Method for monitoring slump of poured and stirred concrete on line

Abstract

The invention provides an online slump monitoring method for poured and stirred concrete, which comprises the following steps: arranging a plurality of concrete surface moisture sensors and wind speed sensors on a concrete pouring site, presetting the sampling frequency of the sensors, and measuring the surface and environment of covered concrete within a preset range; obtaining water evaporation rate data and wind speed data of the concrete surface; based on the statistical characteristic quantities of the different wind speed modes, the statistical characteristic quantities of the wind speed modes are used as input, the statistical characteristic quantities of the evaporation rates in the corresponding wind speed modes are used as output, and prediction models in the different wind speed modes are constructed; and after the collected wind speed data is newly added, judging a wind speed mode corresponding to the newly added wind speed data, and obtaining a predicted value of the evaporation rate of the water on the surface of the concrete according to the prediction model in the corresponding wind speed mode.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of information technology, and in particular to a method for online monitoring of slump of casted and mixed concrete. BACKGROUND

[0002] Slump of concrete is an important indicator of workability of concrete, which reflects the plasticizing performance and pumpability of concrete, and is crucial to ensure normal construction. Evaporation of water on the surface of concrete will directly lead to a decrease in free water in the concrete clinker, and further cause slump loss of concrete. Especially in the early stage of pouring, if water evaporation is too fast, it will lead to surface water loss and cause plastic shrinkage cracks, thereby reducing the slump and affecting the workability. On the contrary, proper water retention helps to maintain the slump and ensure good construction performance and final strength. Factors such as humidity, natural wind speed and temperature in the natural environment will affect the evaporation rate of water in concrete. Therefore, it is crucial to monitor the surface water evaporation rate and natural wind speed fluctuation in real time at the construction site of concrete pouring. However, the wind speed in the natural environment has complex non-stationary patterns such as gust and turbulence, and the relationship between these patterns and the evaporation rate is not clear. How to accurately identify these non-stationary natural wind speed patterns and correlate them with evaporation rate data has become a technical problem to be solved. Specifically, an algorithm needs to be developed to extract high-frequency and low-frequency components from continuous natural wind speed time series and determine the occurrence time and duration of gust and turbulence according to a preset threshold. At the same time, a model needs to be established to link the statistical characteristics under different natural wind speed patterns with the evaporation rate in the corresponding period to realize the prediction of evaporation rate. In order to effectively solve this problem, research not only needs to go deep into the fields of signal processing and pattern recognition, but also needs to overcome the technical obstacles of time synchronization and short-time wind speed change capture. At the same time, in the complex construction site environment, how to ensure that the data collected by the sensor is accurate and real-time is also one of the factors that must be considered. The ultimate goal is to dynamically adjust the construction strategy through an optimized monitoring system to maintain the ideal slump of concrete and ensure that the engineering quality and progress are not affected. SUMMARY

[0003] The present application provides a method for online monitoring of slump of casted and mixed concrete, mainly comprising:

[0004] A plurality of concrete surface moisture sensors and wind speed sensors are arranged at the concrete pouring site, a preset sensor sampling frequency is set, the concrete surface and the environment within a preset range are measured, and concrete surface water evaporation rate data and wind speed data are obtained;

[0005] The collected concrete surface moisture evaporation rate data and wind speed data are uploaded to a data processing server in real time and stored in a time sequence format to form moisture evaporation time sequence data and wind speed time sequence data, and the wind speed time sequence data is processed into wind speed data of different scales, and the wind speed data is divided into high-frequency components and low-frequency components according to different scales;

[0006] The high-frequency components and low-frequency components of the wind speed data are compared with preset gust identification thresholds and turbulence identification thresholds respectively to determine whether gusts and turbulence occur in the wind speed data, and if so, the start and end times and the duration of the gusts and turbulence are recorded, and the time distribution of the gust wind speed mode and the turbulence wind speed mode is obtained according to the recorded start and end times and the duration of the wind speed, to form a wind speed mode subsequence;

[0007] The moisture evaporation time sequence data is divided according to the time distribution of different wind speed modes to obtain evaporation rate subsequences corresponding to different wind speed modes, and the influence of different wind speed modes on the moisture evaporation rate is quantified by comparing the mean and standard deviation of the evaporation rate subsequences under different wind speed modes to obtain statistical characteristic quantities of the evaporation rate under different wind speed modes;

[0008] Based on the statistical characteristic quantities of different wind speed modes, a prediction model under different wind speed modes is constructed with the statistical characteristic quantities of the wind speed mode as input and the evaporation rate statistical characteristic quantities under the corresponding wind speed mode as output;

[0009] After new wind speed data is added, the wind speed mode corresponding to the new wind speed data is determined, and the prediction value of the concrete surface moisture evaporation rate is obtained according to the prediction model under the corresponding wind speed mode;

[0010] A concrete slump prediction model is constructed in advance, the prediction value of the concrete surface moisture evaporation rate is taken as input, the moisture evaporation rate data is analyzed and predicted, and the prediction value of the concrete slump is obtained, and if the predicted concrete slump value is outside or below the preset qualified range, the concrete slump is determined to be unqualified, and a warning signal is issued;

[0011] After the warning signal is triggered, the adjustment scheme of the concrete proportioning is determined according to the value of the concrete slump prediction value outside or below the preset qualified range, the concrete proportioning parameters of the adjustment scheme are input into the concrete production equipment, and the surface moisture evaporation rate data of the adjusted concrete is continuously obtained, returned to the concrete slump prediction model input, and the online prediction of the concrete slump is performed next time.

[0012] The technical scheme provided by the embodiment of the application can include the following beneficial effects:

[0013] The application discloses a kind of online monitoring of pouring and mixing concrete slump degree method.The method is by laying sensor in concrete pouring site, real-time acquisition concrete surface moisture evaporation rate and wind speed data, and utilize wavelet transform to identify different wind speed mode.Based on these data, the application constructs the correlation model of wind speed mode and moisture evaporation rate, and then predicts concrete slump degree.When the predicted slump degree is unqualified, the application automatically triggers early warning and gives concrete proportion adjustment scheme, realizes the closed-loop control to concrete quality.This method innovatively combines environmental factors, material characteristics and production process closely, realizes the accurate prediction and real-time regulation of concrete slump degree, effectively improves concrete construction quality and efficiency, and provides new technical path for intelligent concrete production. BRIEF DESCRIPTION OF DRAWINGS

[0014] Fig. 1 The flow chart of the online monitoring of pouring and mixing concrete slump degree method of the application.

[0015] Fig. 2 The schematic diagram of the online monitoring of pouring and mixing concrete slump degree method of the application.

[0016] Fig. 3 The schematic diagram of the online monitoring of pouring and mixing concrete slump degree method of the application. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical scheme and advantages of the application more clear, the application is described in detail below with reference to the drawings and specific embodiments.

[0018] As Figs. 1-3 , the online monitoring of pouring and mixing concrete slump degree method of the embodiment can specifically include:

[0019] Step S101, several concrete surface moisture sensors and wind speed sensors are laid in concrete pouring site, preset sensor sampling frequency, measure the covering concrete surface and environment in preset range, obtain concrete surface moisture evaporation rate data and wind speed data.

[0020] According to sensor sampling area grid division diagram, concrete surface is divided into measurement sub-regions, and time series monitoring values of sensor array are acquired using data acquisition module;The water distribution map and the wind speed distribution map of the concrete surface are obtained by performing spatial interpolation operation on the time series monitoring values using the Kriging interpolation algorithm;The water evaporation rate data of the measurement points are obtained by dividing the water content difference of adjacent sampling time in the water distribution map by the sampling time interval;If the wind speed sensor detection value is greater than the preset wind speed threshold value, then the water evaporation rate data are denoised using Gaussian filtering to obtain the corrected water evaporation rate data.

