolution is used based on time series data to perform comprehensive feature learning from the perspective of time series and data morphology, and the trained features are stored; when prediction is actually required, feature extraction is performed through the decoder unit, and the stored features of the encoder unit are called for comprehensive learning to improve the accuracy of the prediction; this method can be used to perform modular splitting in the training and prediction process, and the training model and the prediction model can be independently optimized to adapt to the data deduction and prediction work in different actual scenarios, so as to improve the accuracy of the service life simulation deduction prediction, and also facilitate the subsequent independent optimization of the model and parameters for training and prediction work, so as to further improve the accuracy of the service life simulation deduction prediction.

[0010] 3. The present invention obtains prediction results through service life simulation and deduction of the prediction model, combines Bayesian inference method and genetic algorithm to fit the data to generate a curve, combines the predicted slope data to perform data correction of the curve, and combines the data relationship between the usage time and the wear degree to perform data correction of the quadratic curve; this method can visualize the data based on the prediction and deduction results, so as to assist relevant technical personnel in generating corresponding solutions or corrections to the sampled solutions based on the visualization results; at the same time, other life-influencing data are also generated into corresponding curves for visualization operations, so as to further facilitate the auxiliary work of relevant technical personnel in data analysis, and at the same time, the independent visualization based on each data is also convenient for the optimization of the entire data deduction and prediction work in the later stage, so as to improve the accuracy of the service life simulation and deduction prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 It is a schematic diagram of the method flow of the present invention; Figure 2 It is a schematic diagram of the digital twin of a hydropower station of the present invention; Figure 3It is a schematic diagram of the process flow of the detection gate data encoding module of the present invention; Figure 4 It is a schematic diagram of the process flow of the detection gate data decoding module of the present invention; Figure 5 A schematic diagram of a model of a detection gate data encoding module and a detection gate data decoding module of the present invention; Figure 6 Schematic diagram of the residual convolution group structure of the present invention; Figure 7 It is a model flow diagram of the comparative model M03 of the present invention; Figure 8 It is the first durable variation curve of the large-span assemblable detection door of the present invention; Fig. 9 The third durability variation curve diagram of the large-span assemblable detection door of the present invention; Fig.10 It is a durability curve diagram of the large-span assemblable detection door of the present invention. DETAILED DESCRIPTION

[0012] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0013] Due to its modular structural characteristics, the large-span assemblable inspection door is easy to transport and suitable for a variety of scenarios. It can be used in a large span, is convenient for water flow control in multiple areas, and realizes maintenance and transportation of relevant personnel; but due to its modular structural characteristics, the stability requirements of the large-span assemblable inspection door are also relatively high. Due to the long-term influence of water pressure and water corrosion, its stability will also decrease with the increase of usage time. Therefore, it is necessary to perform data deduction on the durability of the large-span assemblable inspection door in order to accurately measure the service life of the large-span assemblable inspection door. For this purpose, the present invention provides a data deduction method for a large-span assemblable inspection door, referring to Figure 1 As shown, the technical solution is as follows: Obtain material data of the large-span assemblable inspection door to be simulated, and obtain historical wear data, historical water quality data, and historical water pressure data based on the large-span assemblable inspection door with the same material data; generate a historical change curve based on the historical wear data, and obtain historical slope data and a first usage duration of the historical change curve; Establishing a detection door influencing parameter prediction model, and performing model training on the detection door influencing parameter prediction model according to the first usage time, the historical slope data, the historical water quality data, the historical water pressure data, and the historical wear data; Acquire the first usage time, water quality data and water pressure data corresponding to the location information of the water conservancy detection target, and acquire wear prediction data and slope prediction data according to the detection door influencing parameter prediction model, and fit the first durability change curve according to the wear prediction data; and correct the first durability change curve according to the slope prediction data; According to the wear data of the large-span assemblable detection door with the same material data and the second use time, a second durability change curve is generated by fitting; and the first durability change curve is corrected twice to generate a third durability change curve; The durability of the large-span assemblable inspection door is simulated and deduced based on the third durability change curve and the second durability change curve.

