A method for managing the pouring of ready-mixed concrete at construction sites

Abstract

This invention relates to the field of construction management technology and discloses a method for managing the pouring of ready-mixed concrete on construction sites, including the following steps: Step 1: Data Acquisition Stage: S1.1 By scanning the QR code on the concrete raw material packaging, the type, batch number, and production date information of cement, aggregate, fly ash, and admixtures are automatically entered. This invention constructs a full-dimensional data acquisition network by deploying temperature and humidity sensors, slump sensors, pressure sensors, and vibration sensors. Combined with LoRa wireless transmission technology, data integrity is achieved: temperature data is collected every minute (initial setting stage), and slump data is collected every 5 minutes (final setting stage), forming a continuous temperature curve; rapid response to anomalies: abnormal data is reconstructed through cubic spline interpolation to ensure monitoring continuity; decision support: a 128×128 pixel infrared thermal image and a T×N sensor matrix are provided for the CNN model, improving the accuracy of quality prediction.

Description

Technical Field

[0001] This invention relates to the field of construction management technology, specifically to a method for monitoring and managing the quality of ready-mixed concrete pouring at construction sites based on intelligent sensing technology and artificial intelligence algorithms. This invention integrates multi-source sensor data with deep learning algorithms to achieve real-time monitoring and quality early warning of the entire concrete pouring process, and is applicable to ready-mixed concrete construction scenarios in various construction projects. Background Technology

[0002] Ready-mixed concrete, as a core material in modern construction engineering, directly impacts structural safety and durability through its pouring quality. Currently, construction sites commonly employ a management approach combining manual inspection and basic sensors: Manual inspection is dominant, relying on the experience of quality inspectors to judge indicators such as concrete slump and temperature through visual observation or simple tools (such as slump cones), which suffers from high subjectivity and low efficiency. Basic sensor application: Some sites use temperature and humidity sensors to monitor environmental parameters, but the data collection range is limited and lacks correlation analysis with concrete performance. Delayed response mechanism: Abnormal situations (such as sudden temperature rises or slump loss) require manual reporting, with an average response time exceeding 2 hours, making it difficult to prevent the expansion of quality defects.

[0003] However, existing technologies cannot achieve real-time monitoring and intelligent early warning of the entire concrete pouring process. They are prone to false alarms and missed alarms, data silos and decision-making rely on experience. They cannot link key variables such as concrete mix proportions and pouring speed, resulting in one-sided quality analysis. The frequency of manual inspections is low (usually 2-3 times a day), making it impossible to capture dynamic changes during the pouring process. Quality records rely heavily on paper documents, making it difficult to achieve full-process traceability.

[0004] Therefore, the present invention provides a method for managing the pouring of ready-mixed concrete at construction sites. Summary of the Invention

[0005] The purpose of this invention is to provide a method for managing the pouring of ready-mixed concrete at construction sites, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for managing the pouring of ready-mixed concrete at a construction site, comprising the following steps:

[0007] Step 1: Data Acquisition Phase

[0008] S1.1 By scanning the QR code on the concrete raw material packaging, the type, batch number, and production date information of cement, aggregate, fly ash, and admixtures are automatically entered;

[0009] S1.2 Deploy temperature and humidity sensors, slump sensors, pressure sensors and vibration sensors to collect concrete hardening temperature curves, drying shrinkage rate, fluidity index, peak mixing pressure and pouring vibration frequency data in real time.

[0010] S1.3 uploads the data collected by the multi-source sensors to the central processing unit via the LoRa wireless transmission module for data fusion processing, generates a real-time monitoring report, and then intelligently analyzes the concrete pouring quality based on the real-time monitoring report.

[0011] Step Two: Quality Inspection Stage

[0012] S2.1 Construct a convolutional neural network model. The input layer receives a 128×128 pixel infrared thermal image of the concrete surface and a sensor time series data matrix. The sensor time series data matrix contains time series data of temperature, humidity, pressure and vibration sensors. The dimension is T×N, where T is the time step and N is the number of sensors.

[0013] S2.2 sets up four convolutional layers with a kernel size of 3×3 and a number of kernels of 64, 128, 256 and 512 respectively. Each layer is followed by a batch normalization layer and a ReLU activation function.

[0014] S2.3 Insert two max pooling layers with a pooling window of 2×2 and a stride of 2, and a global average pooling layer to reduce the spatial dimension of the feature map.

[0015] The S2.4 output layer uses the Softmax function to classify concrete quality grades and generates a quality report that includes slump deviation rate, temperature gradient and vibration uniformity coefficient.

[0016] Step 3: Early Warning and Control Phase

[0017] S3.1 When the system detects any of the following abnormal conditions in real time through temperature and humidity sensors and slump sensors deployed at the construction site, it will automatically activate the graded early warning procedure:

[0018] When the internal temperature of concrete is between 30°C and T, or slump loss rate is between 10% and SLR, a Level 1 Yellow Alert is issued, and structured alert information is pushed to the construction worker's handheld terminal via a LoRa wireless communication module.

