Intelligent construction site energy-saving control method and system based on Internet of Things

Abstract

The invention discloses an intelligent construction site energy-saving control method and system based on the Internet of Things, and the method comprises the steps: building a dynamic network model for image feature extraction through obtaining a real-time image of a construction site, the operation state of energy storage equipment and personnel positioning data, and obtaining a scene recognition result through a scene recognition model; and then extracting energy demand characteristics, and obtaining predicted demand power through a power prediction model. The system adaptively adjusts the operation time sequence of the non-key equipment according to a comparison result of the demand matching degree and a preset threshold value; meanwhile, a Kalman filtering algorithm is adopted to carry out de-noising processing on personnel positioning data, and the operation distance between personnel and equipment is calculated in combination with image features. When the operation distance exceeds the preset threshold value, the system automatically triggers the standby energy-saving mode, and precise control over the energy-saving mode is achieved through energy consumption monitoring. According to the method, the problems of energy waste and low operation efficiency caused by insufficient scene recognition precision, unreasonable switching time of the energy storage equipment and misjudgment of the standby energy-saving mode are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of Internet of Things, and in particular to an energy-saving control method and system for a smart construction site based on the Internet of Things. Background Art

[0002] At present, with the continuous advancement of smart construction site construction, energy-saving control systems, as a core component, are facing new technical challenges and development opportunities. In the smart construction site environment, accurate energy management requires real-time monitoring and identification of construction site scenes, which involves the in-depth application of image acquisition and scene recognition technology. However, with the increase in the complexity of construction site scenes, how to achieve real-time processing while ensuring recognition accuracy has become a key technical problem that needs to be solved urgently. At the same time, the dynamic adjustment of energy based on scene recognition introduces the demand for intelligent management of energy storage equipment. How to achieve precise coordination between energy storage equipment and electrical equipment has become an important breakthrough in improving energy utilization efficiency. In this context, the system also needs to be able to intelligently judge the standby conditions of the equipment, which will put higher requirements on personnel positioning technology. The solution to these technical problems is not only related to the performance improvement of smart construction site energy-saving control systems, but also will provide important technical support for the intelligent transformation of the construction industry.

[0003] In a traditional technical approach, first, the system collects temperature and light data from various areas within the construction site through a network of fixed sensors, which are connected to a central controller by wire. The central controller determines the real-time status of the construction site area based on preset thresholds, such as turning on the area lighting equipment when the light is below a certain threshold. Next, the system controls the operating time of large equipment through a simple timer mechanism, such as shutting down non-essential equipment during the preset night time period to save energy. For the management of energy storage equipment, the system adopts a fixed charging and discharging strategy, such as charging during the low-peak electricity price period at night and discharging during the peak period during the day, but this strategy does not take into account the dynamic changes in the actual electricity demand of the construction site. Finally, the system obtains personnel location information through manual inspections and manual recording, which is used to determine the standby conditions of the equipment, such as shutting down the corresponding equipment when there is no one in a certain area.

[0004] In the energy-saving control system of smart construction sites, traditional technical means have obvious defects. Scene recognition relies on simple rules or fixed thresholds, with low accuracy and poor dynamic adaptability. It cannot reflect complex scene changes in real time, affecting the accuracy of energy adjustment. The energy storage device switching adopts a fixed charging and discharging strategy or a preset schedule, which cannot dynamically respond to changes in energy demand. It may cause the switching timing to not match the actual demand, resulting in energy waste or equipment interruption. The standby energy-saving mode is triggered based on simple sensors or time delays to judge the presence of personnel, which is susceptible to environmental interference and insufficient positioning accuracy. It may lead to misjudgment and affect normal operations or energy-saving effects. Traditional technical means lack real-time, accuracy and dynamic optimization capabilities, and it is difficult to meet the energy-saving needs of smart construction sites. Summary of the invention

[0005] The present invention provides an energy-saving control method and system for a smart construction site based on the Internet of Things, so as to solve the problems of energy waste and low operating efficiency caused by insufficient scene recognition accuracy, unreasonable switching timing of energy storage equipment and misjudgment of standby energy-saving mode.

[0006] In the first aspect, in order to solve the above technical problems, the present invention provides an energy-saving control method for a smart construction site based on the Internet of Things, comprising: Obtain real-time images of the construction site, the operating status of the energy storage equipment, and personnel location data, where the operating status includes the discharge power; Using a pre-built dynamic network model to perform preliminary feature extraction on the real-time image to obtain image features; Inputting the image features into a pre-trained scene recognition model to obtain a scene recognition result; Extracting energy demand characteristics from the scene recognition results and inputting them into a pre-built power prediction model to obtain predicted demand power; Calculating a demand matching degree according to the discharge power and the demand power, and if the demand matching degree is less than a preset matching degree threshold, adjusting the operating timing of non-critical equipment according to the criticality value of the equipment; A Kalman filter algorithm is used to perform denoising on the personnel positioning data to obtain accurate position data; Calculating the working distance between the person and the equipment according to the precise location data and the image features, and triggering the standby energy-saving mode if the working distance exceeds a preset distance threshold; After the standby energy-saving mode is triggered, the energy consumption of the equipment is monitored. If the energy consumption of the equipment is lower than the preset energy consumption threshold and the position of the personnel has not changed, the standby state is maintained, otherwise the normal operating mode is restored.

[0007] In an optional implementation, the pre-built dynamic network model is used to perform preliminary feature extraction on the real-time image to obtain image features, including: Using a dynamic weight adjustment method to perform multi-scale feature extraction on the real-time image to obtain preliminary image features; Inputting the preliminary image features into a pre-built convolutional neural network model for feature optimization to obtain optimized features; Inputting the optimized features into a pre-built support vector machine classifier to perform image classification to obtain image features; Wherein, the dynamic network model includes a convolutional neural network model and a support vector machine classifier.