[0014] In order to be able to perform data deduction of the large-span assembled inspection door, the material data of the large-span assembled inspection door to be simulated is used to obtain the historical wear data of the large-span assembled inspection door that has been used and made of the same material, and the historical water quality data and historical water pressure data under the corresponding use environment are obtained. A change curve of the historical wear data is established through these data. The change curve can be used to understand the change of wear during actual use. Therefore, the slope data of each point is collected through the curve as the actual change data, and the first use time data is obtained through the curve as one of the parameters of the wear change under different use times, so that in the process of data deduction, the data can be corrected to achieve a more accurate data deduction result; At the same time, based on the collected data, the parameter prediction model of the detection door can be used to conduct the training phase of data prediction and deduction. By using the duration data, wear data, water quality data, water pressure data and slope data as time series data input, the model training and learning parameters can directly hide the data characteristics and the change characteristics of the data itself, and store the data characteristics learned by the model; in combination with the need for deduction scenarios, the large-span assembleable detection door can be used at various positions of the target for use. Based on the water pressure data and water quality data collected in the real environment, combined with the set first usage duration data, data prediction has been achieved. The change relationship of the wear degree is calculated based on multiple prediction results to generate the first durability curve, so as to intuitively understand the wear data of the large-span assembleable detection door, facilitate the generation of subsequent related plans and the correction of the generated plans, and provide more intuitive visual data; In order to further improve the accuracy, the second durability change curve is established through the relationship between the second usage time and the wear degree. The data is further corrected to eliminate the influence of the second usage time on the wear degree to display the durability change curve in the actual process. This method can visualize the usage data under different circumstances to further assist the work of relevant technical personnel. It can also intuitively display the service life, facilitate the debugging of the existing model, and improve the accuracy and precision of the data prediction and deduction results.