[0019] When the internal temperature of the concrete is between 35°C and T, and the slump loss rate is between 15% and SLR, and the slump loss rate is between 20% and 15%, a Level II orange alert is issued. The speed of the concrete delivery pump is adjusted in conjunction with the system, and the adjustment amount is calculated using the following formula:

[0020] ;

[0021] When the internal temperature of the concrete T ≥ 40°C or the slump loss rate SLR ≥ 20%, a Level III red alert is issued. The power supply to the delivery pump is cut off via a relay module, and the pouring valve is simultaneously closed (pneumatic actuator response time ≤ 0.5s). The cooling water pipe system embedded in the concrete structure is then activated, and the cooling water flow rate is controlled according to the following formula:

[0022] This ensures rapid cooling in emergency situations, preventing the concrete structure from losing strength due to high temperatures, and effectively guaranteeing construction quality and safety.

[0023] In one embodiment of the present invention, preferably, the process of automatically entering information by scanning the QR code on the concrete raw material packaging includes: using a handheld terminal device to read the QR code data, transmitting it to a central database in real time, and then the system automatically comparing the entered information with the construction plan requirements to ensure material compliance. The system also performs real-time analysis of the entered data and generates a material usage report. If the entered information does not match the construction plan requirements, the generated material usage report will have a "mismatch" mark and provide adjustment suggestions to ensure that construction personnel verify the information and suspend the use of the batch of materials. Conversely, if the information matches the construction plan requirements, a "match" mark will be generated, allowing the use of the batch of materials.

[0024] In one embodiment of the present invention, preferably, the process of acquiring the concrete hardening temperature curve in real time by deploying temperature and humidity sensors includes:

[0025] Before concrete pouring, temperature and humidity sensors are buried at a predetermined depth in a quincunx pattern with a spacing of ≤1.5m to ensure coverage of different depths and locations. For large-scale projects, a multi-point arrangement is adopted, with one main sensor and one backup sensor set at each monitoring point to avoid single-point failure.

[0026] During installation, ensure the temperature and humidity sensor is in close contact with the concrete to avoid air bubbles or gaps, and use thermally conductive adhesive to fix it to reduce thermal resistance; surface sensors are fixed to the structural surface by magnetic attraction or adhesive to ensure long-term stability.

[0027] The temperature and humidity sensor collects data at a preset frequency, packages and transmits it to the gateway via the LoRa protocol, and the gateway verifies the data integrity through CRC check. If the verification fails, a retransmission is triggered. The collection frequency is divided into:

[0028] Initial condensation stage: Samples are collected once per minute to capture temperature peaks;

[0029] Final condensation stage: Samples were collected every 5 minutes to monitor the cooling rate;

[0030] Late hardening stage: Samples were collected every 30 minutes to assess long-term stability;

[0031] During data preprocessing, short-term fluctuations and random noise are eliminated by sliding a fixed-size window across the data sequence and calculating the average value of the data within the window. Then, specific upper and lower thresholds are set to determine whether data points are abnormal, such as temperatures >60℃ or <-5℃, and abnormal data are marked.

[0032] For marked anomalous data, cubic spline interpolation is used for reconstruction first. If the anomalous data continues for more than 3 cycles, the backup sensor is activated and the event is recorded to the blockchain.

[0033] Generate temperature curve reports, including fitting formulas and key parameters such as maximum temperature and heating rate. The data is stored on the blockchain to ensure traceability and support quality traceability and auditing.

[0034] In one embodiment of the present invention, preferably, the temperature data point during the initial condensation stage is set as follows: The fitted polynomial is:

[0035] ,in, It is the order of the polynomial. These are the coefficients to be solved. To find the best-fit polynomial, we minimize the sum of squared errors. Solve for the coefficients;

[0036] make The Vandermonde matrix has the following elements. , A column vector of temperature data. Let be the column vector of polynomial coefficients, then we have Solving for the polynomial coefficients yields... ;

[0037] The final setting stage is converted into a quadratic polynomial (k=2) to simplify computational resource consumption. To ensure the temperature rise process of concrete under adiabatic conditions, its mathematical expression is: ,in, Indicates time The concrete temperature at that time; This represents the final temperature rise of the concrete, that is, the difference between the concrete temperature and the initial temperature as time approaches infinity. It is the heating rate constant, which determines the rate at which the concrete temperature rises. The larger the value, the faster the concrete temperature rises; The ambient temperature means that the influence of the external environment on the concrete temperature during the insulation process can be considered constant.

[0038] To solve for the parameters in the model , and The specific process includes the following:

[0039] Collect temperature data of concrete at different time points and corresponding time ;

[0040] Take the natural logarithm of both sides of the mathematical expression of the model. This allows the original model to be converted into a linear form;

[0041] The transformed linear equation was fitted using the least squares method, and the collected temperature data was used as the basis for the calculation. and time data Substitute into the equation, and adjust , and The value of minimizes the sum of squared errors between the fitted result and the actual data. After linear fitting, we can obtain... , and The estimated value.

[0042] In one embodiment of the present invention, preferably, the process of calculating the slump value in real time by deploying a slump sensor includes:

[0043] Place the slump sensor near the concrete mixer truck outlet or pouring point to ensure that the sensor container is in full contact with the concrete. During installation, the sensor should be placed horizontally to avoid external vibration interference and there should be no obstructions around it to ensure that the concrete can be poured in smoothly.

