Building engineering construction water consumption prediction method and system

Abstract

The invention relates to a building engineering construction water consumption prediction method and system, and the method comprises the steps: collecting second sample data of other construction sites for a scene with insufficient data volume; and then analyzing a target influence factor associated with the water consumption by using the first sample data of the current construction site, and then constructing a statistical model capable of reflecting a general water consumption rule of the current construction site. And performing confidence evaluation on the second sample data of other construction sites by using the statistical model, thereby reserving the sample data meeting the current construction site. And the BP neural network is trained in combination with the existing sample data of the current construction site, so that a prediction model capable of accurately predicting the construction water supply amount is obtained. According to the method, target influence factor analysis is carried out on different construction sites, and proper sample data is screened from other data sources, so that even if the sample data is insufficient, the prediction model capable of accurately predicting the water consumption can be trained based on the existing data, and the prediction result is more accurate.

Description

Technical Field

[0001] The present invention relates to the field of water consumption prediction, and specifically to a method and system for predicting construction water consumption in construction projects. Background Art

[0002] Scientific and effective water resource management is an important guarantee for social and economic development. Construction water consumption in construction projects is an important part of urban water use, especially in large cities in developing countries. The prediction of construction water demand in construction projects is one of the important contents of water resource management. Accurate and effective water resource prediction can effectively ensure urban water resource security.

[0003] However, the problem of predicting construction water consumption is relatively complex. On the one hand, almost all construction operations such as concrete construction and formwork construction at the construction site of a construction project require water, and have the characteristic of large water consumption. On the other hand, construction projects have complex uncertainties, and the water consumption in different construction stages shows great differences. Therefore, it is relatively difficult to accurately predict construction water consumption.

[0004] In most existing water demand prediction scenarios, neural models are used for prediction. However, for the construction site, there are not many data samples for training at the beginning. Applying the historical data of other construction sites may have different influencing factors, and the established model cannot accurately predict water consumption. Summary of the Invention

[0005] In view of this, the purpose of the present invention is to provide a method and system for predicting construction water consumption in construction projects to solve the above technical problems.

[0006] To achieve the above purpose, the present invention adopts the following technical solutions:

[0007] A method for predicting construction water consumption in a construction project of the present invention includes the steps of:

[0008] Obtain the first sample data of the current construction site and the second sample data of other construction sites, and obtain the construction plan and weather forecast of the construction site. Among them, both the first sample data and the second sample data include the water consumption at multiple time points and the values of various influencing factors at multiple time points;

[0009] Based on the first sample data, perform a correlation analysis on the water consumption and various influencing factors of the current construction site to obtain the target influencing factors affecting the construction site;

[0010] Construct a statistical model based on the first sample data and the target impact factor; and use the statistical model to evaluate the confidence level of the second sample data to obtain the confidence level of the second sample data, where the statistical model includes the value relationship between the water consumption and the target impact factor;

[0011] Train a BP neural network based on the first sample data and the second sample data with a confidence level higher than the preset confidence level threshold to obtain a prediction model;

[0012] Extract the predicted value of the target impact factor from the construction plan and weather forecast of the construction site, and predict the water consumption of the current construction site based on the prediction model and the predicted value of the target impact factor.

[0013] In an embodiment of the present application, perform a correlation analysis on the water consumption of the current construction site and various impact factors based on the first sample data to obtain the target impact factors affecting the construction site, including:

[0014] Normalize the water consumption at multiple time points in the first sample data to obtain a water consumption time series {X1, X2,..., X t}; and normalize the values of various impact factors at multiple time points in the first sample data to obtain a time series of various impact factors {n1, n2,..., n t}, where X1 is the normalized water consumption at the first time point, X2 is the normalized water consumption at the second time point, X t is the normalized water consumption at the t-th time point, n1 is the normalized value of the impact factor at one time point, n2 is the normalized value of the impact factor at the second time point, and n t is the normalized value of the impact factor at the t-th time point;

[0015] Align the water consumption time series and the time series of various impact factors in the time dimension; and calculate the difference between the aligned water volume time series and the time series to obtain a difference series {|X1 - n1|, |X2 - n2|,..., |X t - n t |};

[0016] Calculate the correlation degree r of the impact factor based on the difference series. The mathematical expression of the correlation degree R is:

[0017]

[0018] In the formula, N is the number of multiple time points;

[0019] Use the impact factor with a correlation degree greater than the preset correlation degree threshold as the target impact factor affecting the construction site.