[0021] Specifically, according to the sensor sampling area grid division map, the concrete surface is divided into a plurality of measurement sub-areas, a concrete surface moisture sensor and a wind speed sensor are arranged at the center point of each measurement sub-area, and a data acquisition module is used to periodically collect data of the sensor array in the area at a preset sampling time interval to obtain time series monitoring values. The sensor calibration module is used to correct the zero drift and compensate the linearity of the collected concrete surface moisture sensor and wind speed sensor data, and the Kriging interpolation algorithm is used to calculate the spatial interpolation of the data between the measurement sub-areas to obtain the moisture distribution map and the wind speed distribution map of the entire concrete surface. The time series monitoring values are subjected to moving average filtering to eliminate random fluctuations, and the moisture evaporation rate data of each measurement point is calculated according to the difference between the moisture contents of the concrete surface at adjacent sampling times divided by the sampling time interval. If the wind speed sensor detection value is greater than the preset wind speed threshold value, the moisture evaporation rate data is denoised by Gaussian filtering to obtain the corrected moisture evaporation rate data, wherein the preset wind speed threshold value is determined based on the risk of rapid loss of moisture on the concrete surface. The recursive least squares method is used to perform regression operation on the corrected moisture evaporation rate data and wind speed data to establish a linear regression equation of the moisture evaporation rate and the wind speed, wherein the independent variable is the wind speed data and the dependent variable is the moisture evaporation rate data. According to the established linear regression equation, the slope coefficient and the intercept coefficient between the moisture evaporation rate and the wind speed of the concrete surface are calculated to form a quantitative relationship parameter matrix reflecting the moisture evaporation characteristics. In the concrete construction site, the sensor arrangement adopts a grid division method, and the grid is divided according to a length of 0.5 meters for each measurement sub-area. For a concrete surface of 10 meters x 10 meters, a total of 400 measurement sub-areas are divided. A concrete surface moisture sensor produced by the German Vika Company and a three-dimensional ultrasonic wind speed sensor produced by the American Campbell Company are arranged at the center point of each measurement sub-area, and the sensor sampling time interval is set to 1 minute. During the sensor calibration process, a standard solution is used to calibrate the concrete surface moisture sensor at three points, and the moisture content of the standard solution is 5%, 10%, and 15% respectively, to obtain a linear relationship curve between the sensor output voltage and the moisture content. The wind speed sensor is calibrated using a standard wind tunnel, and the wind speed range is 0 to 20 meters / second, with a calibration point every 2 meters / second to obtain the corresponding relationship between the output signal and the wind speed. The spatial interpolation uses the Kriging algorithm, which takes into account the spatial correlation between measurement points and describes the spatial correlation characteristics through a variogram model. The variogram model uses a spherical model with a mathematical expression of γ(h) = C O +C[1.5(h / a)-0.5(h / a) 3 ], where h is the sampling point spacing, C OThe block gold value is 0.01, the base value of C is 0.95, and a is the variable range of 5 meters. For the collected time series data, a 5-point moving average method is used to eliminate random fluctuations. In calculating the water evaporation rate, the difference between the water content of the adjacent two sampling times is divided by the sampling time interval of 60 seconds to obtain the instantaneous evaporation rate. When the wind speed exceeds 5 meters / second, a Gaussian filter is used to denoise the water evaporation rate data, and the standard deviation of the Gaussian filter is set to 1.2. In establishing the relationship between water evaporation rate and wind speed, recursive least squares method is used for fitting, and the regression equation form is y=kx+b, where y is the water evaporation rate (grams / meter2hour), x is the wind speed (meters / second), k is the slope coefficient, and b is the intercept coefficient. The typical parameter values obtained by fitting the experimental data are k=2.35, b=0.82, and the correlation coefficient R 2 =0.94, indicating that the wind speed and water evaporation rate have strong linear correlation. In practical application, for a measurement sub-area, when the wind speed is 3 meters / second, the water evaporation rate calculated according to the regression equation is 7.87 grams / meter2hour. When the wind speed increases to 8 meters / second, the water evaporation rate rises to 19.62 grams / meter2hour, which is within ±5% of the relative error of the actual measured value, verifying the accuracy of the regression model.

[0022] Step S102, the collected concrete surface water evaporation rate data and wind speed data are uploaded to the data processing server in real time, and are stored in time series format to form water evaporation time series data and wind speed time series data. The wind speed time series data is processed into wind speed data of different scales, and the wind speed data is divided into high-frequency components and low-frequency components according to different scales.

[0023] The concrete surface monitoring device is used to obtain water evaporation rate data and wind speed data, and the water evaporation rate data and wind speed data are marked with time stamps according to a preset sampling period and then compressed and transmitted to a data processing server; the compressed data is received and decompressed using a time series database, and missing data points are interpolated using a cubic spline interpolation algorithm according to the time stamp to obtain complete time series data; the complete time series data is preprocessed using a median filter, and measurement noise and outliers in the time series data are removed through filtering operations to obtain preprocessed wind speed data; the preprocessed wind speed data is decomposed using a db4 wavelet basis function, the instantaneous frequency characteristics of the wind speed data are calculated through Hilbert transform, and the wind speed data is divided into high-frequency wind speed components and low-frequency wind speed components using a Butterworth band-pass filter bank.

[0024] Specifically, according to the concrete surface moisture evaporation rate data and wind speed data in the data temporary storage area in the acquisition device, time stamp marks are added to the data according to a preset sampling period, the data with time stamps is compressed and packaged by using a lossless compression algorithm, and then transmitted to a data processing server through a network. The compressed data packet is decompressed and restored by using a time series database, and moisture evaporation rate time series data and wind speed time series data are respectively generated according to the time stamp marks, and missing data points are interpolated by using a cubic spline interpolation algorithm to obtain complete time series. The time series database stores data by using a time stamp index structure. The complete time series data is preprocessed by median filtering to remove measurement noise and outliers, and the preprocessed wind speed time series data is obtained. The preprocessed wind speed time series data is subjected to discrete wavelet transform multi-scale decomposition by using a db4 wavelet basis function, and the signal is subjected to four-layer iterative decomposition according to a preset decomposition layer, and wavelet coefficients and scale coefficients corresponding to each layer of decomposition are obtained. The wind speed component data of each scale is reconstructed according to the obtained wavelet coefficients and scale coefficients, and the instantaneous frequency characteristics of the wind speed components of each scale are calculated by using Hilbert transform. The wind speed component data is divided into frequency ranges according to the instantaneous frequency characteristics, and the wind speed components of different frequency ranges are screened by using a Butterworth bandpass filter set, to obtain high-frequency wind speed component data and low-frequency wind speed component data, and the energy proportion corresponding to each frequency component is recorded. During the data acquisition process of the concrete surface moisture evaporation rate and wind speed, the data temporary storage area adopts a ring buffer structure, the buffer size is set to 1024 data points, the sampling period is 1 second, and data compression transmission is triggered when the buffer data amount reaches 512 data points. The data compression adopts run-length encoding in the lossless compression algorithm, and for continuous numerical values such as 3.5, 3.5, 3.5, 3.6 and 3.6 of the wind speed data, the compressed values are (3.5, 3) and (3.6, 2), achieving a compression ratio of 2:1. The time series database stores data by using a time stamp index tree structure, the time stamp accuracy is millisecond level, and the leaf nodes of the index tree store specific data values. When data is missing, data completion is performed by using cubic spline interpolation, and the interpolation polynomial form is f(x) = ax 3 +bx 2 + cx+ d, where a, b, c, d are undetermined coefficients, which are obtained by solving a linear equation set. For the adjacent 4 data points (0, 1.2), (1, 1.5), (3, 2.1), (4, 2.4), the obtained interpolation function can be used to calculate the missing value 1.8 at x = 2. The data preprocessing adopts 5-point median filtering. For the sequence data [1.2, 1.5, 5.8, 1.8, 2.1], the original value 5.8 is replaced by the middle value 1.8 to eliminate the mutation interference. The db4 wavelet basis function is selected for four-layer wavelet decomposition, and the basis function has good time-frequency localization characteristics with a support length of 7. For wind speed data with a length of 1024 points, the first layer decomposition obtains 512 low-frequency coefficients and 512 high-frequency coefficients, the second layer decomposition obtains 256 low-frequency coefficients and 256 high-frequency coefficients, and so on. In the frequency feature analysis, the Hilbert transform is used to calculate the instantaneous frequency. Assuming that the original signal is x(t), its Hilbert transform is y(t), then the analytic signal z(t) = x(t) + jy(t), j is the imaginary unit, the instantaneous phase φ(t) = arctan(y(t) / x(t)), and the instantaneous frequency f(t) = (1 / 2π)·dφ(t) / dt. For the wind speed signal, the calculated instantaneous frequency distribution range is 0-0.5Hz. Based on the instantaneous frequency characteristics, a 4th order Butterworth band-pass filter set is designed to divide 0-0.1Hz into a low frequency band and 0.1-0.5Hz into a high frequency band. The filter frequency response function H(ω) = 1 / [1+(ω / ωc)8], where ωc is the cutoff frequency and ω is the angular frequency. Through energy statistics, it is found that the low-frequency component accounts for 75% of the total energy, reflecting that the wind speed change is mainly low-frequency and slow.