[0044] Real-time collection of sensor output signals and current temperature Calculate slump using the basic model Its mathematical expression is: Set a reasonable slump range, such as 80mm ≤ S ≤ 180mm. Data points outside this range will be marked as abnormal. For transient abnormalities (such as sensor noise), use median filtering to replace the abnormal values. If there are 3 consecutive abnormal points, notify the construction personnel to check the concrete mix ratio or sensor status.

[0045] By calculating the rate of change of slump If the rate of change suddenly exceeds 5 mm / s, an abnormal alarm will be triggered.

[0046] Calculate the mean (μ) and standard deviation (σ) of the data. If a data point exceeds the mean ± 3 times the standard deviation (μ ± 3σ), it is considered an outlier.

[0047] In one embodiment of the present invention, preferably, the process of calculating the drying shrinkage rate by real-time acquisition of concrete drying shrinkage rate includes:

[0048] Set the drying shrinkage rate of concrete to be... The length of the concrete specimen in its initial state is The length of the concrete specimen after drying is The reference length selected when measuring the change in concrete length is Through formula This allows for the quantification of the degree of shrinkage of concrete during the drying process.

[0049] In one embodiment of the present invention, preferably, the method for calculating the concrete fluidity index includes:

[0050] When concrete cannot flow to the far right of the box, its fluidity is restricted, and its fluidity index decreases. Calculated using the following formula:

[0051] ,in, This refers to the actual flow length, which is the farthest distance the concrete can flow to within a specified time. This is the total length of the box;

[0052] When the concrete can flow to the far right of the box, its fluidity index is... Calculated using the following formula:

[0053] ,in, This refers to the height of the bottom of the box or a specific location after the concrete flow has stopped. This refers to the height at the same position when the concrete is in its initial state.

[0054] In one embodiment of the present invention, preferably, the process of acquiring real-time data on peak stirring pressure and pouring vibration frequency by deploying pressure sensors and vibration sensors includes:

[0055] During the concrete mixing process, peak mixing pressure detection is used to first collect pressure data over a period of time and then calculate the peak pressure during that period. and average pressure Then calculate the peak-to-average ratio. ,when >2.5 and When the pressure exceeds 500 kPa, a Level 3 alarm is triggered. Simultaneously, a CNN model is used to analyze the current data, and adjustments are made dynamically based on the model's output. The threshold;

[0056] When concrete is poured into the formwork, vibration and impact signals are generated due to the concrete's fluidity and gravity. Vibration sensors use internal piezoelectric crystals to monitor these vibration signals during the concrete pouring process and convert them into electrical signals for output. During this process, a fast Fourier transform is used for spectral analysis, converting the time-domain signal into the frequency domain, which involves the vibration uniformity coefficient. It can be calculated using the following formula:

[0057] ,in, For frequency, The vibration frequency at the measuring point, The average vibration frequency, To determine the number of measurement points, compare each measurement point. Vibration frequency at the measuring point With average vibration frequency The degree of deviation can quantify the uniformity of vibration.

[0058] In one embodiment of the present invention, preferably, in process S2.1, the two-dimensional feature map of the infrared image and the one-dimensional time-series vector of the sensor data are fused through channel concatenation to form a multimodal input tensor, which is expressed by the following formula:

[0059] ,in, For the fused multimodal input tensor, For the feature map tensor of the infrared image, For sensor time-series data;

[0060] The formula for the convolution operation in process S2.2 is expressed as follows:

[0061] ,in, For the first The output of the layer at position (i,j) For the first The convolutional kernel weights of the layer at position (m,n) For the first The input at position (i+m, j+n) of the previous layer. For the first Layer bias terms, Batch normalization is used to accelerate training and improve model stability. A nonlinear transformation is introduced for the activation function;

[0062] In process S2.3, for each position (i,j) in the output feature map, its value... It is the maximum value within the corresponding 2×2 window in the input feature map, expressed by the following formula:

[0063] ;

[0064] In process S2.4, the Softmax function is: ,in, The feature vector after global average pooling. , These are the weight vector and bias term for the k-th class, respectively. , These are the weight vector and bias term for class c, respectively, where C is the total number of classes. Here, C=3, representing three classes: qualified, warning, and unqualified. The probability distribution of each quality level is calculated using the Softmax function to finally determine the grade classification of the concrete sample.

[0065] The formula for slump deviation rate is: ,in, This indicates the final value of the parameter or quantity. This indicates the initial value of the parameter or quantity;

[0066] The temperature gradient formula is: ,in, Indicates the final temperature. This indicates the initial temperature.

[0067] This invention provides a method for managing the pouring of ready-mixed concrete at construction sites. It has the following beneficial effects:

[0068] (1) This invention constructs a full-dimensional data acquisition network by deploying temperature and humidity sensors, slump sensors, pressure sensors, and vibration sensors. Combined with LoRa wireless transmission technology, it achieves data integrity: temperature data is collected every minute (initial condensation stage) and slump data is collected every 5 minutes (final condensation stage), forming a continuous temperature curve; rapid response to anomalies: abnormal data is reconstructed through cubic spline interpolation to ensure monitoring continuity; decision support: a 128×128 pixel infrared thermal image and a T×N sensor matrix are provided for the CNN model, improving the accuracy of quality prediction.