[0020] In one embodiment of the present application, constructing a statistical model based on the first sample data and the target influencing factor includes:

[0021] Dividing the first sample data based on multiple water consumption ranges to obtain multiple sample data units;

[0022] Performing density clustering on the values of the target influencing factor in multiple sample data units at multiple time points to obtain multiple clusters;

[0023] Regarding the clusters with the number of data points within the cluster greater than a preset threshold as target clusters, and calculating the average value of the target influencing factor within the target clusters;

[0024] Eliminating the data points within the target clusters with a deviation value from the average value greater than a preset deviation rate threshold to obtain updated target clusters;

[0025] Calculating the average value A and standard deviation σ of the target influencing factor within the updated target clusters;

[0026] Based on the average value A and standard deviation σ of the target influencing factor within the target clusters, constructing a first reference value range (A - α1σ, A + α1σ), a second reference value range {(A - α2σ, A - α1σ), (A + α1σ, A + α2σ)}, and a third reference value range {(A - α3σ, A - α2σ), (A + α2σ, A + α3σ)} of the target influencing factor under the current water consumption range, where α1, α2, and α3 are all range adjustment parameters and satisfy α1 < α2 < α3;

[0027] Constructing a statistical model based on multiple first reference value ranges, second reference value ranges, and third reference value ranges of the target influencing factor corresponding to multiple water consumption ranges.

[0028] In one embodiment of the present application, evaluating the confidence level of the second sample data using the statistical model to obtain the confidence level of the second sample data, including:

[0029] Extracting the water consumption and the values of the influencing factor at each time point in the second sample data to obtain multiple samples to be evaluated;

[0030] Determining the water consumption range, the first reference value range, the second reference value range, and the third reference value range in the statistical model based on the water consumption of the samples to be evaluated;

[0031] Comparing the values of the target influencing factor in the samples to be evaluated with the corresponding first reference value range, second reference value range, and third reference value range, and obtaining the confidence level C of the samples to be evaluated based on the comparison results, where the mathematical expression of the confidence level C is:

[0032]

[0033] Wherein, N1 is the number of target influencing factors in the to-be-evaluated sample that fall within the first reference value range, C1 is the confidence level when falling within the first reference value range, N2 is the number of target influencing factors in the to-be-evaluated sample that fall within the second reference value range, C2 is the confidence level when falling within the second reference value range, N3 is the number of target influencing factors in the to-be-evaluated sample that fall within the third reference value range, and C3 is the confidence level when falling within the third reference value range.

[0034] In an embodiment of the present application, based on the first sample data and the second sample data with a confidence level higher than the preset confidence threshold, the BP neural network is trained to obtain a prediction model, including:

[0035] Taking the water consumption at multiple time points in the first sample data and the second sample data with a confidence level higher than the preset confidence threshold as labels, and taking the values of the target influencing factors at multiple time points in the first sample data and the second sample data with a confidence level higher than the preset confidence threshold as training data to construct a training dataset;

[0036] Based on the training dataset, the BP neural network is trained in combination with the particle swarm optimization algorithm, and after the loss function is minimized, a prediction model is obtained.

[0037] In an embodiment of the present application, the mathematical expression of the loss function F(ψ,ω,θ,r) is:

[0038]

[0039] Wherein, ψ is the error value in the BP neural network, ω, θ, and r are the input layer, hidden layer, and output layer of the neural network respectively; N1 and M are the number of weight and threshold nodes respectively; y t (s) is the expected output of the neural network; is the actual output; t is a node for optimizing the connection weight, and s is a node for optimizing the threshold.

[0040] In an embodiment of the present application, based on the prediction model and the predicted values of the target influencing factors, the water consumption at the current construction site is predicted, including:

[0041] Inputting the predicted values of the target influencing factors into the prediction model to obtain the water consumption predicted values corresponding to the predicted value time of the target influencing factors.

[0042] The present application also provides a construction engineering construction water consumption prediction system, including:

[0043] An acquisition module, configured to acquire first sample data of the current construction site and second sample data of other construction sites, and acquire the construction plan and weather forecast of the building construction site, wherein both the first sample data and the second sample data include the water consumption at multiple time points and the values of multiple influencing factors at multiple time points;

[0044] A correlation analysis module, configured to perform a correlation analysis on the water consumption and multiple influencing factors of the current construction site based on the first sample data to obtain target influencing factors that affect the building construction site;

[0045] A sample screening module, configured to construct a statistical model based on the first sample data and the target influencing factors; and use the statistical model to evaluate the confidence level of the second sample data to obtain the confidence level of the second sample data, wherein the statistical model includes the value relationship between the water consumption and the target influencing factors;

[0046] A model training module, configured to train a BP neural network based on the first sample data and the second sample data with a confidence level higher than a preset confidence threshold to obtain a prediction model;

[0047] A prediction module, configured to extract the predicted values of the target influencing factors from the construction plan and weather forecast of the building construction site, and predict the water consumption of the current construction site based on the prediction model and the predicted values of the target influencing factors.