[0025] Step S103, comparing the high-frequency component and the low-frequency component of the wind speed data with the preset gust recognition threshold and the turbulence recognition threshold respectively, judging whether gust and turbulence appear in the wind speed data, if so, recording the start and end time and the duration of the gust and turbulence, obtaining the time distribution of the gust wind speed mode and the turbulence wind speed mode according to the recorded start and end time and the duration of the wind speed, and forming a wind speed mode subsequence.

[0026] A frequency separator is used to separate the high-frequency component and the low-frequency component of the wind speed data, and a gust event marker is obtained according to the comparison of the high-frequency component with the threshold value, and a turbulence event marker is obtained according to the comparison of the low-frequency component with the threshold value; a time window processor is used to segment the wind speed data, and an event count value in each time window is obtained according to the gust event marker and the turbulence event marker; feature extraction is performed on the wind speed data in the time window, and a gust feature sequence and a turbulence feature sequence are calculated according to the event count value; a dynamic programming algorithm is used to time-align the gust feature sequence and the turbulence feature sequence, and a wind speed mode feature vector is generated according to the aligned feature sequences.

[0027] Specifically, the amplitude comparator is used to compare the high-frequency component of the wind speed data with the preset gust identification threshold value, and when the amplitude of the high-frequency component exceeds the threshold value, the start point of the gust is marked, and when the amplitudes of the high-frequency components of five consecutive sampling points are lower than the threshold value, the end point of the gust is marked, and the peak value and the duration of the high-frequency component of the wind speed in this time period are recorded. The low-frequency component of the wind speed data is compared with the preset turbulence identification threshold value, and the root mean square value of the low-frequency component is calculated as the turbulence cumulative intensity index, and when the turbulence cumulative intensity index exceeds the preset turbulence identification threshold value, the start point of the turbulence is marked, and when the index decreases below the threshold value, the end point of the turbulence is marked, and the root mean square value and the duration of the low-frequency component in this time period are recorded. According to the recorded time information of the gust and turbulence events, a fixed-length time window is set, and the wind speed data is processed in segments, and the number of occurrences and the cumulative duration of gusts and turbulence in each time window are counted. In each time window, the wind speed data is extracted for features, the gust peak value coefficient and the turbulence intensity coefficient are calculated, and the gust feature sequence and the turbulence feature sequence are generated. The dynamic programming algorithm is used to time-align the gust feature sequence and the turbulence feature sequence, and the wind speed mode sub-sequence is generated according to the time distribution law of the feature sequence. According to the combination relationship of the gust features and the turbulence features in the wind speed mode sub-sequence, the wind speed mode feature vector is established, and the time distribution of various wind speed modes is recorded. In the wind speed data processing, the determination of the gust identification threshold value and the turbulence identification threshold value is based on a large amount of historical data statistical analysis. The gust identification threshold value is set to 2.5 times the standard deviation of the wind speed, and for an average wind speed of 5 meters / second, the standard deviation is 0.8 meters / second, and the threshold value is set to 2.0 meters / second. The turbulence identification threshold value is based on the turbulence intensity index, and the value is 0.15, that is, when the wind speed fluctuation amplitude exceeds 15% of the average wind speed, it is determined as turbulence. In the actual application scenario, the original wind speed data sequence [4.8, 4.9, 7.2, 7.5, 7.3, 4.7, 4.8] meters / second, after high-frequency component extraction, [0.1, 0.2, 2.4, 2.7, 2.5, -0.1, 0.0] meters / second is obtained. When the amplitudes of three consecutive high-frequency components exceed 2.0 meters / second, a gust event is identified, and the duration is 3 sampling periods. At the same time, the low-frequency component sequence [4.7, 4.7, 4.8, 4.8, 4.8, 4.8, 4.8] meters / second represents the slow-changing trend of the wind speed. The turbulence cumulative intensity is calculated by the root mean square value, and for the above low-frequency component sequence, the root mean square value in a 10-second time window is 0.05, which is lower than the turbulence identification threshold value 0.15, indicating that there is no turbulence in this period. In another data sequence, when there is a low-frequency fluctuation such as [4.5, 4.8, 5.2, 5.5, 5.2, 4.8, 4.5], the root mean square value reaches 0.18, which exceeds the threshold value, and a turbulence event is identified. In the feature extraction process, the gust peak value coefficient is defined as the ratio of the maximum instantaneous wind speed to the average wind speed, such as the maximum wind speed 7.5 meters / second and the average wind speed 4.8 meters / second, the peak value coefficient is 1.56.The turbulence intensity coefficient is calculated by dividing the standard deviation of the wind speed fluctuation by the average wind speed. In the sample data, the standard deviation is 0.8 m / s, the average wind speed is 5.0 m / s, and the turbulence intensity coefficient is 0.16. The dynamic programming algorithm realizes the time alignment of the feature sequence by constructing a state transition matrix. Taking 2-minute data as an example, 12 10-second time windows are set, and the gust feature [1.56, 1.52, 1.48] is detected in the 3rd, 4th, and 5th windows, and the turbulence feature [0.16, 0.15, 0.14] is detected in the 4th, 5th, and 6th windows, indicating that the gust and turbulence have time overlap. The generated wind speed pattern feature vector contains gust peak coefficient, duration, turbulence intensity coefficient, and overlap degree, which are used to characterize the time-varying characteristics of wind speed.

[0028] In step S104, the water evaporation time series data is divided according to the time distribution of different wind speed patterns to obtain evaporation rate sub-sequences corresponding to different wind speed patterns. By comparing the mean and standard deviation of the evaporation rate sub-sequences under different wind speed patterns, the influence of different wind speed patterns on the evaporation rate of water is quantified to obtain statistical characteristic quantities of the evaporation rate under different wind speed patterns.