[0069] (2) The convolutional neural network model constructed by this invention includes: a four-layer convolutional architecture: 3×3 convolutional kernels extract features layer by layer; a hybrid pooling strategy: two layers of max pooling (2×2 stride) to reduce spatial dimension; global average pooling to integrate global features; and multi-task output: the Softmax classifier outputs quality level and generates a quantitative report simultaneously, thereby realizing the automation and accuracy of quality assessment, improving construction efficiency, and reducing human error.

[0070] (3) This invention achieves risk management in advance by using a graded early warning mechanism to monitor data in real time, dynamically classify early warning levels, and automatically push early warning information to management personnel, thereby effectively preventing potential problems and ensuring construction safety and quality. Attached Figure Description

[0071] Figure 1 This is a flowchart of the method steps of the present invention;

[0072] Figure 2 This is a flowchart of the data acquisition phase of the present invention;

[0073] Figure 3 This is a flowchart of the quality inspection stage of the present invention;

[0074] Figure 4 Configure a table view for the convolutional layers of this invention. Detailed Implementation

[0075] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0076] Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the invention, and should not be construed as limiting the invention.

[0077] A preferred embodiment of the on-site management method for ready-mixed concrete pouring provided by the present invention is as follows: Figures 1-4 As shown: A method for managing the pouring of ready-mixed concrete at a construction site, comprising the following steps:

[0078] Step 1: Data Acquisition Phase

[0079] S1.1 By scanning the QR code on the concrete raw material packaging, the type, batch number, and production date information of cement, aggregate, fly ash, and admixtures are automatically entered;

[0080] The process of automatically entering information by scanning the QR code on the concrete raw material packaging includes: using a handheld terminal device to read the QR code data, transmitting it to a central database in real time, and then the system automatically comparing the entered information with the construction plan requirements to ensure material compliance. The system also performs real-time analysis of the entered data and generates a material usage report. If the entered information does not match the construction plan requirements, the generated material usage report will have a "mismatch" mark and provide adjustment suggestions to ensure that construction personnel verify the information and suspend the use of the batch of materials. Conversely, if the information matches, a "match" mark will be generated, allowing the use of the batch of materials.

[0081] S1.2 Deploy temperature and humidity sensors (measurement range -40℃~85℃, accuracy ±0.3℃), slump sensors (resolution 0.1mm), pressure sensors (range 0~500kPa, accuracy 0.1%FS) and vibration sensors (frequency response range 0~500Hz) to collect concrete hardening temperature curves, drying shrinkage rate, fluidity index, peak mixing pressure, and pouring vibration frequency data in real time.

[0082] The process of acquiring the concrete hardening temperature curve in real time by deploying temperature and humidity sensors includes:

[0083] Before concrete pouring, temperature and humidity sensors are embedded at a predetermined depth (e.g., 50mm from the surface) in a quincunx pattern with a spacing of ≤1.5m to ensure coverage of different depths and locations. For large-scale projects, multi-point deployment is adopted, with each monitoring point equipped with one main sensor and one backup sensor to avoid single-point failure.

[0084] During installation, ensure the temperature and humidity sensor is in close contact with the concrete to avoid air bubbles or gaps, and use thermally conductive adhesive to fix it to reduce thermal resistance; surface sensors are fixed to the structural surface by magnetic attraction or adhesive to ensure long-term stability.

[0085] The temperature and humidity sensor collects data at a preset frequency, packages and transmits it to the gateway via the LoRa protocol, and the gateway verifies the data integrity through CRC check. If the verification fails, a retransmission is triggered. The collection frequency is divided into:

[0086] Initial coagulation stage (0-6 hours): Samples are collected once per minute to capture temperature peaks;

[0087] Final condensation stage (6-24 hours): Samples are collected every 5 minutes to monitor the cooling rate;

[0088] Late hardening stage (>24 hours): Samples were collected every 30 minutes to assess long-term stability;

[0089] During data preprocessing, short-term fluctuations and random noise are eliminated by sliding a fixed-size window across the data sequence and calculating the average value of the data within the window. Then, specific upper and lower thresholds are set to determine whether data points are abnormal, such as temperatures >60℃ or <-5℃, and abnormal data are marked.

[0090] For marked anomalous data, cubic spline interpolation is used for reconstruction first. If the anomalous data continues for more than 3 cycles, the backup sensor is activated and the event is recorded to the blockchain.

[0091] Generate temperature curve reports, including fitting formulas and key parameters such as maximum temperature and heating rate. Data is stored on the blockchain to ensure traceability and support quality traceability and auditing.

[0092] The temperature data point is set as follows during the initial setting stage. The fitted polynomial is:

[0093] ,in, It is the order of the polynomial. These are the coefficients to be solved. To find the best-fit polynomial, we minimize the sum of squared errors. Solve for the coefficients;

[0094] make The Vandermonde matrix has the following elements. , A column vector of temperature data. Let be the column vector of polynomial coefficients, then we have Solving for the polynomial coefficients yields... ;

[0095] The final setting stage is converted into a quadratic polynomial (k=2) to simplify computational resource consumption. To ensure the temperature rise process of concrete under adiabatic conditions, its mathematical expression is: ,in, Indicates time The concrete temperature at that time; This represents the final temperature rise of the concrete, that is, the difference between the concrete temperature and the initial temperature as time approaches infinity. It is the heating rate constant, which determines the rate at which the concrete temperature rises. The larger the value, the faster the concrete temperature rises; The ambient temperature means that the influence of the external environment on the concrete temperature during the insulation process can be considered constant.