[0015] Embodiment 1 In order to elaborate on the model training and prediction process, the following examples are used for illustration: Obtain material data (such as concrete, metal, etc.) of the large-span assemblable inspection door to be simulated, and obtain historical wear data, historical water quality data, and historical water pressure data based on the large-span assemblable inspection door with the same material data; generate a historical change curve based on the historical wear data, and obtain historical slope data and a first usage time of the historical change curve; the first usage time is the usage time of the large-span assemblable inspection door during maintenance and use; the second usage time is the usage time that generates wear data under assembly and transportation conditions; The historical wear data is obtained by using the yolov5 model to identify the wear characteristics on the large-span assemblable inspection door caused by water pressure or water corrosion. The specific calculation is as follows: ; in, is the wear data, M is the total number of images, is the number of characteristic pixels of water pressure or water corrosion wear of the i-th image identified by yolov5, is the total number of pixels in the i-th image, Based on a large amount of data combined with expert experience, the wear data of water pressure or water corrosion is and is the wear weight, which defaults to 0.5 and 0.5, and can be modified according to the specific implementation situation; the minimum wear data is 0, which represents no wear, and the maximum wear data is 1, which represents complete wear; Establishing a detection door influencing parameter prediction model, and performing model training on the detection door influencing parameter prediction model according to the first usage time, the historical slope data, the historical water quality data, the historical water pressure data, and the historical wear data; Furthermore, the detection gate influence parameter prediction model includes a detection gate data encoding module and a detection gate data decoding module, and the model training process is performed in the detection gate data encoding module; The detection gate data encoding module includes a detection gate influence data processing unit, a detection gate influence data feature unit and a detection gate influence data training output unit. Figure 3 As shown; the detection door impact data processing unit performs data preprocessing on the historical slope data, the first usage duration, the historical water quality data, the historical water pressure data and the historical wear data to generate a plurality of detection door impact preprocessing data; The detection gate influence data feature unit extracts the feature of each detection gate influence preprocessed data to generate detection gate influence feature data; the detection gate influence data training output unit predicts slope data and wear data based on the detection gate influence feature data, and stores the detection gate influence feature data; In order to facilitate the training and prediction process of the model, the model is split into a coding module part and a decoding module part. The coding module can be trained and extract features based on existing historical data, and the decoding part performs real training. Considering that in the prediction simulation and deduction of actual data, only water pressure data and water quality data simulating the real environment are input, the training and prediction processes are split. During the training process, the first usage time data, historical water pressure data, historical water quality data, historical slope data and historical wear are used to extract features through the model training process to obtain the hidden feature connections between the data and the change trends of each data, so as to deeply mine the features of each data to improve the accuracy of training prediction and the depth of storage data features, so as to provide a solid model feature foundation for the subsequent simulation and deduction data process and improve the accuracy of the subsequent simulation and deduction process; Furthermore, the detection gate influence data feature unit includes: Transformer network, residual convolution network, LSTM network and feature fusion network, refer to Figure 5 The detection gate affects the data encoding module as shown in the part; Inputting a plurality of the detection gate impact preprocessed data into the Transformer network for feature enhancement encoding; The number of the residual convolutional networks is 8, each of which includes 3 1*1 convolutional networks and 1 1*3 convolutional network; the residual convolutional network refers to Figure 6 As shown; the number of LSTM networks is 12; the number of feature fusion networks is 2; The three residual convolutional networks and the six LSTM units are used to extract the first correlation feature of the detection gate influence preprocessing data of the first usage duration, the historical slope data, the historical water quality data, the historical water pressure data and the historical wear data after feature encoding; The three residual convolution networks and the four LSTM unit networks are used to extract the second associated features of the detection gate impact preprocessing data corresponding to the first usage duration, the historical water quality data, the historical water pressure data and the historical wear data after feature encoding; The two residual convolutional networks, the two LSTM units and the one feature fusion network are used to extract the third correlation feature of the detection gate influence preprocessing data corresponding to the first usage duration, the historical slope data and the historical wear data; The first correlation feature and the second correlation feature are fused through one of the feature fusion networks; The first correlation feature and the third correlation feature are fused through one of the feature fusion networks; In this embodiment, we obtained historical wear data, historical water quality data and historical water pressure data through relevant technical personnel. After the slope data was obtained through processing by the above method, the training process of the model detection gate data encoding module was carried out, and multiple groups of data were used for training. Each group of data will generate multiple results. Therefore, the average of the results of each group of data will be used as the accuracy of the group of data, and the model parameters with the highest accuracy in the training process are taken for storage. Due to too much data, this table only shows part of the results. The training accuracy results are shown in Table 1: Table 1. Accuracy of simulation training of detection gate data encoding module ; From Table 1, we can see that except for individual data with an accuracy rate lower than 90%, the overall data training accuracy rate is higher than 90%, and the training results are good; During the training process of the model, the global hidden correlation features on the wear impact correlation and the wear change trend are first obtained through the first usage time, the historical slope data, the historical water quality data, the historical water pressure data and the historical wear data; then the local impact correlation weight features of the first usage time, the historical water pressure data, the historical water quality data and the historical wear data on the wear are obtained, and at the same time, the local impact change trend features of the wear data that changes with time between the first usage time, the wear data and the slope are learned during the training process; through this training process, the data can be trained from the global data and local data dimensions to achieve the purpose of more accurately grasping the data characteristics, not only from the feature correlation between the data, but also from the change dimension of the data itself to improve the accuracy of the subsequent simulation and deduction process; Acquire the first usage time, water quality data and water pressure data corresponding to the location information of the water conservancy detection target; and acquire wear prediction data and slope prediction data according to the detection door influencing parameter prediction model, and fit the first durability change curve according to the wear prediction data; and correct the first durability change curve according to the slope prediction data; Further, the wear degree prediction data and the slope prediction data are predicted according to the detection gate data decoding module of the detection gate influencing parameter prediction model; the first durability change curve is nonlinearly fitted by Bayesian inference method and genetic algorithm; The detection gate data decoding module includes a detection gate durability prediction processing unit, a detection gate durability prediction coding connection unit and a detection gate durability prediction feature extraction and output unit. Figure 4 As shown; The durability prediction processing unit performs data preprocessing on the first usage duration, the historical water quality data and the historical water pressure data to generate durability prediction preprocessing data; The detection gate durability prediction coding connection unit is connected to the detection gate data coding module; the detection gate durability prediction feature extraction and output unit extracts the durability prediction feature of the durability prediction preprocessing data in combination with the detection gate data coding module; and predicts the wear degree prediction data and the slope prediction data based on the durability prediction feature; In the decoding unit, since the data for prediction and deduction are only water quality data and water pressure data, in addition to extracting the environmental data for simulation and deduction, the data features stored in the training process are also added, and the data prediction work is performed based on the associated feature data of wear training added to the environmental data for simulation and deduction, so as to predict the wear data corresponding to the feature association of the environmental data, so as to achieve the purpose of predicting the wear data and realize the requirements for the life data deduction of the large-span assemblable detection door; The detection gate durability prediction feature extraction and output unit includes: 1 Transformer network, 5 residual convolution networks and 6 LSTM networks; 1 Transformer network extracts the feature enhancement coding of the durability prediction preprocessing data, and performs feature association with the features of the detection gate durability prediction coding link unit; 5 residual convolution networks and 6 LSTM networks extract the features of the durability prediction preprocessing data after feature association, and perform prediction output; In this embodiment, some data that have not participated in the training are selected to simulate the real environment data for deduction and prediction. The prediction accuracy is obtained by the difference with the real data. Combined with the expert experience, the prediction accuracy is calculated as 100% if the error is within 2%. By analogy, the accuracy calculation formula is as follows: ; in, is the accuracy, is the predicted value, is the true value; This embodiment uses 5 groups of data that have not participated in the training as data for data deduction in a simulated real environment. The prediction accuracy results of the detection gate data decoding module are shown in Table 2: Table 2 Prediction accuracy results of the detection gate data decoding module ; This embodiment also compares the model M01 of this article, and only uses the detection gate data encoding module M02, the detection gate data encoding module and the detection gate data decoding module as the overall model M03, M01 refers to Figure 5 As shown, M02 reference Figure 5 The detection door affects the data encoding module as shown, M03 reference Figure 7 As shown, the prediction data is compared, and the specific results using data A01 and A02 are shown in Table 3: Table 3 Comparison results of model data prediction and deduction