[0029] According to the time stamp alignment method, the water evaporation time series data is segmented, and the boundary data is supplemented at the segmentation points by using the cubic spline interpolation method to obtain the water evaporation rate sub-sequence satisfying the first derivative continuity constraint; the trapezoidal integral method is used to perform integral operation on the water evaporation rate sub-sequence, and the arithmetic mean of the water evaporation rate sub-sequence is calculated by the integral operation result to obtain the instantaneous evaporation rate in the corresponding time period; the water evaporation rate sub-sequence is identified according to the box plot method, and the linear interpolation method is used to correct the abnormal points to obtain the corrected water evaporation rate sub-sequence; the central moment of the corrected water evaporation rate sub-sequence is calculated by using the moment estimation method, the central moment is standardized by using the maximum and minimum value normalization method, the standardized feature vector is grouped according to the hierarchical clustering method, and the water evaporation rate statistical feature combination under different wind speed patterns is obtained.

[0030] Specifically, based on the temporal distribution data of wind speed patterns, the water evaporation time series data is segmented using timestamp alignment. At the segmentation points, cubic spline interpolation is used to supplement boundary data, with the interpolation function satisfying the first derivative continuity constraint at the boundary points, generating water evaporation rate subsequences corresponding to different wind speed patterns. The trapezoidal integral method is used to calculate the instantaneous evaporation rate of each water evaporation rate subsequence within the corresponding time period, and the arithmetic mean of the water evaporation rate subsequences is calculated based on the integration results. Outlier removal is performed on the water evaporation rate subsequences, using box plots to identify outlier data points, which are then corrected using linear interpolation to obtain the corrected water evaporation rate subsequences. The moment estimation method is used to calculate the central moments of the corrected subsequences, including the mean, variance, third central moment, and fourth central moment. The central moment data is standardized using the minimax normalization method. Based on the standardized central moment data, water evaporation rate feature vectors are constructed, and Euclidean distance is used to measure the similarity between feature vectors. A bottom-up hierarchical clustering method was used to group the feature vectors, and the inter-class and intra-class distances were calculated using the Ward minimum variance criterion. Based on the clustering results, statistical feature combinations of water evaporation rates under different wind speed modes were determined, and the mean and standard deviation of each combination were recorded. For the segmented processing of the water evaporation rate time series, cubic spline interpolation was used at boundary points to ensure data continuity. Taking the wind speed mode switching point t = 120 seconds as an example, two data points were taken before and after the switching point to form a data group.

[0031] Given the polynomial {(118,2.5),(119,2.6),(120,2.8),(121,3.2),(122,3.3)}, the interpolation function f(t) = 0.002t is obtained by solving the coefficients of the third-order polynomial. 3 -0.015t 2+ 1.8, which satisfies the first-order derivative continuity at the boundary points. In the transient evaporation rate calculation, the trapezoidal integration method is used to process the sub-sequence. For the sequence data [2.5, 2.8, 3.1, 2.9, 2.7] with a time interval of 1 second, the integral values of adjacent points are (2.5 + 2.8) * 0.5 = 2.65, (2.8 + 3.1) * 0.5 = 2.95, and so on, to obtain the integral value of the complete sequence, and then the average evaporation rate 2.82 g / m2·s is calculated. The outlier identification uses the box plot method, and the quartiles of the sequence data are calculated as Q1 = 2.6, Q2 = 2.8, Q3 = 3.0, the interquartile range IQR = 0.4, and the outlier judgment range is set as [Q1 - 1.5IQR, Q3 + 1.5IQR], i.e. [2.0, 3.6]. For the value 4.2 in the original sequence, which exceeds the upper limit value 3.6, it is determined as an outlier, and is corrected to 3.3 by linear interpolation. The central moment calculation reflects the statistical characteristics of the data. For the sequence [2.5, 2.8, 3.1, 2.9, 2.7], the first-order central moment (mean) μ = 2.8, the second-order central moment (variance) σ 2 = 0.045, the third-order central moment (skewness) γ = 0.12, and the fourth-order central moment (kurtosis) κ = 2.3. The maximum and minimum value normalization method is used to map each order central moment to the interval [0, 1]. The Euclidean distance calculation of the feature vector reflects the difference in evaporation characteristics under different wind speed modes. The Euclidean distance d between two feature vectors v1 = [0.3, 0.5, 0.2, 0.4] and v2 = [0.5, 0.6, 0.3, 0.5] is d = √[(0.3 - 0.5) 2 + (0.5 - 0.6) 2 + (0.2 - 0.3) 2 + (0.4 - 0.5) 2 ] = 0.283, and the smaller the distance, the more similar the characteristics. In the hierarchical clustering process, the ward minimum variance criterion is used to merge the categories. For categories A, B, and C composed of three feature vectors, the inter-class distances are D(A, B) = 5.2, D(B, C) = 3.8, and D(A, C) = 6.4, and the categories B and C with the smallest distance are selected for merging. Through the iterative merging process, the final water evaporation characteristic combination under the stable, turbulent, and gust wind speed modes is obtained, in which the mean evaporation rate under the stable wind speed mode is 2.8 g / m2·s and the standard deviation is 0.2 g / m2·s.

[0032] Step S105, based on the statistical characteristic quantities of different wind speed modes, the statistical characteristic quantities of wind speed modes are taken as input, and the evaporation rate statistical characteristic quantities under the corresponding wind speed mode are taken as output, to construct a prediction model under different wind speed modes.

[0033] The feature quantity dataset is grouped according to the wind speed mode category label, the maximum and minimum value standardization method is used to process the feature quantity dataset, and the standardized wind speed mode feature quantity and evaporation rate feature quantity are obtained; the correlation between the standardized feature quantities is calculated using the Pearson correlation coefficient, and if the absolute value of the correlation coefficient is greater than a preset threshold, the reduced wind speed mode feature quantity and evaporation rate feature quantity are obtained; the reduced feature quantity dataset is stratified sampled, and the training set and the validation set are obtained by dividing according to the training set proportion, the training validation data combination is obtained by resampling the dataset using the cross-validation method; and a random forest prediction model is constructed based on the training validation data combination, the optimal parameter combination is searched in the preset parameter grid space, and the root mean square error and the determination coefficient of the validation set are calculated by using the optimal parameter combination.