[0096] To solve for the parameters in the model , and The specific process includes the following:

[0097] Collect temperature data of concrete at different time points and corresponding time ;

[0098] Take the natural logarithm of both sides of the mathematical expression of the model. This allows the original model to be converted into a linear form;

[0099] The transformed linear equation was fitted using the least squares method, and the collected temperature data was used as the basis for the calculation. and time data Substitute into the equation, and adjust , and The value of minimizes the sum of squared errors between the fitted result and the actual data. After linear fitting, we can obtain... , and The estimated value;

[0100] The process of calculating slump values ​​in real time by deploying slump sensors includes:

[0101] Place the slump sensor near the concrete mixer truck outlet or pouring point to ensure that the sensor container is in full contact with the concrete. During installation, the sensor should be placed horizontally to avoid external vibration interference and there should be no obstructions around it to ensure that the concrete can be poured in smoothly.

[0102] Real-time collection of sensor output signals and current temperature Calculate slump using the basic model Its mathematical expression is: Set a reasonable slump range, such as 80mm ≤ S ≤ 180mm. Data points outside this range will be marked as abnormal. For transient abnormalities (such as sensor noise), use median filtering to replace the abnormal values. If there are 3 consecutive abnormal points, notify the construction personnel to check the concrete mix ratio or sensor status.

[0103] By calculating the rate of change of slump If the rate of change suddenly exceeds 5 mm / s, an abnormal alarm will be triggered.

[0104] Calculate the mean (μ) and standard deviation (σ) of the data. If a data point exceeds the mean ± 3 times the standard deviation (μ ± 3σ), it is considered an outlier.

[0105] The process of calculating the drying shrinkage rate by collecting concrete drying shrinkage data in real time includes:

[0106] Set the drying shrinkage rate of concrete to be... The length of the concrete specimen in its initial state is The length of the concrete specimen after drying is The reference length selected when measuring the change in concrete length is Through formula This allows for the quantification of the degree of shrinkage of concrete during the drying process;

[0107] Methods for calculating the concrete fluidity index include:

[0108] When concrete cannot flow to the far right of the box, its fluidity is restricted, and its fluidity index decreases. Calculated using the following formula:

[0109] ,in, This refers to the actual flow length, which is the farthest distance the concrete can flow to within a specified time. This is the total length of the box;

[0110] When the concrete can flow to the far right of the box, its fluidity index is... Calculated using the following formula:

[0111] ,in, This refers to the height of the bottom of the box or a specific location after the concrete flow has stopped. The height at the same position when the concrete is in its initial state;

[0112] The process of acquiring real-time data on peak mixing pressure and pouring vibration frequency by deploying pressure and vibration sensors includes:

[0113] During the concrete mixing process, peak mixing pressure detection is used to first collect pressure data over a period of time and then calculate the peak pressure during that period. and average pressure Then calculate the peak-to-average ratio. ,when >2.5 and When the pressure exceeds 500 kPa, a Level 3 alarm is triggered. Simultaneously, a CNN model is used to analyze the current data, and adjustments are made dynamically based on the model's output. The threshold;

[0114] When concrete is poured into the formwork, vibration and impact signals are generated due to the concrete's fluidity and gravity. Vibration sensors use internal piezoelectric crystals to monitor these vibration signals during the concrete pouring process and convert them into electrical signals for output. During this process, a fast Fourier transform is used for spectral analysis, converting the time-domain signal into the frequency domain, which involves the vibration uniformity coefficient. It can be calculated using the following formula:

[0115] ,in, For frequency, The vibration frequency at the measuring point, The average vibration frequency, To determine the number of measurement points, compare each measurement point. Vibration frequency at the measuring point With average vibration frequency The degree of deviation can quantify the uniformity of vibration;

[0116] The criteria for determining the uniformity of vibration are as follows:

[0117] <0.15: The vibration state is judged as "uniform vibration", indicating that the vibration frequency of each measurement point is very close, the overall vibration effect is good, and no adjustment is required;

[0118] 0.15≤ <0.3: The vibration state is considered "slightly uneven". In this case, although there is some vibration imbalance, the degree is not serious. It is possible to consider whether slight optimization or monitoring is needed based on the actual situation.

[0119] ≥0.3: The vibration state is identified as "severely uneven". At this time, the vibration difference between each measurement point is significant, which may lead to uneven structural stress, reduced product quality or other potential problems. The vibration process must be adjusted to improve the vibration distribution and ensure that the expected construction or production effect is achieved.

[0120] S1.3 uploads the data collected by the multi-source sensors to the central processing unit via the LoRa wireless transmission module for data fusion processing, generates a real-time monitoring report, and then intelligently analyzes the concrete pouring quality based on the real-time monitoring report.