[0008] In an optional implementation, the inputting the image features into a pre-trained scene recognition model to obtain a scene recognition result includes: Inputting the image features into a preliminary scene recognition model to perform current scene recognition, and obtaining a preliminary scene recognition result and confidence level; Comparing the confidence level with a preset confidence threshold, and if the confidence level is greater than or equal to the confidence threshold, determining that the preliminary scene recognition result is qualified, and obtaining a scene recognition result; If the confidence level is less than the confidence threshold, the preliminary scene recognition result is judged to be unqualified, and the image features are input into the precise scene recognition model for precise scene recognition to obtain a precise scene recognition result, which is used as the final scene recognition result; The scene recognition model includes a preliminary scene recognition model and a precise scene recognition model.

[0009] In an optional implementation manner, the process of constructing the power prediction model includes: Perform data cleaning and feature extraction on the pre-stored historical energy demand characteristics to obtain the cleaned historical energy characteristics; Inputting the cleaned historical energy characteristics into the input layer of the initially constructed neural network model for training, and obtaining the predicted value output by the output layer of the neural network model; Substitute the predicted value and the pre-stored actual value into the loss function to calculate the loss value and obtain the loss value data; Calculate the gradient of the output layer output of the neural network model according to the loss value data, and pass the gradient forward layer by layer through the chain rule to calculate the gradient of each layer parameter to obtain gradient data; Update the parameters of each layer of the neural network model according to the gradient data and the preset learning rate; The parameters of each layer are updated repeatedly and iteratively until the training times of the neural network model are greater than the preset training times, or the loss value data of the neural network model is less than the preset loss threshold, the training is determined to be completed, and a power prediction model is obtained.

[0010] In an optional implementation, the denoising process of the personnel positioning data using a Kalman filter algorithm to obtain accurate position data includes: Cleaning the personnel positioning data of outliers to obtain pre-processed positioning data; Using a Kalman filter algorithm to perform denoising on the pre-processed positioning data to obtain denoised positioning data; Extracting the personnel ID, coordinate value and movement trajectory according to the denoised positioning data, and calculating the movement speed according to the time series of the coordinate value; The personnel ID, the coordinate value, the moving track and the moving speed are used as precise location data; Among them, the outlier is determined by the following formula: The moving speed is calculated by the following formula: in, Indicated in The moment Personnel location data Axis coordinates, Indicates personnel location data The mean of the axis coordinates, Indicates personnel location data The standard deviation of the axis coordinates, Indicates the preset abnormal threshold. Indicated in Moment The precise location of a person Axis coordinates, Indicated in Moment The precise location of a person Axis coordinates, represents the pre-stored sampling time step, Indicates the movement speed.

[0011] In an optional implementation, the calculating the operating distance between the person and the equipment according to the precise location data and the image features includes: The working distance between personnel and equipment is calculated by the following formula: in, Indicates the preset device location, Indicated in Moment The precise location of a person, Indicates The working distance of each person.

[0012] In an optional implementation, the calculating the demand matching degree according to the discharge power and the demand power, and if the demand matching degree is less than a preset matching degree threshold, adjusting the operating timing of non-critical equipment according to the criticality value of the equipment, includes: Calculating according to the discharge power and the required power to obtain a demand matching degree; If the demand matching degree is less than a preset matching degree threshold, the construction site equipment is adjusted according to the predicted demand power, including re-planning the operating timing of non-critical equipment; The demand matching degree is calculated by the following formula: in, Indicates the degree of demand matching. Indicates the required power, Indicates the discharge power.

[0013] In a second aspect, the present invention provides a smart construction site energy-saving control system based on the Internet of Things, comprising: A data acquisition module is used to obtain real-time images of the construction site, the operating status of the energy storage equipment, and personnel positioning data, where the operating status includes the discharge power; An image feature analysis module, used to perform preliminary feature extraction on the real-time image using a pre-built dynamic network model to obtain image features; A scene recognition module, used to input the image features into a pre-trained scene recognition model to obtain a scene recognition result; A demand prediction module, used to extract energy demand characteristics from the scene recognition results and input them into a pre-built power prediction model to obtain predicted demand power; A demand analysis module, configured to calculate a demand matching degree according to the discharge power and the demand power, and if the demand matching degree is less than a preset matching degree threshold, adjust the operating timing of non-critical equipment according to the criticality value of the equipment; A data filtering module, used to perform denoising on the personnel positioning data using a Kalman filtering algorithm to obtain accurate position data; An operation distance calculation module, used to calculate the operation distance between the personnel and the equipment according to the precise location data and the image features, and trigger a standby energy-saving mode if the operation distance exceeds a preset distance threshold; The energy consumption monitoring module is used to monitor the energy consumption of the equipment after the standby energy-saving mode is triggered. If the energy consumption of the equipment is lower than the preset energy consumption threshold and the position of the personnel has not changed, the standby state is maintained, otherwise the normal operation mode is restored.

[0014] In a third aspect, the present invention further provides an electronic device comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, it implements any one of the above-mentioned smart construction site energy-saving control methods based on the Internet of Things.

[0015] In a fourth aspect, the present invention further provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the above-mentioned smart construction site energy-saving control methods based on the Internet of Things.