[0034] Specifically, the statistical feature quantity dataset is grouped according to the wind speed mode category label, and each group of feature data is normalized. The maximum and minimum value standardization method is used to map the feature quantity to the interval [0, 1] to obtain the standardized wind speed mode feature quantity and evaporation rate feature quantity. The correlation between the feature quantities is calculated using the Pearson correlation coefficient, and the feature combination with an absolute value of the correlation coefficient greater than a preset threshold is selected to generate the reduced wind speed mode feature quantity and evaporation rate feature quantity. The reduced dataset is stratified sampled, and the training set and the validation set are divided according to the training set proportion of 0.8. The training set is resampled using the five-fold cross-validation method to generate multiple training validation data combinations. A random forest prediction model is constructed, the number of decision trees and the maximum depth of the tree parameters are set, and the root mean square error is used as the optimization objective function. Each parameter combination is searched in the preset parameter grid space, the prediction model is trained for each parameter combination, and the validation error is recorded. The parameter combination with the smallest validation error is selected as the optimal parameter. Based on the optimal parameter, a prediction model is constructed for each wind speed mode, the complete training set is used to train the model, and the tree structure parameters and node splitting rules of the model are recorded. The performance of the trained prediction model is evaluated using the validation set, the root mean square error and the determination coefficient between the predicted value and the true value are calculated, and a model validation report is generated. In the standardization process of the wind speed mode feature quantity, the original feature data [15.2, 18.5, 12.8, 22.4, 16.7] is normalized using the maximum and minimum value standardization formula x' = (x-xmin) / (xmax-xmin) to obtain the standardized feature value [0.25, 0.59, 0.0, 1.0, 0.41]. For the evaporation rate feature quantity, the original data

[0035] The values ​​[2.8, 3.2, 2.5, 3.8, 3.0] were standardized to obtain [0.23, 0.54, 0.0, 1.0, 0.38]. During feature selection, the Pearson correlation coefficient matrix was calculated, with the correlation coefficient r = (Σ(x-μx)(y-μy)) / (σxσy), where μx and μy are the means, σx and σy are the standard deviations, and x and y represent the data sequences of the two variables or features. Taking peak wind speed and evaporation rate as an example, the calculated correlation coefficient was 0.85, exceeding the preset threshold of 0.6, so this feature was retained. Conversely, the correlation coefficient between wind speed duration and evaporation rate was only 0.32, below the threshold, so this feature was removed. In dataset partitioning, stratified sampling was used to maintain the proportion of sample categories. The original dataset contains 400 stable wind speed samples, 300 turbulent wind samples, and 200 gust wind samples. After splitting the dataset by a 0.8 ratio, the training set contains 320 stable wind speed samples, 240 turbulent wind samples, and 160 gust wind samples. Five-fold cross-validation divides the training set into five equal parts, selecting four parts for training and one part for validation each time. Parameter optimization of the random forest prediction model uses a grid search method, searching for the number of decision trees within the range of [50, 100, 150, 200] and the maximum tree depth within the range of [3, 5, 7, 9], resulting in 16 parameter combinations. The model is trained for each set of parameters, and the root mean square error is calculated. Where yi is the actual value. Here, n represents the predicted value, and n is the number of samples. The optimal parameter combination was obtained through search, resulting in 150 decision trees with a maximum depth of 7. Taking a stable wind speed model as an example, the random forest model with the optimal parameters contains 150 decision trees, each with a maximum depth of 7 layers. In the first tree, the root node splits based on wind speed peak value x1 ≤ 0.5, the left child node splits based on turbulence intensity x2 ≤ 0.3, the right child node splits based on duration x3 ≤ 0.8, and so on, forming a complete tree structure. Model validation uses two metrics: root mean square error (RMSE) and coefficient of determination (COP). On the validation set, the prediction results for the stable wind speed model show RMSE = 0.15 and R² = 0.15. 2 =0.92; the prediction result of the turbulence model is RMSE=0.18, R 2 =0.89; the RMSE of the gust model prediction result is 0.21, R 2 =0.86. The predictive performance of the three models differs, which is related to the different data complexity and sample size of each model.

[0036] Step S106: After adding new wind speed data, determine the wind speed mode corresponding to the new wind speed data, and obtain the predicted value of the water evaporation rate on the concrete surface according to the prediction model under the corresponding wind speed mode.

[0037] The wind speed data is discretely wavelet-decomposed by using a data acquisition device, and the decomposition coefficient values and time indexes of high-frequency sub-signals and low-frequency sub-signals obtained by the wavelet decomposition are obtained; statistical parameters and energy distribution characteristics are calculated according to the decomposition coefficient values, the statistical parameters are processed by using a maximum-minimum value normalization method to obtain a normalized wind speed feature vector; a kernel similarity is calculated according to the wind speed feature vector and support vectors in a support vector machine classifier, and a wind speed mode category to which the wind speed data belongs is judged by using the kernel similarity and a classification weight; a corresponding prediction model is selected according to the wind speed mode category, and the wind speed feature vector is calculated by using the prediction model to obtain a predicted value of the concrete surface moisture evaporation rate.

[0038] Specifically, the newly added wind speed data is acquired by using a data acquisition device, the wind speed data is four-layer discretely wavelet-decomposed based on a preset sampling period, a db4 wavelet base function is used to extract high-frequency sub-signals and low-frequency sub-signals of the wind speed data, and the decomposition coefficient values of each layer and the corresponding time indexes are recorded. The statistical parameters of the wavelet decomposition coefficients of each layer are calculated, including the mean, the standard deviation, the kurtosis and the skewness, the statistical parameters are normalized by using the maximum-minimum value normalization method to generate the normalized wind speed feature parameters. The cumulative energy distribution of the wind speed data is calculated based on a sliding time window, the energy proportion of each frequency band signal is extracted, and the feature vector is constructed in combination with the normalized wind speed feature parameters. The pre-trained support vector machine classifier is called, the kernel similarity of the feature vector and the support vectors of each category is calculated by using a radial basis kernel function, and the wind speed mode category to which the newly added wind speed data belongs is determined according to the kernel similarity and the classification weight. According to the judged wind speed mode category, the corresponding prediction model is selected, the wind speed feature vector is input into the prediction model, and the moisture evaporation rate prediction value under the normalized scale is obtained. The prediction value under the normalized scale is restored to the original numerical value interval by using the inverse normalization transformation to obtain the moisture evaporation rate prediction value of the concrete surface under the actual physical unit. The confidence interval of the prediction value is calculated, the reliability of the prediction result is judged based on a preset confidence threshold, and the prediction time, the prediction value and the confidence interval are recorded. In the wind speed data processing, the wind speed sequence with a sampling period of 1 second is analyzed by using four-layer discrete wavelet decomposition. Taking the measured wind speed data [5.2, 5.5, 7.8, 7.6, 5.3, 5.4] m / s as an example, the db4 wavelet base function is used for decomposition, the first layer decomposition obtains the low-frequency coefficient

[0039] [5.35, 7.7, 5.35] and the high-frequency coefficient [0.15, -0.1, -0.05], and the second layer decomposition obtains the trend characteristics of lower frequency. In the statistical parameter calculation, the feature values of the wavelet coefficient sequence are calculated. Taking the first layer high-frequency coefficient as an example, the mean μ = 0.0, the standard deviation σ = 0.13, the kurtosis k = 2.8, and the skewness s = 0.15. These statistical values are mapped to the [0, 1] interval after maximum-minimum value normalization to form the normalized feature parameters

[0040] [0.5, 0.65, 0.72, 0.58]. The energy distribution of the wind speed signal reflects the contribution of different frequency components. Using a 20-second sliding time window to calculate the cumulative energy, the low-frequency component accounts for 72% of the total energy, the medium-frequency component accounts for 21%, and the high-frequency component accounts for 7%. Combined with the normalized feature parameters, a 9-dimensional feature vector is formed

[0041] [0.5, 0.65, 0.72, 0.58, 0.82, 0.65, 0.72, 0.21, 0.07]. The support vector machine classification uses the radial basis kernel function K(x, y) = exp(-γ||x-y||2), where γ = 0.1, ||x-y||2 is the squared Euclidean distance between x and y. Calculate the kernel similarity between the feature vector and the support vector of each category, get the similarity value [0.82, 0.35, 0.28], the maximum similarity 0.82 corresponds to the stable wind speed mode, determine that this section of data belongs to the stable wind speed type. The prediction model input is the normalized feature vector, and the output is the predicted value in the normalized scale 0.45. Through the inverse normalization transformation y = y'(ymax-ymin) + ymin, where y' is the normalized predicted value, ymax = 5.0, ymin = 1.0, the actual evaporation rate prediction value is calculated 2.8 grams per square meter per second. The confidence interval calculation is based on the statistical distribution of the prediction error. At the 95% confidence level, the interval estimate of the predicted value is [2.5, 3.1] grams per square meter per second. The interval width is less than the preset threshold 1.0 grams per square meter per second, indicating that the prediction result has good reliability. The complete prediction record contains the prediction time 2024-01-15 10:30:00, the prediction value 2.8, the upper confidence limit 3.1, and the lower confidence limit 2.5, all in grams per square meter per second. Through this multi-level data analysis and processing, the full-process conversion from raw wind speed data to water evaporation rate prediction is realized, and the prediction result contains both point estimate and interval estimate, providing more comprehensive prediction information.