[0121] Step Two: Quality Inspection Stage

[0122] S2.1 Construct a convolutional neural network model. The input layer receives a 128×128 pixel infrared thermal image of the concrete surface and a sensor time series data matrix. The sensor time series data matrix contains time series data of temperature, humidity, pressure and vibration sensors. The dimension is T×N, where T is the time step and N is the number of sensors.

[0123] S2.2 sets up four convolutional layers with a kernel size of 3×3 and a number of kernels of 64, 128, 256 and 512 respectively. Each layer is followed by a batch normalization layer and a ReLU activation function.

[0124] S2.3 Insert two max pooling layers with a pooling window of 2×2 and a stride of 2, and a global average pooling layer to reduce the spatial dimension of the feature map.

[0125] The S2.4 output layer uses the Softmax function to classify concrete quality grades and generates a quality report that includes slump deviation rate (ΔSL≤2%), temperature gradient (ΔT / h≤1.5℃), and vibration uniformity coefficient (VI≥0.85).

[0126] In process S2.1, the two-dimensional feature map of the infrared image and the one-dimensional time-series vector of the sensor data are concatenated and fused through channels to form a multimodal input tensor, which is expressed by the following formula:

[0127] ,in, For the fused multimodal input tensor, For the feature map tensor of the infrared image, For sensor time-series data;

[0128] The formula for the convolution operation in process S2.2 is expressed as follows:

[0129] ,in, For the first The output of the layer at position (i,j) For the first The convolutional kernel weights of the layer at position (m,n) For the first The input at position (i+m, j+n) of the previous layer. For the first Layer bias terms, Batch normalization is used to accelerate training and improve model stability. A nonlinear transformation is introduced for the activation function;

[0130] In process S2.3, for each position (i,j) in the output feature map, its value... It is the maximum value within the corresponding 2×2 window in the input feature map, expressed by the following formula:

[0131] ;

[0132] In process S2.4, the Softmax function is: ,in, The feature vector after global average pooling. , These are the weight vector and bias term for the k-th class, respectively. , These are the weight vector and bias term for class c, respectively, where C is the total number of classes. Here, C=3, representing three classes: qualified, warning, and unqualified. The probability distribution of each quality level is calculated using the Softmax function to finally determine the grade classification of the concrete sample.

[0133] The formula for slump deviation rate is: ,in, This indicates the final value of the parameter or quantity. This indicates the initial value of the parameter or quantity;

[0134] The temperature gradient formula is: ,in, Indicates the final temperature. Indicates the initial temperature;

[0135] Step 3: Early Warning and Control Phase

[0136] S3.1 When the system detects any of the following abnormal conditions in real time through temperature and humidity sensors and slump sensors deployed at the construction site, it will automatically activate the graded early warning procedure:

[0137] When the internal temperature of the concrete is between 30°C and 35°C or the slump loss rate is between 10% and 15%, a Level 1 Yellow Alert is issued. Structured alert information is pushed to the construction worker's handheld terminal (running a customized construction management app) via a LoRa wireless communication module (operating frequency band 433MHz, transmission distance ≥2km). The alert includes: anomaly type (temperature exceeding limits / slump loss), real-time data (current temperature value / SLR percentage), location information (based on GPS module positioning, accuracy ±2m), and suggested remedial measures (such as increasing shading measures / adjusting the amount of water-reducing agent).

[0138] When the internal temperature of the concrete is between 35°C and T, and the slump loss rate is between 15% and SLR, and the slump loss rate is between 20% and 15%, a Level II orange alert is issued. The speed of the concrete delivery pump (model HBT80.16.110SZ) should be adjusted using the following formula:

[0139] ;

[0140] When the internal temperature of the concrete (T) ≥ 40°C or the slump loss rate (SLR) ≥ 20%, a Level III red alert is issued. The power supply to the delivery pump is cut off via a relay module (model G3NA-210B), and the pouring valve is simultaneously closed (pneumatic actuator response time ≤ 0.5s). The cooling water pipe system embedded in the concrete structure (DN50 diameter, 1.5m spacing) is then activated. The cooling water flow rate is controlled according to the following formula:

[0141] This ensures rapid cooling in emergency situations, preventing the concrete structure from losing strength due to high temperatures, and effectively guaranteeing construction quality and safety.