[0048] This application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the method described in any one of the above is implemented.

[0049] This application also provides an electronic terminal, including: a processor and a memory;

[0050] The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory so that the terminal executes the method described in any one of the above.

[0051] The beneficial effects of the present invention are as follows: A method and system for predicting construction water consumption in a construction project according to the present invention, compared with the traditional neural network training method, for the scenario of insufficient data volume, collect second sample data from other construction sites. Then, analyze the target influencing factors related to water consumption using the first sample data of the current construction site, and then construct a statistical model that can reflect the general water consumption pattern of the current construction site. Use the statistical model to evaluate the confidence of the second sample data of other construction sites, so as to retain the sample data that meets the current construction site. Combine the existing sample data of the current construction site to train the BP neural network, so as to obtain a prediction model that can accurately predict the construction water supply. The present application analyzes the target influencing factors for different construction sites and screens out appropriate sample data from other data sources, so that even when the sample data is insufficient, a prediction model that can accurately predict water consumption can be trained based on the existing data, and the prediction result is more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The present invention will be further described below with reference to the drawings and embodiments:

[0053] Figure 1 is a flowchart of a method for predicting construction water consumption shown in an embodiment of the present application;

[0054] Figure 2 is a schematic diagram of a convergence curve in an embodiment of the present application;

[0055] Figure 3 is a structural diagram of a system for predicting construction water consumption shown in an embodiment of the present application;

[0056] Figure 4 shows a schematic diagram of the structure of a computer system of an electronic device suitable for implementing the embodiments of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0057] The following specific examples illustrate the embodiments of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0058] It should be noted that the diagrams provided in the following embodiments only illustrate the basic concept of the present invention schematically. Therefore, only the layers related to the present invention are shown in the diagrams, rather than being drawn according to the number, shape, and size ratio of the layers in actual implementation. The types, quantities, and arrangements of the layers in actual implementation may be arbitrarily changed, and the layer layout pattern may also be more complex.

[0059] In the following description, a large number of details are explored to provide a more thorough explanation of the embodiments of the present invention. However, it is obvious to those skilled in the art that the embodiments of the present invention can be implemented without these specific details.

[0060] In recent years, the emergence of artificial intelligence algorithms such as artificial neural networks has provided new ideas for solving the real-time results of the complex model system of water demand prediction in construction engineering. Artificial neural network algorithms have good adaptive learning ability and non-linear mapping ability, can fully explore the potential laws of input data, and show great superiority in the research and analysis of complex systems with multi-factor coupling. When applying artificial neural networks to water demand prediction in construction engineering, the model solves the non-linear problems in water demand prediction by simulating the structural characteristics and action mechanisms of biological neurons, and uses limited data measured on-site instead of a large amount of complete statistical data, and adopts a data-driven method to predict water demand. At present, relevant scholars have carried out research on water demand prediction and achieved rich research results.

[0061] However, there are many problems in applying artificial neural networks to specific construction sites. For example, the influencing factors of water consumption vary among different construction sites, and these influencing factors will change according to factors such as the type, scale, geographical location, climate conditions, and specific construction stages of construction projects. This results in the inability to accumulate sufficient data samples at the beginning of construction and also unable to directly use data samples from other construction sites for training.

[0062] To solve the above problems, the present application proposes the following method and system for predicting water consumption in construction engineering.

[0063] Figure 1 is a flowchart of a method for predicting water consumption in construction engineering shown in an embodiment of the present application. As Figure 1 shown: A method for predicting water consumption in construction engineering in this embodiment may include steps S110 to S150:

[0064] S110, obtain the first sample data of the current construction site and the second sample data of other construction sites, and obtain the construction plan and weather forecast of the construction site. Among them, both the first sample data and the second sample data include the water consumption at multiple time points and the values of various influencing factors at multiple time points;

[0065] The first sample data is recorded from the current construction site, including multiple factors affecting the water consumption during the construction period of the building project and the water consumption. The influencing factors include the number of workers, the amount of concrete poured on the current day, the highest temperature on the current day, the weather on the current day, the amount of wood used on the current day, the amount of steel used on the current day, etc.

[0066] The second sample data is the historical sample data of construction sites of the same type, and the influencing factors are the same as those of the first sample data.