[0042] Step S107, a concrete slump prediction model is constructed in advance, the predicted value of the concrete surface water evaporation rate is taken as the input, the input water evaporation rate data is analyzed and predicted, and the predicted value of the concrete slump is obtained. If the predicted concrete slump value is outside or below the preset qualified range, it is determined that the concrete slump is unqualified, and a warning signal is issued.

[0043] The cumulative moving average method is used to smooth the evaporation rate prediction value sequence, and the smoothed evaporation rate change rate data is obtained; the exponential weighted fusion is carried out according to the evaporation rate change rate data, and the time sequence feature vector reflecting the evaporation rate evolution characteristic is constructed; the time sequence feature vector is input into the pre-trained long short-term memory network, and the slump prediction value is obtained through the memory unit output state of the network; if the confidence interval width of the slump prediction value exceeds the preset threshold, the local linear regression method is used to correct the prediction value, and the corresponding level warning mark is obtained by comparing the corrected prediction value with the preset qualified range.

[0044] Specifically, the cumulative moving average method is used to smooth the sequence of water evaporation rate prediction values, the smoothed evaporation rate time series data is calculated based on the preset sampling period, and the first and second order change rates of the evaporation rate are extracted using the difference algorithm. Calculate the cumulative change of the evaporation rate, fuse the change rate data based on the exponential weighting method, and construct a time series feature vector reflecting the evolution characteristics of the evaporation rate. The time series feature vector is input into the pre-trained long short-term memory network, which includes an input layer, a hidden layer, and an output layer, where the hidden layer consists of memory cells and forget gates, recording the time series variation of the concrete slump. According to the output state of the memory cell, the slump prediction value is calculated, and the preset confidence interval method is used to evaluate the reliability of the prediction result. When the prediction interval width exceeds the threshold, it is marked as a low confidence prediction result. For low confidence prediction results, the local linear regression method is used to correct the prediction value, and the corrected slump prediction value is calculated based on the historical data in the neighborhood of the prediction point. Compare the corrected slump prediction value with the upper and lower limits of the preset qualified range, and generate corresponding level warning marks when the prediction value deviates from the qualified range. According to the severity of the warning marks, determine the warning level, mild deviation corresponds to yellow warning, serious deviation corresponds to red warning, and transmit the warning information to the monitoring terminal through the data communication interface. In the water evaporation rate data processing, the original sequence [2.8, 3.2, 3.5, 3.3, 2.9, 2.7] g / m2·s is smoothed by 5-point moving average to obtain the smoothed sequence [3.14, 3.12]. Calculate the first order difference [-0.02] to reflect the trend of change, and the second order difference to reflect the acceleration of change. Use the exponential weight α = 0.7 to weight the change rate, and construct a three-dimensional vector [3.12, -0.02, 0] reflecting the dynamic characteristics of evaporation. The core of the long short-term memory network lies in the state control of the memory cells. Taking the predicted slump as an example, the network input layer receives the feature vector, and the hidden layer contains 32 memory cells. Each memory cell controls the flow of information through the input gate, the forget gate, and the output gate. The input gate determines the proportion of new information received, the forget gate controls the retention degree of historical information, and the output gate adjusts the output amount of information. When new evaporation rate data is input, the memory cell will weigh the importance of new and old information. In slump prediction, the confidence interval uses 3 times the mean square error as the interval radius. For a prediction value of 180 mm, the prediction standard deviation is 5 mm, and the confidence interval is [165, 195] mm. When the interval width exceeds the preset threshold of 40 mm, it indicates that the prediction result has low credibility and needs to be corrected. Correction uses local linear regression, selecting 5 historical data points before and after the prediction point to fit the local linear equation y = kx + b, where k is the slope and b is the intercept. The fitted value replaces the original prediction value. The warning mechanism is based on the qualified range of concrete slump.For the preset qualified range [160, 200] mm, a yellow warning is triggered when the predicted value is in the interval [140, 160) or (200, 220], and a red warning is triggered when the predicted value exceeds the interval [140, 220]. In an actual case analysis, the corrected slump prediction value is 145 mm, which falls into the yellow warning interval, and the system generates a warning information containing the timestamp 2024-01-15 14:30:00, the predicted value 145, and the warning level yellow. The warning information is transmitted to the monitoring terminal in JSON format through the data communication interface, containing the necessary warning elements: {"timestamp": "2024-01-15 14:30:00", "slump_predict": 145, "confidence_interval": [140, 150], "alert_level": "yellow"}. After receiving the warning information, the monitoring terminal displays the corresponding visual prompt according to the warning level, with yellow flashing prompt for yellow warning and red flashing prompt with sound alarm for red warning. Through this multi-level data analysis and warning mechanism, the prediction from evaporation rate to slump is realized, and a complete warning response process is established, providing timely warning information for concrete quality control. The hierarchical setting and display method of the warning mechanism intuitively reflect the severity of quality risk.

[0045] Step S108, after triggering the warning signal, according to the value of the concrete slump prediction value exceeding or being lower than the preset qualified range, determine the adjustment scheme of the concrete proportion, input the concrete proportion parameters of the adjustment scheme into the concrete production equipment, and continuously obtain the surface moisture evaporation rate data of the adjusted concrete, return to the concrete slump prediction model input, and perform online prediction of the next concrete slump.

[0046] The deviation value of the concrete slump prediction value from the preset qualified range is fuzzy processed by using a triangular membership function, and the water-binder ratio adjustment amount is obtained according to the deviation value and a preset fuzzy rule; according to the water-binder ratio adjustment amount, the adjustment values of the cement dosage, aggregate dosage and admixture dosage are obtained by a regression equation of proportioning parameters constructed by a weighted least squares method; the adjustment values are converted into equipment control instructions, packaged through an industrial Ethernet protocol and transmitted to the concrete production equipment; the surface moisture data of the adjusted concrete production equipment is obtained by using a data acquisition device, and a sequence of moisture evaporation rate data is calculated according to a preset sampling period, and the sequence of moisture evaporation rate data is returned to the slump predictor after digital filtering processing.