[0142] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0143] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A construction site ready-mixed concrete placement management method, characterized by, The method comprises the following steps: Step one data acquisition stage: S1.1 By scanning the two-dimensional code on the packaging of concrete raw materials, automatically input the variety, batch number, and production date information of cement, aggregate, fly ash, and admixtures; S1.2 Deploy temperature and humidity sensors, slump sensors, pressure sensors, and vibration sensors to collect real-time data on concrete hardening temperature curves, drying shrinkage, fluidity index, mixing pressure peak, and pouring vibration frequency data; S1.3 Upload the data collected by the multi-source sensors to the central processing unit through the LoRa wireless transmission module for data fusion processing to generate real-time monitoring reports, and then intelligently analyze the concrete pouring quality according to the real-time monitoring reports; Step two quality detection stage: S2.1 Build a convolutional neural network model, the input layer receives a 128x128 pixel infrared thermal imaging image of the concrete surface and a sensor time series data matrix, which includes temperature and humidity, pressure, and vibration sensor time series data with a dimension of T1xN, T1 being the time step and N being the number of sensors; S2.2 Set four convolutional layers with a kernel size of 3x3, with 64, 128, 256, and 512 respectively, each followed by a batch normalization layer and a ReLU activation function; S2.3 Insert two layers of maximum pooling layer with a pooling window of 2x2 and a step of 2, and a global average pooling layer to reduce the spatial dimension of the feature map; S2.4 The output layer uses the Softmax function to classify the concrete quality grade and generates a quality report containing the slump deviation rate, temperature gradient, and vibration uniformity coefficient; Step three early warning control stage: When the system detects any of the following abnormal conditions in real time through the temperature and humidity sensors and slump sensors deployed on the construction site, it will automatically start the hierarchical warning program: When the internal temperature of the concrete is 30°C≤T<35°C or the slump loss rate is 10%≤SLR<15%, it is a first-level yellow warning, and the structured warning information is pushed to the construction worker's handheld terminal through the LoRa wireless communication module; When the internal temperature of the concrete is 35°C≤T<40°C or the slump loss rate is 15%≤SLR<20%, it is a second-level orange warning, and the speed adjustment is executed through the linked concrete delivery pump, with the adjustment amount calculated according to the following formula: ； When the internal temperature of the concrete is T≥40°C or the slump loss rate is SLR≥20%, it is a third-level red warning, and the power supply of the delivery pump is cut off through the relay module, and the pouring valve is closed, with a response time of the pneumatic actuator ≤0.5s, and the cooling water pipe system embedded in the concrete structure is started, with a cooling water flow controlled according to the following formula: Q=5±0.5 liters / second, to ensure rapid cooling in emergency situations and prevent the concrete structure from being damaged due to high temperature, effectively ensuring construction quality and safety.

2. A construction site ready-mix concrete placement management method according to claim 1, characterized in that: The process of automatically entering information by scanning the two-dimensional code on the packaging of concrete raw materials includes: reading the two-dimensional code data using a handheld terminal device, transmitting it to the central database in real time, then the system automatically compares the entered information with the construction plan requirements to ensure material compliance, analyzes the entered data in real time, and generates a material usage report. If the entered information does not match the construction plan requirements, the generated material usage report will have a "mismatch" label and will prompt adjustment suggestions to ensure that construction personnel check and temporarily suspend the use of the batch of materials. Otherwise, a "match" label will be generated, allowing the use of the batch of materials.

3. A construction site ready-mix concrete placement management method according to claim 1, characterized in that: The process of real-time collection of concrete hardening temperature curve by deploying temperature and humidity sensors includes: Before pouring concrete, bury the temperature and humidity sensors at a predetermined depth, arrange them in a quincunx shape, and ensure a spacing of ≤1.5m to cover different depths and positions. For large projects, use multiple point arrangements, with one main sensor and one backup sensor at each monitoring point to avoid single-point failure. When installing, ensure that the temperature and humidity sensors are in close contact with the concrete to avoid air bubbles or gaps, and use thermal conductive glue to fix them to reduce thermal resistance. Surface sensors are fixed to the structure surface through magnetic attraction or adhesive to ensure long-term stability. The temperature and humidity sensors collect data at a preset frequency, package the data through LoRa protocol, and transmit it to the gateway. The gateway verifies the data integrity through CRC check, and triggers retransmission if it fails. The collection frequency is divided into: Initial setting stage: collect data every minute to capture temperature peak value; Final setting stage: collect data every 5 minutes to monitor cooling rate; Hardening later stage: collect data every 30 minutes to evaluate long-term stability; In the data preprocessing process, a fixed size window is slid on the data sequence, and the average value of the data in the window is calculated to eliminate short-term fluctuations and random noise. Then, specific upper and lower threshold values are set to determine whether the data points are abnormal, and the abnormal data is marked. For the marked abnormal data, preferentially use cubic spline interpolation reconstruction, and if continuous abnormality exceeds 3 periods, start the backup sensor and record the event to the blockchain. Generate a temperature curve report containing the fitting formula, key parameters, and data storage on the chain to ensure traceability and support quality traceability and audit. To solve the parameters A, B, and C in the model, the specific process includes the following:

4. A construction site ready-mix concrete placement management method according to claim 3, characterized in that: The temperature data points for the initial set stage are , and the fitted polynomial is: where k is the order of the polynomial, are the coefficients to be solved for, in order to find the best fitting polynomial by minimizing the sum of squared errors solving for the coefficients; Let be the Vandermonde matrix with elements , T be the temperature data column vector, be the polynomial coefficient column vector, then , the polynomial coefficients are solved as ; The final setting stage is converted to a quadratic polynomial k = 2, which simplifies the calculation of resource consumption. In order to ensure the temperature rise process of concrete under adiabatic conditions, the mathematical expression is: Wherein, T(t) represents the temperature of concrete at time t; A represents the final temperature rise of concrete, that is, the difference between the temperature of concrete and the initial temperature when the time tends to infinity; B is the temperature rise rate constant, which determines the speed of the temperature rise of concrete. The greater the value of B, the faster the temperature rise of concrete; C is the ambient temperature, that is, the influence of the outside world on the temperature of concrete during the adiabatic process can be regarded as constant; The process of real-time calculation of slump value by deploying slump sensors includes: Collecting temperature data of the concrete at different points in time and the corresponding time t; Taking natural logarithm on both sides of the mathematical expression of the model This can convert the original model into linear form; The converted linear equations are fitted using the least square method, and the collected temperature data and time data t are substituted into the equations, and the values of A, B and C are adjusted to minimize the sum of the square errors between the fitting results and the actual data. After linear fitting, the estimated values of A, B and C are obtained.