[0016] As can be seen from Table 3, the model M01 in this paper has the highest wear prediction accuracy and slope prediction accuracy, and can be stabilized at around 90%, with good prediction effect; Furthermore, the detection gate data decoding module generates a plurality of prediction intervals based on the first usage duration, and each prediction interval is associated with the corresponding historical water quality data and the historical water pressure data; The nth prediction will use the wear prediction data of the n-1th time as prediction input data; the nth prediction input data includes the wear prediction data, the first usage time, the historical water quality data and the historical water pressure data, and the total number of the prediction intervals is N. , ; In the actual prediction process, multiple first usage time parameters can be set in combination with expert experience, and these first usage time parameters can be formed into a new time series and combined with water quality data and water pressure data for input. In the actual operation of the model, the model will use the first usage time as the prediction interval, and make predictions based on the water quality data and water pressure data under each first usage time; and in the prediction process of the first usage time of the second node, the predicted wear data of the previous node will be substituted into the next prediction interval for input. This method can realize multiple result predictions to obtain multiple data for the convenience of subsequent curve generation and visualization operations. At the same time, by inputting the first usage time to generate the number of predictions, it is also convenient for relevant technical personnel to perform quantitative operations, improve the operating space of model prediction, and make it more convenient for relevant technical personnel to deduce the durability data of large-span assembleable inspection doors, and can more accurately measure the deduction of its service life, so as to facilitate the modification or generation of corresponding technical solutions. Based on the wear data and the second usage time of the large-span assembled inspection door with the same material data, a second durability change curve is fitted and generated; and the first durability change curve is corrected for the second time to generate a third durability change curve; based on the third durability change curve and the second durability change curve, the durability of the large-span assembled inspection door is simulated and deduced.