[0016] Compared with the prior art, the present invention has the following beneficial effects: (1) The present invention uses a dynamic network model to perform preliminary feature extraction on real-time images, and combines it with a pre-trained scene recognition model to perform scene recognition, which can more accurately identify changes in complex construction site scenes and improve the accuracy and dynamic adaptability of scene recognition.

[0017] (2) The present invention calculates the demand matching degree based on the predicted demand power and the discharge power of the energy storage device, and dynamically adjusts the charging and discharging strategy of the energy storage device, which can more accurately meet the actual power demand of the construction site, improve energy utilization efficiency, and avoid energy waste or equipment interruption.

[0018] (3) The present invention uses the Kalman filter algorithm to perform denoising on the personnel positioning data to obtain accurate location data, and combines the image features to calculate the operating distance between the personnel and the equipment, which can more accurately determine the presence of personnel, thereby more reliably triggering the standby energy-saving mode, avoiding misjudgment, and improving energy-saving effects. When the operating distance exceeds the preset threshold, the standby energy-saving mode is triggered to avoid misjudgment caused by the temporary departure of personnel or environmental interference, thereby improving energy-saving effects and not affecting normal operations.

[0019] (4) The present invention collects various data of the construction site in real time through the Internet of Things technology, and uses dynamic network models, scene recognition models and power prediction models to process and analyze the data in real time, which can quickly respond to changes in the construction site scene and energy demand. For example, when the scene recognition model detects that the energy demand in the construction site area has changed, the power prediction model can predict the required power in real time and dynamically adjust the charging and discharging strategy of the energy storage device to achieve dynamic optimization of energy management, thereby improving the real-time performance and dynamic optimization capabilities of the system.

[0020] In summary, the present invention improves and optimizes traditional technical means from multiple aspects by introducing technical means such as the Internet of Things, dynamic network models, scene recognition models, power prediction models, and Kalman filtering algorithms. For example, in scene recognition, the accuracy and dynamic adaptability are improved; in energy storage equipment management, dynamic response to changes in energy demand is achieved; in personnel positioning, the positioning accuracy and reliability of the standby energy-saving mode are improved. These improvements and optimization measures jointly enhance the overall performance of the smart construction site energy-saving control system, enabling it to better meet the energy-saving needs of smart construction sites and provide strong support for the intelligent transformation of the construction industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a schematic flow chart of a smart construction site energy-saving control method based on the Internet of Things provided by the first embodiment of the present invention; Figure 2 It is a schematic diagram of the structure of a smart construction site energy-saving control system based on the Internet of Things provided in the second embodiment of the present invention. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] Reference Figure 1 The first embodiment of the present invention provides an energy-saving control method for a smart construction site based on the Internet of Things, comprising the following steps: S11, obtaining a real-time image of the construction site, the operating status of the energy storage equipment, and personnel location data, wherein the operating status includes the discharge power; S12, using a pre-built dynamic network model to perform preliminary feature extraction on the real-time image to obtain image features; S13, inputting the image features into a pre-trained scene recognition model to obtain a scene recognition result; S14, extracting energy demand characteristics from the scene recognition results, and inputting them into a pre-built power prediction model to obtain predicted demand power; S15, calculating a demand matching degree according to the discharge power and the demand power, and if the demand matching degree is less than a preset matching degree threshold, adjusting the operating timing of non-critical equipment according to the criticality value of the equipment; S16, using a Kalman filter algorithm to perform denoising on the personnel positioning data to obtain accurate position data; S17, calculating the working distance between the person and the equipment according to the precise location data and the image features, and triggering the standby energy-saving mode if the working distance exceeds a preset distance threshold.

[0024] S18, after the standby energy-saving mode is triggered, monitor the energy consumption of the equipment. If the energy consumption of the equipment is lower than the preset energy consumption threshold and the position of the personnel has not changed, maintain the standby state, otherwise resume the normal operation mode.

[0025] In step S11 , a real-time image of the construction site, the operating status of the energy storage device, and personnel positioning data are acquired, wherein the operating status includes the discharge power.

[0026] In this embodiment, real-time images are collected by high-definition cameras deployed at key locations on the construction site. The cameras capture the environmental images of the construction site at fixed time intervals and transmit the data to the central processing system. The operating status of the energy storage device is obtained through sensors installed on the equipment. The sensors monitor the charging power, discharge power and charge state of the battery pack in real time, and transmit these data to the central database through the Internet of Things communication protocol. Personnel positioning data is collected through smart terminal devices (such as smart bracelets or positioning tags). The terminal devices have built-in GPS modules and Bluetooth beacons, support high-precision positioning, and upload personnel location information to the central system in the form of time series. Real-time images are used to identify the current activity scenes, equipment distribution and personnel flow of the construction site; the operating status data of the energy storage device is used to evaluate the current energy supply capacity of the equipment and its health status; and the personnel positioning data is used to analyze the location distribution and movement trajectory of personnel. By integrating these data, the system can fully grasp the real-time status of the construction site.

[0027] In step S12, a pre-built dynamic network model is used to perform preliminary feature extraction on the real-time image to obtain image features.

[0028] In a specific implementation, the pre-built dynamic network model is used to perform preliminary feature extraction on the real-time image to obtain image features, including: Using a dynamic weight adjustment method to perform multi-scale feature extraction on the real-time image to obtain preliminary image features; Inputting the preliminary image features into a pre-built convolutional neural network model for feature optimization to obtain optimized features; Inputting the optimized features into a pre-built support vector machine classifier to perform image classification to obtain image features; Wherein, the dynamic network model includes a convolutional neural network model and a support vector machine classifier.