5. A construction site ready-mix concrete placement management method according to claim 1, characterized in that: Place the slump sensors near the concrete mixer truck outlet or pouring point to ensure that the sensor container is in full contact with the concrete. When installing, the sensor should be placed horizontally to avoid external vibration interference and ensure that the concrete is poured smoothly without any obstructions. Calculate the mean μ and standard deviation σ of the data. If a data point exceeds ±3 times the standard deviation μ ± 3σ, it is considered an outlier. The sensor output signal D and the current temperature T are collected in real time, and the slump S is calculated using a basic model, and the mathematical expression is: A reasonable slump range of 80mm≤S≤180mm is set, and data points exceeding this range will be marked as abnormal. For temporary abnormal sensor noise, the median filter is used to replace the abnormal value. If there are three consecutive abnormal points, the construction personnel is notified to check the concrete proportioning or the sensor state. By calculating the rate of change of slump If the rate of change jumps more than 5 mm / s, an anomaly alert is triggered The process of calculating the drying shrinkage rate by real-time collection of concrete drying shrinkage rate includes:

6. A construction site ready-mix concrete placement management method according to claim 1, characterized in that: The method for calculating the concrete fluidity index includes: The dry shrinkage rate of the concrete is set as , the length of the concrete test piece in the initial state is , the length of the concrete test piece after drying is , the reference length selected when measuring the length change of the concrete is , and the shrinkage degree of the concrete in the drying process is quantified through the formula .

7. A construction site ready-mix concrete placement management method according to claim 1, characterized in that: The process of real-time collection of mixing pressure peak value and pouring vibration frequency data by deploying pressure sensors and vibration sensors includes: When the concrete cannot flow to the right end of the box, in this case, the flowability of the concrete is limited, and its flowability index This is calculated by the following equation: wherein, is the actual flow length, i.e. the furthest distance the concrete can flow in a given time, is the total length of the box. When the concrete is able to flow to the right end of the box, its flowability index This is calculated by the following equation: wherein, H1 is the height of the bottom of the box or a specific location after the concrete flow has stopped, H2 is the height of the same location when the concrete is in the initial state.

8. A construction site ready-mix concrete placement management method according to claim 1, characterized in that: ​ In the concrete mixing process, the peak value of mixing pressure is detected, the pressure data in a period of time is collected, the peak value of pressure in the period of time is calculated , and the average pressure , then the peak-to-average ratio is calculated When > 2.5 and > 500kPa, trigger the third level alarm, at the same time, use the CNN model to analyze the current data, and dynamically adjust the threshold value according to the output of the model; When the concrete is poured into the formwork, due to the fluidity of the concrete and the action of gravity, vibration and impact signals are generated, and the vibration sensor utilizes the internal piezoelectric crystal to monitor the vibration signals during the concrete pouring process and converts them into electrical signals for output. In the process, the fast Fourier transform is used for frequency spectrum analysis, converting the time domain signal into the frequency domain, involving the vibration uniformity coefficient , which is calculated by the following formula: wherein, is the frequency, is the frequency of vibration at the measurement point, is the average frequency of vibration, n is the number of measurement points, by comparing the degree of deviation of the frequency of vibration at each measurement point n from the average frequency of vibration the uniformity of the vibration can be quantified.

9. A construction site ready-mix concrete placement management method according to claim 1, characterized in that: In the S2.1 process, the two-dimensional feature map of the infrared image and the one-dimensional time sequence vector of the sensor data are fused through channel splicing to form a multi-modal input tensor, which is represented by the following formula: wherein, is the fused multi-modal input tensor, is the feature map tensor for the infrared image, is the tensor for sensor timing data; In the S2.2 process, the formula of the convolution operation is represented as: where, is the output of the layer at position (i, j), is the convolution kernel weight of the layer at position (m, n), is the input of the layer at position (i + m, j + n), is the bias term of the layer, is the batch normalization, used to accelerate training and improve model stability, is the activation function, which introduces a non-linear transformation; In the S2.3 process, for each position (i,j) in the output feature map, its value is the maximum value within the corresponding 2x2 window in the input feature map, represented by the following equation: ； In the S2.4 process, the Softmax function is wherein, is the feature vector after global average pooling, , is the weight vector and bias item of the kth class respectively, , is the weight vector and bias item of the cth class respectively, C is the total number of classes, here C=3, representing three classes of qualified, early warning and unqualified, the probability distribution of each quality grade is calculated through the Softmax function, and finally the grade attribution of the concrete sample is determined; The slump deviation rate formula is: wherein, represents the final value of the parameter or quantity, represents the initial value of the parameter or quantity; The temperature gradient formula is: wherein, represents the final temperature, represents the initial temperature.

Images