[0017] Embodiment 2 The curve fitting process is described by the following examples: Obtain material data of the large-span assemblable inspection door to be simulated, and obtain historical wear data, historical water quality data, and historical water pressure data based on the large-span assemblable inspection door with the same material data; generate a historical change curve based on the historical wear data, and obtain historical slope data and a first usage duration of the historical change curve; The historical wear data, the historical water quality data and the historical water pressure data obtain corresponding data according to the fixed time recording point T, and the data collection point of the historical slope data is the same as the fixed time recording point T; by default, the first usage duration is the time interval between each time recording point; In order to obtain the change of wear data, so that the wear data of the large-span assemblable detection door can be visualized in the process of data deduction, water quality data, wear data and water pressure data are formed into time series data through multiple recording points during data collection, so as to be displayed on the coordinates. At the same time, the corresponding slope data and the first use time data are obtained based on the curve fitting data as additional parameters for later data deduction. This method can quantify the data, facilitate analysis and prediction, and provide a data basis for subsequent data deduction calculations. At the same time, more influencing parameters can be obtained from the existing data, which improves the analysis dimension of data deduction and the accuracy of subsequent deductions. Furthermore, the wear change coordinates are generated with the historical wear data as the vertical axis coordinate and the time as the horizontal axis coordinate, and the wear change coordinates are nonlinearly fitted to generate the historical change curve; the fitted curve can be used to understand the influence of different water pressure data and water quality data on the wear during actual use of the wear data, and based on the data, the slope data at different time points can be extracted to explore the hidden data relationship between the wear data and the data affecting it, so as to increase the dimension of the analyzable data and improve the accuracy of subsequent deductions; The historical change curve is nonlinearly fitted to the wear degree change coordinates by using Bayesian inference method and genetic algorithm; The Bayesian inference method can be used to perform nonlinear fitting based on existing data, and the genetic algorithm can optimize the fitted nonlinear curve, making the fitted curve more accurate. At the same time, the curve optimized by the Bayesian inference method and the genetic algorithm can obtain a more accurate slope change relationship, which is convenient for providing more accurate slope data, so as to facilitate the subsequent feature training and data correction, so as to improve the data accuracy of the subsequent deduction; Establishing a detection door influencing parameter prediction model, and performing model training on the detection door influencing parameter prediction model according to the first usage time, the historical slope data, the historical water quality data, the historical water pressure data, and the historical wear data; Acquire the first usage time, water quality data and water pressure data corresponding to the location information of the water conservancy detection target; Furthermore, the water conservancy detection target (such as dams, hydropower stations, etc., refer to Figure 2 The digital twin of the hydropower station shown in the figure), the location information is all the installable locations of the large-span assembleable detection door; the water conservancy detection target is simulated by the digital twin model to facilitate the establishment of a data model, and all actual data are digitized based on the simulated virtual data to facilitate the subsequent deduction work. At the same time, based on the position annotation of the digital twin model, it is also convenient for relevant technical personnel to simulate the data deduction process of the large-span assembleable detection door to achieve visual and digital display; and obtaining wear prediction data and slope prediction data according to the detection door influencing parameter prediction model, fitting a first durability change curve according to the wear prediction data; and correcting the first durability change curve according to the slope prediction data; Further, the first durability change curve is nonlinearly fitted by Bayesian inference method and genetic algorithm; According to the wear data of the large-span assemblable detection door with the same material data and the second use time, a second durability change curve is generated by fitting; and the first durability change curve is corrected twice to generate a third durability change curve; Furthermore, the second use time is used as the horizontal axis, and the wear data is used as the vertical axis, and the second durability change curve is fitted by using Bayesian inference method and genetic algorithm; Classify according to historical assembly time and historical transportation time, perform data fitting based on historical assembly time and wear data to generate a second durability change curve for assembly time, perform data fitting based on historical transportation time and wear data to generate a second durability change curve for transportation time, combine the average of multiple historical data and expert experience to determine the organization time and transportation time under different prediction intervals, and obtain corresponding wear data according to the corresponding second durability change curve; The wear data generated during assembly and transportation is calculated in the same way as the wear data of water pressure or water corrosion; that is, the wear data generated during assembly and transportation is obtained by comprehensive weighting through the yolov5 recognition features and the expert experience of a large amount of data; The first durability change curve is corrected twice to generate a third durability change curve as follows: ; in, The third durability change curve is the first use time The degree of wear, The first durability change curve is the first usage time The degree of wear, The second usage time on the second durability change curve Corresponding wear degree; In this embodiment, relevant technicians used some untrained data to perform data deduction and simulation, and set a total of three first use times, the first first use time was 185 minutes, the second first use time was 95 minutes, and the third first use time was 90 minutes, to simulate the deduction and simulation of the same large-span assembled detection door under continuous use, and the generated first durability change curve is as follows: Figure 8 As shown, the third durability change curve generated after the second durability change curve is corrected is as follows Fig. 9 As shown; A second durability change curve is generated based on the second usage time and wear data to correct the first durability change curve. Based on this solution, the wear of the large-span assembleable inspection door data caused by non-water pressure and water corrosion, such as wear during transportation or wear during organization, can be excluded. Through this method, the wear of the large-span assembleable inspection door under different conditions can be further understood to estimate the wear of the large-span assembleable inspection door, so as to evaluate the service life of the large-span assembleable inspection door after the current first usage time. At the same time, the change of the second usage time such as transportation can also be visualized, which is convenient for the relevant technical personnel to formulate and modify the plan and improve the accuracy of data deduction; The durability of the large-span assemblable inspection door is simulated and deduced according to the third durability change curve and the second durability change curve; durability The calculation is: ; In this embodiment, 1 represents the initial durability in the unworn state; the wear degree corresponds to the consumed durability; In this embodiment, based on Figure 8 and Fig. 9 The results of the long-span assembly inspection door are used to generate a visual durability curve. Fig.10 shown.