[0029] Specifically, firstly, a dynamic weight adjustment method is used to extract multi-scale features from the real-time image to obtain preliminary image features. The dynamic weight adjustment method fuses multi-scale features by weighted summation, and the formula is: in, is a multi-scale feature, For the Dynamic weights of layer features, For the Layer features, is the total number of layers. The calculation of dynamic weights is based on the importance and consistency of features, and the formula is: in, is the preset smoothing coefficient, For the The saliency score of the layer feature is calculated by the mean square error and gradient value of the feature map. Multi-scale feature extraction can capture the detailed information of different scales in the image and enhance the robustness of feature expression.

[0030] The preliminary image features are input into a pre-built convolutional neural network model for feature optimization to obtain optimized features. The convolutional neural network model includes multiple layers of convolution and pooling operations. The convolution operation extracts local features through a sliding window. The formula is: in, is the convolution result, is the convolution kernel, represents the convolution operation, is the bias term. The pooling operation reduces the feature dimension through maximum pooling or average pooling, and the formula is: in, is the pooling result, is the pooling window size. Through multi-layer convolution and pooling operations, the convolutional neural network model can extract high-level abstract features. The formula is: in, To optimize the features, Represents the feature extraction function of a convolutional neural network.

[0031] The optimized features are input into a pre-built support vector machine classifier for image classification to obtain image features. The support vector machine classifier divides the feature space into different categories by solving the hyperplane, and the formula is: in, is the hyperplane normal vector, is the penalty parameter, is the slack variable, is the number of samples. The classification decision maps low-dimensional features to high-dimensional space through the kernel function, and the formula is: in, is the kernel function, is the bias term. Through the support vector machine classifier, the optimized features are mapped to specific scene categories and finally the image features are obtained.

[0032] The convolutional neural network model and the support vector machine classifier optimize parameters through back propagation and gradient descent during the training phase. The loss function of the convolutional neural network is the cross entropy loss, and the formula is: in, is the true label, is the predicted probability. The loss function of the support vector machine is hinge loss, and the formula is: Through training, the model can accurately extract and classify image features, providing precise input for scene recognition and energy demand prediction. Through multi-level feature extraction and optimization, detailed scene information of the construction site is obtained, providing reliable data support for subsequent energy management decisions.

[0033] In step S13, the image features are input into a pre-trained scene recognition model to obtain a scene recognition result.

[0034] In a specific implementation, inputting the image features into a pre-trained scene recognition model to obtain a scene recognition result includes: Inputting the image features into a preliminary scene recognition model to perform current scene recognition, and obtaining a preliminary scene recognition result and confidence level; Comparing the confidence level with a preset confidence threshold, and if the confidence level is greater than or equal to the confidence threshold, determining that the preliminary scene recognition result is qualified, and obtaining a scene recognition result; If the confidence level is less than the confidence threshold, the preliminary scene recognition result is judged to be unqualified, and the image features are input into the precise scene recognition model for precise scene recognition to obtain a precise scene recognition result, which is used as the final scene recognition result; The scene recognition model includes a preliminary scene recognition model and a precise scene recognition model.

[0035] Specifically, the scene recognition model includes a preliminary scene recognition model and a precise scene recognition model, and adopts a two-level recognition mechanism to improve recognition accuracy and robustness. First, the image features are input into the preliminary scene recognition model to perform current scene recognition, and obtain preliminary scene recognition results and confidence. The preliminary scene recognition model is implemented based on a lightweight convolutional neural network, and its output layer uses the Softmax function to calculate the probability distribution of each category, and the formula is: in, For the The probability value of the image feature, is the output layer logarithm of the model, is the total number of features. The category of the preliminary scene recognition result is determined by the maximum probability value, and the formula is: in, is the preliminary scene recognition result. The role of is to find the largest 𝑝𝑖 from the probability values ​​of all categories and return the category label 𝑖 corresponding to the maximum value. The confidence is the maximum probability value corresponding to the preliminary scene recognition result, and the formula is: in, is the confidence level, Indicates the maximum value of the probability value of each category.

[0036] The confidence level is compared with a preset confidence threshold. If the confidence level is greater than the confidence threshold, the preliminary scene recognition result is judged to be qualified, and the scene recognition result is obtained. The confidence threshold is set based on the statistical analysis of historical data and the model performance evaluation, and the formula is: in, is the confidence threshold, is the mean of historical confidence levels, is the standard deviation, is the preset adjustment coefficient. , then the scene recognition result is .

[0037] If the confidence level is less than the confidence threshold, the preliminary scene recognition result is judged to be unqualified, and the image features are input into the precise scene recognition model for precise scene recognition to obtain a precise scene recognition result, which is used as the final scene recognition result. The precise scene recognition model is implemented based on a deep residual network, and its loss function is the cross entropy loss, and the formula is: in, is the loss value, is the true label, is the predicted probability, is the number of samples. The category of the accurate scene recognition result is determined by the maximum probability value, and the formula is: in, For accurate scene recognition results, the final scene recognition result is .

[0038] The preliminary scene recognition model and the precise scene recognition model optimize parameters through back propagation and gradient descent during the training phase. The training data set includes typical scene images of the construction site and their corresponding annotation information. During the training process, the model gradually learns the scene features by minimizing the loss function, which is: in, is the updated parameter, is the original parameter, is the learning rate, is the gradient of the loss function with respect to the parameter. After training, the model can accurately classify different types of scenes and determine whether to enable precise recognition based on the confidence level when in use. The two-level recognition mechanism improves the accuracy and robustness of scene recognition, providing reliable basic data for subsequent energy demand forecasting and energy-saving control.

[0039] In step S14, energy demand characteristics are extracted from the scene recognition results and input into a pre-built power prediction model to obtain predicted demand power.