[0067] In addition, in order to predict the water consumption at future time points, it is also necessary to obtain the values of the influencing factors at future time points from the construction plan and the weather forecast.

[0068] S120, perform a correlation analysis on the water consumption and various influencing factors of the current construction site based on the first sample data to obtain the target influencing factors affecting the construction site;

[0069] Before model training, it is necessary to clarify which influencing factors are the main influencing factors affecting the water consumption of the current construction site. Using the sample data of the target influencing factors for model training can avoid phenomena such as overfitting, reduction of model performance, and increase of calculation cost.

[0070] In this application, the following method is used to determine which influencing factors are the target influencing factors affecting the water consumption of the construction site, specifically including:

[0071] S121, perform normalization processing on the water consumption at multiple time points in the first sample data to obtain the water consumption time series {X1, X2,..., X t}; and perform normalization processing on the values of various influencing factors at multiple time points in the first sample data to obtain the time series of various influencing factors {n1, n2,..., n t}, where X1 is the normalized water consumption at the first time point, X2 is the normalized water consumption at the second time point, X t is the normalized water consumption at the t-th time point, n1 is the normalized value of the influencing factor at one time point, n2 is the normalized value of the influencing factor at the second time point, n t is the normalized value of the influencing factor at the t-th time point;

[0072] Among them, the mathematical expression of the normalization processing is:

[0073]

[0074] In the formula, X min is the minimum water consumption in the first sample data, X max is the maximum water consumption in the first sample data; nmin is the minimum value of the influencing factors in the first sample data, n max is the maximum value of the influencing factors in the first sample data.

[0075] Normalization can remove the dimensions of different influencing factors and water consumption, making the values of different influencing factors fall within [0, 1], which is convenient for subsequent comparison and correlation analysis.

[0076] S122. Align the water consumption time series and the time series of multiple influencing factors in the time dimension; and take the difference between the aligned water volume time series and the time series to obtain a difference sequence {|X1 - n1|, |X2 - n2|,..., |X t - n t |};

[0077] This application uses the difference between the normalized water consumption sequence and the influencing factor value sequence, and uses the difference sequence to reflect their differences.

[0078] S123. Calculate the correlation degree R of the influencing factors based on the difference sequence. The mathematical expression of the correlation degree R is:

[0079]

[0080] In the formula, N is the number of multiple time points;

[0081] Finally, take the average value of the absolute value of each element value in the difference sequence to obtain and take the difference with 1 to obtain the correlation degree.

[0082] S124. Take the influencing factors with a correlation degree greater than the preset correlation degree threshold as the target influencing factors affecting the construction site. Specifically, the following table shows the values of each influencing factor and the water consumption at multiple time points of a construction site;

[0083] Table 1 Numerical values of each factor and actual water consumption

[0084]

[0085] After the above correlation analysis and calculation, the results shown in Table 2 are obtained:

[0086] Table 2 Correlation coefficients of influencing factors

[0087]

[0088] As can be seen from Table 2, the correlation degrees of the influencing factors are in the following order: the concrete pouring volume on the same day (r2) > the weather on the same day (r4) > the number of workers (r1) > the wood usage on the same day (r5) > the highest temperature on the same day (r3) > the steel usage on the same day (r6). This ranking can be explained by the construction operation content and characteristics of the building project. Concrete pouring is a typical wet operation that requires a large amount of water; the weather is another important factor affecting construction. When it rains, most operations at the construction site will stop, and the construction water consumption will decrease significantly; the more workers there are, the more domestic water for construction will be; the construction wood needs to be watered and moistened to ensure that its moisture content is near the optimal moisture content, which also requires a large amount of water.

[0089] When the correlation degree is less than 0.6, it is considered that the two sequences are irrelevant. If the correlation degree is greater than 0.8, the two sequences have a good correlation; when it is between 0.6 and 0.8, the correlation degree between the two is good. There are 4 factors in Table 2 with a correlation degree greater than 0.8, namely the concrete pouring volume on the same day, the weather on the same day, the number of workers, and the wood usage on the same day. That is, the key factors affecting the construction water consumption of the building project are the concrete pouring volume on the same day, the weather on the same day, the number of workers, and the wood usage on the same day.

[0090] S130. Based on the first sample data and the target influencing factor, construct a statistical model; and use the statistical model to evaluate the confidence level of the second sample data to obtain the confidence level of the second sample data, where the statistical model includes the value relationship between the water consumption and the target influencing factor.

[0091] Since this application intends to use the sample data of other construction sites, it is necessary to ensure that the sample data of other construction sites meet the basic water usage rules of the current construction site.