[0018] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for deducing data of a large-span assembleable detection door, characterized in that: include: Obtain material data of the large-span assemblable inspection door to be simulated, and obtain historical wear data, historical water quality data, and historical water pressure data based on the large-span assemblable inspection door with the same material data; generate a historical change curve based on the historical wear data, and obtain historical slope data and a first usage duration of the historical change curve; Establishing a detection door influencing parameter prediction model, and performing model training on the detection door influencing parameter prediction model according to the first usage duration, the historical slope data, the historical water quality data, the historical water pressure data, and the historical wear data; Acquire the first usage time, water quality data and water pressure data corresponding to the location information of the water conservancy detection target, and acquire wear prediction data and slope prediction data according to the detection door influencing parameter prediction model, and fit the first durability change curve according to the wear prediction data; and correct the first durability change curve according to the slope prediction data; According to the wear data of the large-span assemblable detection door with the same material data and the second use time, a second durability change curve is generated by fitting; and the first durability change curve is corrected twice to generate a third durability change curve; The durability of the large-span assemblable inspection door is simulated and deduced based on the third durability change curve and the second durability change curve.

2. A method for deducing data of a large-span assembleable detection door according to claim 1, characterized in that: The historical wear data, the historical water quality data and the historical water pressure data obtain corresponding data based on a fixed time recording point T, and the data collection point of the historical slope data is the same as the fixed time recording point T.

3. A method for deducing data of a large-span assembleable detection door according to claim 1, characterized in that: Generating a historical change curve by fitting the historical wear data includes: generating wear change coordinates with the historical wear data as the vertical axis coordinate and time as the horizontal axis coordinate, and performing nonlinear fitting on the wear change coordinates to generate the historical change curve.

4. A method for deducing data of a large-span assembleable detection door according to claim 3, characterized in that: The wear degree variation coordinates are nonlinearly fitted by using Bayesian inference method and genetic algorithm.

5. The method for deducing data of a large-span assembleable detection door according to claim 1 is characterized in that: The water conservancy detection target is simulated by a digital twin model, and the position information is all installable positions of the large-span assembled detection door.

6. A method for deducing data of a large-span assembleable detection door according to claim 1, characterized in that: Establishing a detection door influencing parameter prediction model, and training the detection door influencing parameter prediction model according to the first usage time, the historical slope data, the historical water quality data, the historical water pressure data, and the historical wear data includes: The detection gate influence parameter prediction model includes a detection gate data encoding module and a detection gate data decoding module, and the model training process is performed in the detection gate data encoding module; The detection gate data encoding module includes a detection gate influence data processing unit, a detection gate influence data feature unit and a detection gate influence data training output unit; The detection door influence data processing unit performs data preprocessing on the historical slope data, the first usage duration, the historical water quality data, the historical water pressure data and the historical wear data to generate a plurality of detection door influence preprocessing data; The detection gate influence data feature unit extracts the features of each detection gate influence preprocessed data to generate detection gate influence feature data; the detection gate influence data training output unit predicts slope data and wear data based on the detection gate influence feature data, and stores the detection gate influence feature data.