[0040] In a specific implementation, the process of constructing the power prediction model includes: Perform data cleaning and feature extraction on the pre-stored historical energy demand characteristics to obtain the cleaned historical energy characteristics; Inputting the cleaned historical energy characteristics into the input layer of the initially constructed neural network model for training, and obtaining the predicted value output by the output layer of the neural network model; Substitute the predicted value and the pre-stored actual value into the loss function to calculate the loss value and obtain the loss value data; Calculate the gradient of the output layer output of the neural network model according to the loss value data, and pass the gradient forward layer by layer through the chain rule to calculate the gradient of each layer parameter to obtain gradient data; Update the parameters of each layer of the neural network model according to the gradient data and the preset learning rate; The parameters of each layer are updated repeatedly and iteratively until the training times of the neural network model are greater than the preset training times, or the loss value data of the neural network model is less than the preset loss threshold, the training is determined to be completed, and a power prediction model is obtained.

[0041] Specifically, the power prediction model is built based on a neural network, whose input is the energy demand characteristics and output is the predicted demand power. The construction process of the power prediction model includes data preprocessing, feature extraction, model training and parameter optimization.

[0042] First, the pre-stored historical energy demand characteristics are cleaned and feature extracted to obtain the cleaned historical energy characteristics. Data cleaning includes processing missing values, removing outliers and standardizing data. The formula is: in, is the standardized historical energy demand characteristics, is the historical energy demand characteristics, is the mean of the feature, is the standard deviation of the feature. Feature extraction obtains the cleaned historical energy features by calculating the statistics and correlation of energy demand features.

[0043] The cleaned historical energy features are input into the input layer of the initially constructed neural network model for training, and the predicted values ​​output by the output layer of the neural network model are obtained. The structure of the neural network model includes an input layer, a hidden layer, and an output layer. The activation function of the hidden layer is ReLU, and the formula is: in, is the output of the hidden layer, is the input of the hidden layer. The activation function of the output layer is a linear function, and the formula is: in, is the predicted value, is the weight matrix of the output layer, is the output of the hidden layer, is the bias term of the output layer.

[0044] Substitute the predicted value and the actual value stored in advance into the loss function to calculate the loss value and obtain the loss value data. The loss function uses the mean square error (MSE), and the formula is: in, is the loss value, is the actual value, is the predicted value, is the sample size.

[0045] The gradient of the output layer of the neural network model is calculated based on the loss value data, and the gradient is passed forward layer by layer through the chain rule to calculate the gradient of each layer parameter to obtain the gradient data. The gradient formula of the output layer is: The gradient of the hidden layer is calculated using the chain rule, and the formula is: in, is the weight matrix of the hidden layer, is the bias term of the hidden layer.

[0046] According to the gradient data and the preset learning rate, update the parameters of each layer of the neural network model. The parameter update formula is: in, is the learning rate, and is the updated parameter, and is the original parameter.

[0047] The parameters of each layer are updated repeatedly and iteratively until the number of training times of the neural network model is greater than the preset number of training times, or the loss value data of the neural network model is less than the preset loss threshold, the training is determined to be completed, and the power prediction model is obtained. The number of training times and the loss threshold are set based on the model performance evaluation to ensure the convergence and prediction accuracy of the model. After the training is completed, the power prediction model can accurately predict the required power according to the input energy demand characteristics, and provide data support for the energy management of the construction site. Through the data-driven modeling method, accurate prediction of energy demand is achieved, thereby optimizing energy utilization efficiency.

[0048] In step S15, the demand matching degree is calculated according to the discharge power and the demand power. If the demand matching degree is less than a preset matching degree threshold, the operating sequence of non-critical equipment is adjusted according to the criticality value of the equipment.

[0049] In a specific implementation, the calculating the demand matching degree according to the discharge power and the demand power, and if the demand matching degree is less than a preset matching degree threshold, adjusting the operating timing of non-critical equipment according to the criticality value of the equipment, includes: Calculating according to the discharge power and the required power to obtain a demand matching degree; If the demand matching degree is less than a preset matching degree threshold, the construction site equipment is adjusted according to the predicted demand power, including re-planning the operating timing of non-critical equipment; The demand matching degree is calculated by the following formula: in, Indicates the degree of demand matching. Indicates the required power, Indicates the discharge power.

[0050] Specifically, first, according to the discharge power and the required power Calculate and get the demand matching degree The demand matching calculation formula is: in, Indicates the degree of demand matching. Indicates the required power, Indicates the discharge power. The demand matching degree is used to measure the matching degree between the current discharge power and the demand power. Its value range is The closer the value is to 1, the higher the matching degree is; the closer the value is to 0, the lower the matching degree is. The demand matching degree can reflect whether the current energy supply meets the demand and provide a basis for subsequent equipment adjustments.

[0051] If the requirements match Less than the preset matching threshold , then adjust the construction site equipment according to the predicted power demand, including re-planning the operation schedule of non-critical equipment. The adjustment of equipment is based on its criticality value. ,in Indicates the device number. The higher the criticality value, the more important the device. For non-critical devices, the operation timing adjustment strategy is as follows: For non-critical equipment, the operating timing adjustment strategy is as follows: if the criticality value of the equipment is less than or equal to the preset criticality threshold, the operating time of the equipment will be delayed; if the criticality value of the equipment is greater than the preset criticality threshold, the current operating status of the equipment will be maintained.

[0052] By calculating the demand matching degree and evaluating the criticality of equipment, the operating sequence of non-critical equipment can be dynamically adjusted, energy distribution can be optimized, and the efficiency and stability of system operation can be improved. The calculation of demand matching degree provides a quantitative basis for energy management and equipment adjustment, and the introduction of criticality value ensures the priority operation of key equipment and guarantees the continuity and safety of construction site operations.