[0092] The solution idea is to first construct a statistical model that reflects the basic water usage rules of the current construction site. Use the statistical model to screen the sample data of other construction sites to obtain the sample data that can meet the basic water usage rules of the current construction site, and then use these real sample data that meet the basic water usage rules of the current construction site to train the neural network, so as to obtain a real and accurate prediction model.

[0093] Based on the above idea, the construction method of the statistical model in this application includes:

[0094] S1301. Divide the first sample data based on multiple water consumption ranges to obtain multiple sample data units.

[0095] In this application, the average value and variance of all water consumption can be calculated first, a water consumption range of 3 times the standard deviation is constructed, and then the water consumption range is reasonably divided to obtain the corresponding multiple sample data units.

[0096] By dividing the water consumption range, the sample data of different water consumption ranges can be separated, and in the subsequent analysis process, the values ​​of the target impact factors corresponding to different water consumption levels can be obtained. And these discrete values ​​are used to analyze the distribution law to obtain the value range of the target impact factor associated with the current water consumption level.

[0097] S1302, performing density clustering on the values ​​of the target influencing factors in the multiple sample data units at multiple time points to obtain multiple clusters;

[0098] The present application utilizes density-based clustering to divide the values ​​of target impact factors in multiple sample data units at multiple time points into multiple clusters, thereby clustering data at multiple time points with similar values ​​together.

[0099] In this application, density clustering needs to set the maximum distance within the cluster to limit the dispersion of data points within the cluster. The multiple clusters obtained can reflect the multiple value aggregation ranges of the target impact factor under the current water consumption level.

[0100] S1303, taking a cluster whose number of data points is greater than a preset threshold as a target cluster, and calculating the average value of the target impact factor in the target cluster;

[0101] In this embodiment, if the number of data points in a cluster reaches a certain value, it means that the data points in the cluster are not gathered together by chance, but have corresponding associations. It shows that when the value of the target influencing factor is within the cluster, the water consumption of the current construction site will fall into the corresponding water consumption level.

[0102] Therefore, the present application takes the cluster whose number of data points is greater than a preset threshold as the target cluster.

[0103] S1304, eliminating data points in the target cluster whose deviation from the average value is greater than a preset deviation rate threshold, to obtain an updated target cluster;

[0104] In order to further improve the aggregation of data points within the cluster and obtain a value range that better satisfies regular characteristics, the present application also uses the average value to further screen the data points within the cluster to obtain an updated target cluster.

[0105] S1305, calculating the average value A and standard deviation σ of the target impact factor in the updated target cluster;

[0106] S1306. Construct a first reference value range (A - α1σ, A + α1σ), a second reference value range {(A - α2σ, A - α1σ), (A + α1σ, A + α2σ)}, and a third reference value range {(A - α3σ, A - α2σ), (A + α2σ, A + α3σ)} of the target influencing factor within the current water usage range based on the average value A and the standard deviation σ of the target influencing factor within the target cluster, where α1, α2, and α3 are all range adjustment parameters and satisfy α1 < α2 < α3;

[0107] Finally, construct multiple reference value ranges using the average value A and the standard deviation σ. In this application, three reference value ranges are constructed, namely (A - σ, A + σ), {(A - 2σ, A - σ), (A + σ, A + 2σ)}, and {(A - 3σ, A - 2σ), (A + 2σ, A + 3σ)}. These three ranges respectively correspond to approximately 68.27% of the data points, approximately 95.45% of the data points, and approximately 99.73% of the data points.

[0108] When it falls within the first reference value range (A - α1σ, A + α1σ), it indicates that the value pattern within the target cluster can be satisfied. While falling within the second reference value range {(A - α2σ, A - α1σ), (A + α1σ, A + α2σ)} and the third reference value range {(A - α3σ, A - α2σ), (A + α2σ, A + α3σ)} respectively indicates that the value can barely satisfy the value pattern within the target cluster.

[0109] Not falling within the above three value ranges can indicate data anomalies and inability to satisfy the value pattern within the cluster.

[0110] S1307. Construct a statistical model based on multiple first reference value ranges, second reference value ranges, and third reference value ranges of the target influencing factor corresponding to multiple water usage ranges.

[0111] Based on the constructed statistical model above, this application uses the following method to screen the second sample data, which specifically includes:

[0112] S1311. Extract the water usage amount and the values of the influencing factors at each time point in the second sample data to obtain multiple samples to be evaluated;

[0113] The sample to be evaluated contains the water usage amount at one time point and the values of multiple influencing factors.