7. A method for deducing data of a large-span assembleable detection door according to claim 6, characterized in that: The detection gate impact data feature unit includes: Transformer network, residual convolution network, LSTM network and feature fusion network, specifically: Inputting a plurality of the detection gate impact preprocessed data into the Transformer network for feature enhancement encoding; The number of the residual convolutional networks is 8, each of which includes 3 1*1 convolutional networks and 1 1*3 convolutional network; the number of the LSTM networks is 12; the number of the feature fusion networks is 2; The three residual convolutional networks and the six LSTM units are used to extract the first correlation feature of the detection gate influence preprocessing data of the first usage duration, the historical slope data, the historical water quality data, the historical water pressure data and the historical wear data after feature encoding; The three residual convolution networks and the four LSTM unit networks are used to extract the second associated features of the detection gate impact preprocessing data corresponding to the first usage duration, the historical water quality data, the historical water pressure data and the historical wear data after feature encoding; The two residual convolutional networks, the two LSTM units and the one feature fusion network are used to extract the third correlation feature of the detection gate influence preprocessing data corresponding to the first usage duration, the historical slope data and the historical wear data; The first correlation feature and the second correlation feature are fused through one of the feature fusion networks; The first associated feature and the third associated feature are fused through one of the feature fusion networks.

8. The method for deducing data of a large-span assembleable detection door according to claim 1 is characterized in that: Acquiring the first usage time, water quality data, and water pressure data corresponding to the location information of the water conservancy detection target, acquiring wear prediction data and slope prediction data according to the detection door influencing parameter prediction model, and fitting a first durability change curve according to the wear prediction data; and correcting the first durability change curve according to the slope prediction data includes: Predicting the wear degree prediction data and the slope prediction data according to the detection door data decoding module of the detection door influencing parameter prediction model; nonlinearly fitting the first durability change curve by Bayesian inference method and genetic algorithm; The detection gate data decoding module includes a detection gate durability prediction processing unit, a detection gate durability prediction coding connection unit and a detection gate durability prediction feature extraction and output unit; The durability prediction processing unit performs data preprocessing on the first usage duration, the historical water quality data and the historical water pressure data to generate durability prediction preprocessing data; The detection gate durability prediction coding connection unit is connected to the detection gate data coding module; the detection gate durability prediction feature extraction and output unit extracts the durability prediction feature of the durability prediction preprocessing data in combination with the detection gate data coding module; and predicts the wear degree prediction data and the slope prediction data based on the durability prediction feature; The detection gate durability prediction feature extraction and output unit includes: 1 Transformer network, 5 residual convolution networks and 6 LSTM networks; 1 Transformer network extracts the feature enhancement coding of the durability prediction preprocessing data, and performs feature association with the features of the detection gate durable prediction coding link unit; 5 residual convolution networks and 6 LSTM networks extract the features of the durability prediction preprocessing data after feature association, and perform prediction output.

9. A method for deducing data of a large-span assembleable detection door according to claim 8, characterized in that: The prediction output of the detection gate data decoding module also includes: The detection gate data decoding module generates a plurality of prediction intervals based on the first usage duration, each prediction interval being associated with the corresponding historical water quality data and the historical water pressure data; The nth prediction will use the wear prediction data of the n-1th time as prediction input data; the nth prediction input data includes the wear prediction data, the first usage time, the historical water quality data and the historical water pressure data, and the total number of the prediction intervals is N. , .

10. The method for deducing data of a large-span assembleable detection door according to claim 1, characterized in that: According to the wear data of the large-span assemblable detection door with the same material data and the second use time, a second durability change curve is generated by fitting; And secondarily correcting the first durability change curve to generate a third durability change curve comprises: With the second usage time as the horizontal axis and the wear data as the vertical axis, the second durability change curve is fitted by Bayesian inference method and genetic algorithm; The first durability change curve is corrected twice to generate a third durability change curve as follows: ； in, The third durability change curve is the first use time The degree of wear, The first durability change curve is the first usage time The degree of wear, The second usage time on the second durability change curve The corresponding wear degree.

Images