[0053] In step S16, a Kalman filter algorithm is used to perform denoising on the personnel positioning data to obtain accurate position data.

[0054] In a specific implementation, the use of a Kalman filter algorithm to perform denoising on the personnel positioning data to obtain accurate location data includes: Cleaning the personnel positioning data of outliers to obtain pre-processed positioning data; Using a Kalman filter algorithm to perform denoising on the pre-processed positioning data to obtain denoised positioning data; Extracting the personnel ID, coordinate value and movement trajectory according to the denoised positioning data, and calculating the movement speed according to the time series of the coordinate value; The personnel ID, the coordinate value, the moving track and the moving speed are used as precise location data; Among them, the outlier is determined by the following formula: The moving speed is calculated by the following formula: in, Indicated in Personnel location data at all times Axis coordinates, Indicates personnel location data The mean of the axis coordinates, Indicates personnel location data The standard deviation of the axis coordinates, Indicates the preset abnormal threshold. Indicated in Precise location at the moment Axis coordinates, Indicated in The exact location at the moment Axis coordinates, represents the pre-stored sampling time step, Indicates the movement speed.

[0055] Specifically, in step S16, the Kalman filter algorithm is used to perform denoising on the personnel positioning data to obtain accurate position data. The method achieves refined processing of personnel positioning data through outlier cleaning, Kalman filter denoising and movement speed calculation to obtain accurate position data such as personnel ID, coordinate value, movement trajectory and movement speed.

[0056] First, the personnel positioning data is cleaned of outliers to obtain pre-processed positioning data. The outlier judgment formula is: in, Indicated in Personnel location data at all times Axis coordinates, Indicates personnel location data The mean of the axis coordinates, Indicates personnel location data The standard deviation of the axis coordinates, Indicates the preset abnormal threshold. If this condition is met, the judgment The outliers are removed and the preprocessed positioning data is obtained after the outliers are cleaned.

[0057] The Kalman filter algorithm is used to denoise the pre-processed positioning data to obtain denoised positioning data. The Kalman filter algorithm includes two steps: prediction and update. Assume that the state vector of the personnel position is ,in Indicates location, Represents speed. The state prediction formula is: in, is the predicted state vector, is the state vector at the previous moment, is the state transfer matrix, is the prediction covariance matrix, is the covariance matrix of the previous moment, is the process noise covariance matrix. The state update formula is: in, is the Kalman gain matrix, is the observation matrix, is the observation noise covariance matrix, is the observed value, is the updated state vector, is the updated covariance matrix, is the unit matrix. Through iterative prediction and updating, the Kalman filter algorithm removes observation noise and outputs denoised positioning data.

[0058] The personnel ID, coordinate value and movement trajectory are extracted from the denoised positioning data, and the movement speed is calculated based on the time series of the coordinate value. The movement speed calculation formula is: in, Indicated in Precise location at the moment Axis coordinates, Indicated in The exact location at the moment Axis coordinates, represents the pre-stored sampling time step, Indicates the movement speed.

[0059] The personnel ID, the coordinate value, the moving track and the moving speed are used as accurate location data. The Kalman filter algorithm is used to remove noise and abnormal values ​​in the personnel positioning data to improve the accuracy and reliability of the location data.

[0060] In step S17, the working distance between the person and the equipment is calculated based on the precise location data and the image features. If the working distance exceeds a preset distance threshold, the standby energy-saving mode is triggered.

[0061] In a specific implementation, the calculating the operating distance between a person and a device according to the precise location data and the image features includes: The working distance between personnel and equipment is calculated by the following formula: in, Indicates the preset device location, Indicated in Moment The precise location of a person, Indicates The working distance of each person.

[0062] Specifically, first, through the preset device location and in Precise location at the moment , calculate the working distance between personnel and equipment The calculation formula for the working distance is: in, Indicates the preset device location, Indicated in Moment The precise location of a person, Indicates The formula is based on a two-dimensional plane coordinate system and uses the Euclidean distance to calculate the actual distance between a person and the equipment. The result is a non-negative real number.

[0063] Calculated working distance The distance threshold to the preset For comparison. Greater than , the standby energy-saving mode is triggered, and the device enters a low-power state; if Less than or equal to , then keep the normal operation of the equipment.

[0064] By calculating the distance between people and equipment in real time, the working state of the equipment can be adjusted dynamically to avoid wasting energy in unnecessary situations. The calculation of the working distance provides the system with accurate spatial information, and the introduction of the preset distance threshold ensures that the equipment automatically enters the energy-saving mode when people are away, thereby improving energy utilization efficiency, extending the service life of the equipment, and reducing energy waste.

[0065] In step S18, after the standby energy-saving mode is triggered, the energy consumption of the equipment is monitored. If the energy consumption of the equipment is lower than the preset energy consumption threshold and the position of the personnel has not changed, the standby state is maintained, otherwise the normal operation mode is restored.

[0066] First, after the standby energy-saving mode is triggered, the energy consumption value of the device is monitored in real time and compared with the preset energy consumption threshold. The device energy consumption value represents the power consumption of the device in the standby state, which is obtained through sensors or energy consumption monitoring systems. If the device energy consumption value is less than the energy consumption threshold, it means that the device is in a low power consumption state; if the device energy consumption value is greater than or equal to the energy consumption threshold, it means that the device has high energy consumption and may need to resume normal operation.