[0114] S1312. Determine the water usage range, the first reference value range, the second reference value range, and the third reference value range in the statistical model based on the water usage amount of the sample to be evaluated;

[0115] Based on the water consumption in the sample to be evaluated, the corresponding water consumption range can be queried from the statistical model, and the water consumption in the sample to be evaluated falls within the corresponding water consumption range. Furthermore, the corresponding first reference value range, second reference value range, and third reference value range can be obtained.

[0116] S1313. Compare the value of the target influencing factor in the sample to be evaluated with the corresponding first reference value range, second reference value range, and third reference value range, and obtain the confidence level C of the sample to be evaluated based on the comparison result. Wherein, the mathematical expression of the confidence level C is:

[0117]

[0118] In the formula, N1 is the number of target influencing factors in the sample to be evaluated that fall within the first reference value range, C1 is the confidence level when falling within the first reference value range, N2 is the number of target influencing factors in the sample to be evaluated that fall within the second reference value range, C2 is the confidence level when falling within the second reference value range, N3 is the number of target influencing factors in the sample to be evaluated that fall within the third reference value range, and C3 is the confidence level when falling within the third reference value range.

[0119] Finally, perform confidence level evaluation based on the number of target influencing factors in the sample to be evaluated that fall within the three reference value ranges. The closer the range into which it falls is to the central position, the greater the confidence level. The final confidence level C can then evaluate the degree to which the water consumption in the sample to be evaluated and the values of multiple target influencing factors satisfy the basic water consumption law of the current construction site. The higher the degree, the more it can satisfy the basic water consumption law of the current construction site.

[0120] Therefore, finally, use the confidence level and the confidence level threshold to screen the second data sample to obtain the second data sample that satisfies the basic water consumption law of the current construction site.

[0121] The above-screened data sample is real data and basically satisfies the basic water consumption law of the current construction site. Therefore, real and reliable training data can be obtained.

[0122] S140. Train the BP neural network based on the first sample data and the second sample data with a confidence level higher than the preset confidence level threshold to obtain a prediction model;

[0123] Finally, use the above real and reliable training data to train the BP neural network, specifically including:

[0124] S141. Use the water consumption at multiple time points in the first sample data and the second sample data with a confidence level higher than the preset confidence threshold as labels, and use the values of the target impact factors at multiple time points in the first sample data and the second sample data with a confidence level higher than the preset confidence threshold as training data to construct a training dataset.

[0125] S142. Train the BP neural network based on the training dataset combined with the particle swarm optimization algorithm. After minimizing the loss function, obtain the prediction model.

[0126] The BP neural network is an artificial neural network model with self-learning and self-adaptive capabilities, consisting of two parts: the forward propagation of input data and the backward propagation of error values. The standard neural network topology structure includes input layer nodes, hidden layer nodes, and output layer nodes. The nodes of each layer are interconnected, and the nodes of the same layer do not interact with each other. This algorithm takes n samples X = (x1, x2,..., x n ) of the research object as the input layer nodes of the neural network, and the expected results Y = (y1, y2,..., y n ) as the corresponding output nodes. Through the corresponding weights and thresholds for calculation, comparing the predicted results with the actual results can obtain the error value. The fitness function is the standard for measuring whether the error value meets the requirements. For the calculation results that do not meet the requirements, the network will use the gradient descent method in the weight vector space for error backpropagation. Among them, the correction amount of each step of the weights in the hidden layer and the output layer, through repeated iteration, makes the error reach the expected value, and completes the establishment of the BP neural network calculation model.

[0127] The mathematical expression of the loss function F(ψ, ω, θ, r) of the BP neural network is:

[0128]

[0129] In the formula, ψ is the error value in the BP neural network, ω, θ, and r are the input layer, hidden layer, and output layer of the neural network respectively; N1 and M are the number of weight and threshold nodes respectively; y t (s) is the expected output of the neural network; is the actual output; t is a node whose connection weight needs to be optimized, and s is a node whose threshold needs to be optimized.

[0130] The BP neural network model can effectively solve the prediction of water consumption in construction engineering by discovering internal laws through the classification of a large amount of complex data. However, the traditional BP neural network uses the gradient descent method of the error function. Due to the unknowability of the learning rate and the randomness of the weights and thresholds, the algorithm stops calculating when it finds the local optimal solution, which is not conducive to establishing a correct network model. Considering that PSO has the characteristics of fast convergence speed and easy implementation, using PSO to optimize the initial weights and thresholds of the BP neural network can achieve higher accuracy.