[0047] Specifically, according to the deviation of the concrete slump prediction value from the preset qualified range, the deviation value is processed by a triangular membership function, the water-binder ratio adjustment amount is calculated based on the preset fuzzy rule, and the fuzzy rule includes the corresponding relationship between the deviation degree and the adjustment amplitude. The weighted least squares method is used to construct the concrete proportioning parameter regression equation, the water-binder ratio adjustment amount is used as the independent variable, the adjustment values of the cement content, the coarse and fine aggregate content, and the admixture content are calculated, and the proportioning parameter combination meeting the workability constraint is generated. The proportioning parameter combination is converted into equipment control instructions, the instructions are packaged using the industrial Ethernet protocol, and the adjustment instructions are transmitted to the concrete production equipment through the serial communication interface. The execution state information returned by the production equipment is received, the execution result of the proportioning adjustment instruction is judged according to the state code, and the actual write-in value of the adjustment parameter is recorded. The surface moisture data of the adjusted concrete is obtained by using the data acquisition device, the moisture evaporation rate per unit time is calculated according to the preset sampling period. The moisture evaporation rate data is processed by digital filtering to eliminate the influence of random fluctuations, and a stable evaporation rate data sequence is obtained. The stable evaporation rate data sequence is returned to the slump predictor, and the prediction data is updated in a sliding time window manner to realize the dynamic update of the prediction result. The deviation between the prediction value and the measured value is recorded, and when the prediction deviation exceeds the threshold value for three consecutive times, the online correction of the predictor parameters is triggered, and the predictor internal parameters are updated. When the concrete slump prediction value deviates from the preset qualified range, the triangular membership function is used for fuzzy processing. Taking the measured deviation value of 25mm as an example, the membership function interval [-40, 0] is set, the vertex value is -20, and the deviation membership value 0.75 is calculated. According to the fuzzy rule "if the deviation is negative, the water-binder ratio adjustment amount is positive", the water-binder ratio adjustment amount 0.02 is mapped. When the weighted least squares method is used to construct the proportioning parameter regression equation, the water-binder ratio x is used as the independent variable, the proportioning parameter y is used as the dependent variable, the weight w reflects the importance of the parameter, xi is the water-binder ratio of the i th observation point, and yi is the proportioning parameter of the i th observation point. For the equation y = ax + b, the weighted error sum of squares Σw(yi-axi-b) 2Solve the coefficients a, b. In the example, the cement dosage regression equation is y =-150x + 550, and the water-cement ratio adjustment amount is 0.02. The calculation result is that the cement dosage is reduced by 3 kg per cubic meter. The proportioning parameter adjustment instruction is transmitted using the industrial Ethernet Modbus TCP protocol. The instruction frame contains the function code 0x10 (write multiple registers), the starting address 0x1000, the register number 0x0004, the data length 0x08, and the proportioning parameter value. The device returns a response frame after execution. The status code 0x00 indicates successful execution, and 0x04 indicates an execution exception. The data acquisition device collects surface moisture data every 30 seconds. Five sets of data [12.5%, 12.2%, 11.8%, 11.5%, 11.2%] are continuously collected. The moisture evaporation rate is calculated to be 2.6 g / m2·min by time difference. The 5-point median filter is used to eliminate abnormal fluctuations. The stable sequence after filtering is more suitable for prediction calculation. In the predictor parameter correction, the deviation sequence of the predicted value and the measured value is recorded for three times [-8, -10, -12] mm. The root mean square error 10.2 mm exceeds the preset threshold 8 mm. After triggering the correction, the weight parameters of the predictor are updated to make the prediction result closer to the actual trend. The closed-loop control process shows the complete chain from slump prediction, proportioning adjustment to effect verification. Taking an adjustment as an example, the predicted slump value 145 mm is lower than the qualified lower limit 160 mm. The proportioning adjustment amount is calculated through fuzzy control. After the adjustment is executed, the moisture change is continuously monitored. The new predicted value 162 mm has entered the qualified range, verifying the effectiveness of the adjustment. The whole process realizes the automatic control of prediction-adjustment-verification, ensuring that the concrete performance continuously meets the requirements. In practical application, the control scheme realizes real-time monitoring and automatic adjustment of concrete performance through accurate mathematical models and reliable data transmission, greatly improving the stability of the production process and the consistency of product quality. Through the closed-loop feedback mechanism, deviations can be found and corrected in time, avoiding losses caused by quality fluctuations.

[0048] It should be noted that the above only lists several specific embodiments of the present application. Obviously, the present application is not limited to the above embodiments, but can have many variations. All variations that can be directly derived or inferred from the disclosure of the present application by those of ordinary skill in the art should be considered within the scope of the present application.

Claims

1. A method for online monitoring of slump of casted concrete, characterized in that, The method comprises: a plurality of concrete surface moisture sensors and wind speed sensors are arranged at a concrete pouring site, a preset sensor sampling frequency is set, a concrete surface and an environment within a preset range are measured, and concrete surface moisture evaporation rate data and wind speed data are obtained; the collected concrete surface moisture evaporation rate data and wind speed data are uploaded to a data processing server in real time and stored in a time sequence format to form moisture evaporation time sequence data and wind speed time sequence data, the wind speed time sequence data is processed into wind speed data of different scales, and the wind speed data is divided into high-frequency components and low-frequency components according to different scales; the high-frequency components and the low-frequency components of the wind speed data are compared with preset gust identification thresholds and turbulence identification thresholds respectively, whether gusts and turbulence occur in the wind speed data is judged, if gusts and turbulence occur, the start time, the end time and the duration of the occurrence of the gusts and the turbulence are recorded, the time distribution of gust wind speed modes and turbulence wind speed modes is obtained according to the recorded start time, the end time and the duration of the occurrence of the gusts and the turbulence, and wind speed mode subsequences are formed; the moisture evaporation time sequence data is divided according to the time distribution of different wind speed modes to obtain evaporation rate subsequences corresponding to different wind speed modes, the influence of different wind speed modes on the moisture evaporation rate is quantified by comparing the mean values and standard deviations of the evaporation rate subsequences under different wind speed modes, and statistical characteristic quantities of the evaporation rate under different wind speed modes are obtained; based on the statistical characteristic quantities of different wind speed modes, a prediction model under different wind speed modes is constructed by taking the statistical characteristic quantities of the wind speed modes as inputs and the statistical characteristic quantities of the evaporation rate under the corresponding wind speed modes as outputs; after new wind speed data is collected, the wind speed mode corresponding to the new wind speed data is judged, and the prediction value of the concrete surface moisture evaporation rate is obtained according to the prediction model under the corresponding wind speed mode; a concrete slump prediction model is constructed in advance, the prediction value of the concrete surface moisture evaporation rate is taken as an input, the input moisture evaporation rate data is analyzed and predicted, the prediction value of the concrete slump is obtained, and if the predicted concrete slump value is out of or below a preset qualified range, the concrete slump is determined to be unqualified, and a warning signal is sent out; after the warning signal is triggered, the adjustment scheme of the concrete proportioning is determined according to the value by which the prediction value of the concrete slump exceeds or is below the preset qualified range, the concrete proportioning parameters of the adjustment scheme are input into the concrete production equipment, the surface moisture evaporation rate data of the adjusted concrete is continuously obtained, and the input is returned to the concrete slump prediction model for online prediction of the concrete slump in the next cycle.

2. The method of claim 1, wherein, The method comprises: a plurality of concrete surface moisture sensors and wind speed sensors are arranged at a concrete pouring site, a preset sensor sampling frequency is set, a concrete surface and an environment within a preset range are measured, and concrete surface moisture evaporation rate data and wind speed data are obtained; a sensor sampling region grid division map is used to divide the concrete surface into measurement sub-regions, and a data acquisition module is used to obtain time sequence monitoring values of the sensor array. The time series monitoring values are subjected to spatial interpolation operation by a Kriging interpolation algorithm to obtain a water distribution map and a wind speed distribution map of the concrete surface; The water evaporation rate data of the measurement point is obtained by dividing the difference between the water content at adjacent sampling times in the water distribution map by the sampling time interval; If the wind speed sensor detection value is greater than the preset wind speed threshold value, the water evaporation rate data is subjected to denoising processing by Gaussian filtering to obtain corrected water evaporation rate data.