[0067] At the same time, real-time monitoring of personnel location information and the location of the person when the standby energy saving mode is triggered The judgment condition that the personnel position has not changed is: If the personnel position meets the above conditions, it means that the personnel position has not changed; otherwise, it means that the personnel position has changed.

[0068] By monitoring the energy consumption of equipment and the location of personnel in real time, the working mode of the equipment is dynamically managed. When the energy consumption of the equipment is low and the location of the personnel has not changed, the equipment will remain in standby mode to further reduce energy consumption; when the energy consumption of the equipment is high or the location of the personnel has changed, the normal operation mode will be restored in time to ensure the normal use and working efficiency of the equipment. This mechanism effectively balances the energy-saving needs and equipment function requirements and optimizes energy utilization efficiency.

[0069] Reference Figure 2 The second embodiment of the present invention provides a smart construction site energy-saving control system based on the Internet of Things, including: A data acquisition module is used to obtain real-time images of the construction site, the operating status of the energy storage equipment, and personnel positioning data, where the operating status includes the discharge power; An image feature analysis module, used to perform preliminary feature extraction on the real-time image using a pre-built dynamic network model to obtain image features; A scene recognition module, used to input the image features into a pre-trained scene recognition model to obtain a scene recognition result; A demand prediction module, used to extract energy demand characteristics from the scene recognition results and input them into a pre-built power prediction model to obtain predicted demand power; A demand analysis module, configured to calculate a demand matching degree according to the discharge power and the demand power, and if the demand matching degree is less than a preset matching degree threshold, adjust the operating timing of non-critical equipment according to the criticality value of the equipment; A data filtering module, used to perform denoising on the personnel positioning data using a Kalman filtering algorithm to obtain accurate position data; An operation distance calculation module, used to calculate the operation distance between the personnel and the equipment according to the precise location data and the image features, and trigger a standby energy-saving mode if the operation distance exceeds a preset distance threshold; The energy consumption monitoring module is used to monitor the energy consumption of the equipment after the standby energy-saving mode is triggered. If the energy consumption of the equipment is lower than the preset energy consumption threshold and the position of the personnel has not changed, the standby state is maintained, otherwise the normal operation mode is restored.

[0070] It should be noted that the smart construction site energy-saving control system based on the Internet of Things provided in an embodiment of the present invention is used to execute all the process steps of the smart construction site energy-saving control method based on the Internet of Things in the above-mentioned embodiment. The working principles and beneficial effects of the two correspond one to one, and therefore will not be repeated here.

[0071] The embodiment of the present invention further provides an electronic device. The electronic device includes: a processor, a memory, and a computer program stored in the memory and executable on the processor, such as an energy-saving control program for a smart construction site based on the Internet of Things. When the processor executes the computer program, the steps in the above-mentioned embodiments of the energy-saving control method for a smart construction site based on the Internet of Things are implemented, such as Figure 1 Alternatively, when the processor executes the computer program, the functions of the modules / units in the above-mentioned device embodiments are realized, such as a smart construction site energy-saving control module based on the Internet of Things.

[0072] Exemplarily, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to implement the present invention. The one or more modules / units may be a series of computer program instruction segments capable of implementing specific functions, which are used to describe the execution process of the computer program in the electronic device.

[0073] The electronic device may be a computing device such as a desktop computer, a notebook, a PDA, and a smart tablet. The electronic device may include, but is not limited to, a processor and a memory. Those skilled in the art will appreciate that the above components are merely examples of electronic devices and do not constitute a limitation on the electronic device. The electronic device may include more or fewer components than the above components, or may combine certain components, or different components. For example, the electronic device may also include input and output devices, network access devices, buses, etc.

[0074] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the electronic device, and uses various interfaces and lines to connect various parts of the entire electronic device.

[0075] The memory can be used to store the computer program and / or module, and the processor realizes various functions of the electronic device by running or executing the computer program and / or module stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0076] Wherein, if the module / unit integrated in the electronic device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0077] It should be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art may understand and implement it without paying any creative effort.

[0078] The specific embodiments described above further illustrate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A smart construction site energy-saving control method based on the Internet of Things, characterized in that: include: Obtain real-time images of the construction site, the operating status of the energy storage equipment, and personnel location data, where the operating status includes the discharge power; Using a pre-built dynamic network model to perform preliminary feature extraction on the real-time image to obtain image features; Inputting the image features into a pre-trained scene recognition model to obtain a scene recognition result; Extracting energy demand characteristics from the scene recognition results and inputting them into a pre-built power prediction model to obtain predicted demand power; Calculating a demand matching degree according to the discharge power and the demand power, and if the demand matching degree is less than a preset matching degree threshold, adjusting the operating timing of non-critical equipment according to the criticality value of the equipment; A Kalman filter algorithm is used to perform denoising on the personnel positioning data to obtain accurate position data; Calculating the working distance between the person and the equipment according to the precise location data and the image features, and triggering the standby energy-saving mode if the working distance exceeds a preset distance threshold; After the standby energy-saving mode is triggered, the energy consumption of the equipment is monitored. If the energy consumption of the equipment is lower than the preset energy consumption threshold and the position of the personnel has not changed, the standby state is maintained, otherwise the normal operating mode is restored.

2. The energy-saving control method for smart construction sites based on the Internet of Things according to claim 1 is characterized in that: The pre-built dynamic network model is used to perform preliminary feature extraction on the real-time image to obtain image features, including: Using a dynamic weight adjustment method to perform multi-scale feature extraction on the real-time image to obtain preliminary image features; Inputting the preliminary image features into a pre-built convolutional neural network model for feature optimization to obtain optimized features; Inputting the optimized features into a pre-built support vector machine classifier to perform image classification to obtain image features; Wherein, the dynamic network model includes a convolutional neural network model and a support vector machine classifier.