[0131] The basic concept of the Particle Swarm Optimization (PSO) algorithm originates from the study of the foraging behavior of bird flocks. Inspired by this biological population behavior, the PSO algorithm is used to solve optimization problems. In PSO, each particle represents a solution to the problem and corresponds to a fitness value. The particle velocity determines the distance and direction of the particle's movement and is dynamically adjusted through the movement of itself and other particles, thus realizing the optimization process of the individual in the solvable space.

[0132] The process of using the PSO algorithm first takes the error between the capacity output obtained after the forward learning of the BP neural network and the expected capacity output, and initializes it through the PSO algorithm to find the individual extreme value and the global extreme value, that is, to find the weights and thresholds in the BP neural network; then updates the velocity and position, and updates the original individual extreme value and global extreme value after calculating the fitness; finally, sends the obtained optimal neural network weights and thresholds into the BP neural network for verification.

[0133] Suppose the particle swarm population X = (X1, X2, …, X n ) consists of n particles. Usually, the dimension of the particle is Q, so that there are n particles in the population. Each particle is Q-dimensional, and the population composed of n particles searches Q dimensions (that is, the dimension of each particle). Each particle is represented as X i = (X i1 , X i2 , …, X iQ ), which represents the position of the i-th particle in the Q-dimensional search space and also represents a potential solution to the problem. According to the objective function, the fitness value [ref] corresponding to each particle position X i can be calculated. The velocity corresponding to each particle can be expressed as V = (V i1 , V i2 , …, V iQ ). Each particle should consider two factors when searching:

[0134] 1) The historical optimal value P i that it has searched, P i = (P i1 , P i2 , …, PiQ ), where \(i = 1, 2, \ldots, n\)

[0135] 2) The optimal value \(P\) found by all particles g , \(P\) g = (\(P\) g1 , \(P\) g2 , \ldots, \(P\) gQ ). It is worth mentioning that there is only one \(P\) here. g

[0136] During each iteration, the particle updates its own velocity and position through its individual extreme value and the global extreme value. The updated formulas for the position and velocity optimized by the particle swarm algorithm are as follows:

[0137]

[0138] In the formula: \(\omega\) is the inertia weight; \(d = 1, 2, \ldots, D\); \(i = 1, 2, \ldots, n\); \(k\) is the current iteration number; \(v\) id is the velocity of the particle; \(c1\) is the particle weight coefficient for tracking its own historical optimal value, which represents the particle's own cognition and is called the acceleration factor; \(c2\) is the particle weight coefficient for tracking the group optimal value, which represents the particle's cognition of the knowledge of the entire group and is called the acceleration factor; \(r\) is a random number uniformly distributed in the interval \([0, 1]\).

[0139] Through the analysis of the data, the historical data and several factors that have the greatest impact on the water demand in construction engineering are used as input quantities and input into the neural network. After the neurons in each layer act on the influencing factors, an output quantity is generated. The weights and thresholds of the neural network are optimized by the particle swarm to obtain the fitness value, and the individual with the optimal fitness is found. Then, with the output error as the objective function, the error is corrected until the requirements are met. After training, the neural network can be used for prediction.

[0140] During the training of the training data in the training dataset, the difference in the numerical dimensions is likely to slow down the convergence speed of the algorithm, affecting the model accuracy. The calculation normalization formula normalizes the training data to \([-1, 1]\).

[0141] The mathematical expression of the normalization formula here is:

[0142]

[0143] \(x\) i is the value of the water consumption or the target influencing factor in the \(i\)-th frame of training data, \(x\) min is the minimum water consumption or the minimum value of the target influencing factor in the training data, \(x\) max is the maximum water consumption or the maximum value of the target influencing factor in the training data.

[0144] ​Four main influencing factors of the water demand in the construction area of a construction project are obtained through the correlation analysis method described above. The number of input nodes is m = 4, and n = 2m + 1 = 9, that is, the number of hidden layer nodes is n = 9. The structure of the BP neural network can be obtained as 4-9-1.

[0145] In the process of using the BP neural network to form a model, the available data should be divided into two groups. These two sets represent the training and test sets. The data in the training set is used for training, while the data in the test set is used to check the network. A large number of authors choose data ratios of 90% to 10%, 80% to 20%, or 70% to 30%.

[0146] In this study, the training set is the data of a certain engineering project from May 1st to June 20th, 2018, for a total of 51 days; the test set is the data of a certain engineering project from June 21st to June 30th, 2018, for a total of 10 days. The ratio of the training set data to the test set data is: 83.61%: 16 / 39%.