3. The method of claim 1, wherein, The collected concrete surface water evaporation rate data and wind speed data are uploaded to a data processing server in real time and stored in time series format to form water evaporation time series data and wind speed time series data, and the wind speed time series data is processed into wind speed data of different scales, which is divided into high-frequency components and low-frequency components according to different scales, including: The concrete surface monitoring device is used to obtain water evaporation rate data and wind speed data, which are compressed and transmitted to the data processing server after being marked with time stamps according to a preset sampling period; The compressed data is received and decompressed using a time series database, and the missing data points are interpolated by a cubic spline interpolation algorithm according to the time stamp to obtain complete time series data; The complete time series data is preprocessed using a median filter to remove measurement noise and outliers in the time series data to obtain preprocessed wind speed data; The preprocessed wind speed data is subjected to discrete wavelet transform decomposition using a db4 wavelet basis function, the instantaneous frequency characteristics of the wind speed data are calculated by Hilbert transform, and the wind speed data is divided into high-frequency wind speed components and low-frequency wind speed components using a Butterworth bandpass filter bank.

4. The method of claim 1, wherein, The high-frequency components and low-frequency components of the wind speed data are compared with the preset gust recognition threshold and turbulence recognition threshold, respectively, to determine whether gusts and turbulence occur in the wind speed data, and if so, the start and end times and duration of the gusts and turbulence are recorded, and the time distribution of the gust wind speed pattern and the turbulence wind speed pattern is obtained according to the recorded start and end times and duration of the wind speed, forming a wind speed pattern subsequence, including: The frequency separator is used to separate the high-frequency components and low-frequency components of the wind speed data, and the gust event markers are obtained by comparing the high-frequency components with the threshold, and the turbulence event markers are obtained by comparing the low-frequency components with the threshold; The time window processor is used to segment the wind speed data, and the event count values in each time window are calculated according to the gust event markers and the turbulence event markers; The feature extraction is performed on the wind speed data in the time window, and the gust feature sequence and the turbulence feature sequence are calculated according to the event count values; The dynamic programming algorithm is used to time-align the gust feature sequence and the turbulence feature sequence, and the wind speed pattern feature vector is generated according to the aligned feature sequences.

5. The method of claim 1, wherein, The water evaporation time series data is divided according to the time distribution of different wind speed modes to obtain evaporation rate subsequences corresponding to different wind speed modes. By comparing the mean and standard deviation of the evaporation rate subsequences under different wind speed modes, the influence of different wind speed modes on the water evaporation rate is quantified to obtain statistical characteristic quantities of the evaporation rate under different wind speed modes, including: According to the time stamp alignment method, the water evaporation time series data is segmented, and the boundary data is supplemented at the segmentation points using the cubic spline interpolation method to obtain water evaporation rate subsequences that satisfy the first derivative continuity constraint; The trapezoidal integral method is used to integrate the water evaporation rate subsequences, and the arithmetic mean of the water evaporation rate subsequences is calculated based on the integral operation results to obtain the instantaneous evaporation rate in the corresponding time period; According to the box plot method, the water evaporation rate subsequences are identified for abnormal points, and the linear interpolation method is used to correct the abnormal points to obtain the corrected water evaporation rate subsequences; The central moment of the corrected water evaporation rate subsequences is calculated using the moment estimation method, the central moment is standardized using the maximum and minimum value normalization method, and the standardized feature vectors are grouped using the hierarchical clustering method to obtain the water evaporation rate statistical feature combination under different wind speed modes.

6. The method of claim 1, wherein, The statistical characteristic quantities based on different wind speed modes are used as input, and the evaporation rate statistical characteristic quantities corresponding to the wind speed mode are used as output to construct a prediction model under different wind speed modes, including: The feature quantity dataset is grouped according to the wind speed mode category label, and the maximum and minimum value normalization method is used to process the feature quantity dataset to obtain the standardized wind speed mode feature quantity and evaporation rate feature quantity; The correlation between the standardized feature quantities is calculated using the Pearson correlation coefficient, and if the absolute value of the correlation coefficient is greater than a preset threshold, the reduced wind speed mode feature quantity and evaporation rate feature quantity are obtained; The reduced feature quantity dataset is stratified sampled, and the training set and validation set are obtained by dividing the training set ratio, and the training validation data combination is obtained by resampling the dataset using the cross-validation method; Based on the training validation data combination, a random forest prediction model is constructed, and the optimal parameter combination is searched in the preset parameter grid space. The root mean square error and the determination coefficient of the validation set are calculated using the optimal parameter combination.

7. The method of claim 1, wherein, After adding the newly collected wind speed data, the wind speed mode corresponding to the newly added wind speed data is determined, and the prediction value of the concrete surface water evaporation rate is obtained based on the prediction model under the corresponding wind speed mode, including: The wind speed data is discretely wavelet decomposed using a data acquisition device, and the decomposition coefficient values and time indexes of the high-frequency sub-signals and low-frequency sub-signals obtained by wavelet decomposition are calculated; Statistical parameters and energy distribution characteristics are calculated based on the decomposition coefficient values, and the wind speed feature vectors are standardized using the maximum and minimum value normalization method. According to the wind speed feature vector and the support vector in the support vector machine classifier, a kernel similarity is calculated, and the wind speed data belongs to a wind speed mode category is determined according to the kernel similarity and a classification weight; According to the wind speed mode category, a corresponding prediction model is selected, and the prediction model is used to calculate the wind speed feature vector to obtain a predicted value of the concrete surface water evaporation rate.

8. The method of claim 1, wherein, The pre-constructed concrete slump prediction model takes the predicted value of the concrete surface water evaporation rate as input, analyzes and predicts the input water evaporation rate data, and obtains a predicted value of the concrete slump. If the predicted concrete slump value is outside or below the preset qualified range, it is determined that the concrete slump is unqualified, and a warning signal is sent, including: The cumulative moving average method is used to smooth the water evaporation rate prediction value sequence to obtain smoothed evaporation rate change rate data; According to the evaporation rate change rate data, an exponential weighted fusion is performed to construct a time series feature vector reflecting the evaporation rate evolution characteristics; The time series feature vector is input into a pre-trained long short-term memory network, and a slump prediction value is obtained by outputting the state of the memory unit of the network; If the confidence interval width of the slump prediction value exceeds the preset threshold, the local linear regression method is used to correct the prediction value, and the corrected prediction value is compared with the preset qualified range to obtain a corresponding level of warning mark.

9. The method of claim 1, wherein, After the warning signal is triggered, according to the concrete slump prediction value exceeding or being below the preset qualified range, an adjustment scheme of the concrete proportioning is determined, the concrete proportioning parameters of the adjustment scheme are input into the concrete production equipment, and the surface water evaporation rate data of the adjusted concrete is continuously obtained, returned to the concrete slump prediction model input, and the online prediction of the concrete slump is performed next time, including: The triangular membership function is used to fuzz the deviation value of the concrete slump prediction value and the preset qualified range, and the water-binder ratio adjustment amount is obtained according to the deviation value and the preset fuzzy rule; According to the water-binder ratio adjustment amount, the adjustment values of the cement dosage, the aggregate dosage, and the admixture dosage are obtained through the proportioning parameter regression equation constructed by the weighted least squares method; The adjustment values are converted into equipment control instructions, which are packaged and transmitted to the concrete production equipment through the industrial Ethernet protocol; The data acquisition device is used to obtain the adjusted surface water data of the concrete production equipment, and the water evaporation rate data sequence is calculated according to the preset sampling period. The water evaporation rate data sequence is returned to the slump predictor after digital filtering.

Images