3. The energy-saving control method for smart construction sites based on the Internet of Things according to claim 1 is characterized in that: The step of inputting the image features into a pre-trained scene recognition model to obtain a scene recognition result includes: Inputting the image features into a preliminary scene recognition model to perform current scene recognition, and obtaining a preliminary scene recognition result and confidence level; Comparing the confidence level with a preset confidence threshold, and if the confidence level is greater than or equal to the confidence threshold, determining that the preliminary scene recognition result is qualified, and obtaining a scene recognition result; If the confidence level is less than the confidence threshold, the preliminary scene recognition result is judged to be unqualified, and the image features are input into the precise scene recognition model for precise scene recognition to obtain a precise scene recognition result, which is used as the final scene recognition result; The scene recognition model includes a preliminary scene recognition model and a precise scene recognition model.

4. The energy-saving control method for smart construction sites based on the Internet of Things according to claim 1 is characterized in that: The process of constructing the power prediction model includes: Perform data cleaning and feature extraction on the pre-stored historical energy demand characteristics to obtain the cleaned historical energy characteristics; Inputting the cleaned historical energy characteristics into the input layer of the initially constructed neural network model for training, and obtaining the predicted value output by the output layer of the neural network model; Substitute the predicted value and the pre-stored actual value into the loss function to calculate the loss value and obtain the loss value data; Calculate the gradient of the output layer output of the neural network model according to the loss value data, and pass the gradient forward layer by layer through the chain rule to calculate the gradient of each layer parameter to obtain gradient data; Update the parameters of each layer of the neural network model according to the gradient data and the preset learning rate; The parameters of each layer are updated repeatedly and iteratively until the training times of the neural network model are greater than the preset training times, or the loss value data of the neural network model is less than the preset loss threshold, the training is determined to be completed, and a power prediction model is obtained.

5. The energy-saving control method for smart construction sites based on the Internet of Things according to claim 1 is characterized in that: The Kalman filter algorithm is used to perform denoising on the personnel positioning data to obtain accurate position data, including: Cleaning the personnel positioning data of outliers to obtain pre-processed positioning data; Using a Kalman filter algorithm to perform denoising on the pre-processed positioning data to obtain denoised positioning data; Extracting the personnel ID, coordinate value and movement trajectory according to the denoised positioning data, and calculating the movement speed according to the time series of the coordinate value; The personnel ID, the coordinate value, the moving track and the moving speed are used as precise location data; Among them, the outliers are determined by the following formula: The moving speed is calculated by the following formula: in, Indicated in The moment Personnel location data Axis coordinates, Indicates personnel location data The mean of the axis coordinates, Indicates personnel location data The standard deviation of the axis coordinates, Indicates the preset abnormal threshold. Indicated in Moment The precise location of a person Axis coordinates, Indicated in Moment The precise location of a person Axis coordinates, represents the pre-stored sampling time step, Indicates the movement speed.

6. The energy-saving control method for smart construction sites based on the Internet of Things according to claim 1 is characterized in that: The calculating the operating distance between the personnel and the equipment according to the precise location data and the image features includes: The working distance between personnel and equipment is calculated by the following formula: in, Indicates the preset device location, Indicated in Moment The precise location of a person, Indicates The working distance of each person.

7. The energy-saving control method for smart construction sites based on the Internet of Things according to claim 1 is characterized in that: The calculating the demand matching degree according to the discharge power and the demand power, and if the demand matching degree is less than a preset matching degree threshold, adjusting the operating timing of non-critical equipment according to the criticality value of the equipment, includes: Calculating according to the discharge power and the required power to obtain a demand matching degree; If the demand matching degree is less than a preset matching degree threshold, the construction site equipment is adjusted according to the predicted demand power, including re-planning the operating timing of non-critical equipment; The demand matching degree is calculated by the following formula: in, Indicates the degree of demand matching. Indicates the required power, Indicates the discharge power.

8. A smart construction site energy-saving control system based on the Internet of Things, characterized in that: include: A data acquisition module is used to obtain real-time images of the construction site, the operating status of the energy storage equipment, and personnel positioning data, where the operating status includes the discharge power; An image feature analysis module, used to perform preliminary feature extraction on the real-time image using a pre-built dynamic network model to obtain image features; A scene recognition module, used to input the image features into a pre-trained scene recognition model to obtain a scene recognition result; A demand prediction module, used to extract energy demand characteristics from the scene recognition results and input them into a pre-built power prediction model to obtain predicted demand power; A demand analysis module, configured to calculate a demand matching degree according to the discharge power and the demand power, and if the demand matching degree is less than a preset matching degree threshold, adjust the operating timing of non-critical equipment according to the criticality value of the equipment; A data filtering module, used for performing denoising processing on the personnel positioning data by using a Kalman filtering algorithm to obtain accurate position data; An operation distance calculation module, used to calculate the operation distance between the personnel and the equipment according to the precise location data and the image features, and trigger a standby energy-saving mode if the operation distance exceeds a preset distance threshold; The energy consumption monitoring module is used to monitor the energy consumption of the equipment after the standby energy-saving mode is triggered. If the energy consumption of the equipment is lower than the preset energy consumption threshold and the position of the personnel has not changed, the standby state is maintained, otherwise the normal operation mode is restored.

9. An electronic device, characterized in that: It includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the energy-saving control method for a smart construction site based on the Internet of Things as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the energy-saving control method for a smart construction site based on the Internet of Things as described in any one of claims 1 to 7.

Images