[0147] To obtain better prediction results, the optimal parameters of the BP neural network and the genetic algorithm are set. The parameters of the BP neural network include: the number of training times is 1000 times, the learning rate is 0.1, and the training target is 0.001; the calculation parameters of the PSO include: the number of iterations is 1000 times, the population size is 50, the local learning factor c1 = 1.49445; the global learning factor c2 = 1.49445; the maximum error for terminating the iteration is 0.00001.

[0148] Figure 2 This is a schematic diagram of the convergence curve in an embodiment of this application. After calculation, the convergence curve is as Figure 2 shown. In the case analysis, when the number of iterations reaches about 500, the requirements are met. In this study, the population size and the maximum number of iterations are set to be relatively large to ensure that the model can calculate more complex problems. Figure 2 The convergence curve of 1000 iterations is shown.

[0149] To verify the accuracy of this algorithm, error analysis is introduced in this paper. The relative error value is calculated based on the predicted value and the actual value. The formula is as follows:

[0150]

[0151] In the formula: E is the relative error; γ p is the predicted value; γ a is the true value.

[0152] Table 3 is the data table of the prediction results of the prediction model in this application. As shown in Table 3, compared with the actual water consumption, the errors calculated by the method in this paper are all less than 5%, and the average error is only 2.47%. However, the error calculated by the BP model is relatively large, with a maximum of 46.56% and an average error of 22.39%. This proves the effectiveness and advancement of this method in predicting the water demand for construction engineering construction.

[0153] Table 3. Data Table of Prediction Results of Prediction Model

[0154]

[0155] S150, extract the predicted values of the target influencing factors from the construction plan and weather forecast of the construction site. Based on the prediction model and the predicted values of the target influencing factors, predict the water consumption of the current construction site.

[0156] After obtaining the prediction model, this application extracts the predicted values of the target influencing factors (the amount of concrete poured on the current day, the weather on the current day, the number of workers, the amount of wood used on the current day) from the construction plan and weather forecast, and normalizes them. The normalized predicted values of the target influencing factors are used as the input layer and input into the prediction model to obtain the predicted water consumption values corresponding to the predicted values of the target influencing factors at different times.

[0157] Finally, a case analysis was carried out on the construction water consumption of a certain engineering project with insufficient training data volume. Compared with the actual water consumption, the errors calculated by the method in this paper are all less than 5%, and the average error is only 2.47%. A relatively accurate prediction model can be constructed based on the data analyzed by association and law.

[0158] A method for predicting construction water consumption in a building project according to the present invention. Compared with the traditional neural network training method, in this application, for the scenario of insufficient data volume, the second sample data of other construction sites is collected. Then, the target influencing factors related to the water consumption are analyzed using the first sample data of the current construction site, and a statistical model that can reflect the general water consumption pattern of the current construction site is constructed. The statistical model is used to evaluate the confidence of the second sample data of other construction sites, so as to retain the sample data that meets the current construction site. The BP neural network is trained by combining the existing sample data of the current construction site, so as to obtain a prediction model that can accurately predict the construction water supply. This application analyzes the target influencing factors for different construction sites and screens out appropriate sample data from other data sources, so that even when the sample data is insufficient, a prediction model that can accurately predict the water consumption can be trained based on the existing data, and the prediction result is more accurate.

[0159] As Figure 3 shown, this application also provides a construction water consumption prediction system for building projects, including:

[0160] An acquisition module, configured to acquire first sample data of the current construction site and second sample data of other construction sites, and acquire the construction plan and weather forecast of the building construction site. Wherein, both the first sample data and the second sample data include the water consumption at multiple time points and the values of multiple influencing factors at multiple time points;

[0161] An association analysis module, configured to perform an association analysis on the water consumption and multiple influencing factors of the current construction site based on the first sample data to obtain target influencing factors that affect the building construction site;

[0162] A sample screening module, configured to construct a statistical model based on the first sample data and the target influencing factors; and use the statistical model to evaluate the confidence level of the second sample data to obtain the confidence level of the second sample data. Wherein, the statistical model includes the value relationship between the water consumption and the target influencing factor;

[0163] A model training module, configured to train a BP neural network based on the first sample data and the second sample data with a confidence level higher than a preset confidence threshold to obtain a prediction model;

[0164] A prediction module, configured to extract the predicted values of the target influencing factors from the construction plan and weather forecast of the building construction site, and predict the water consumption of the current construction site based on the prediction model and the predicted values of the target influencing factors.

[0165] A construction project construction water consumption prediction system of the present invention. Compared with the traditional neural network training method, for the scenario of insufficient data volume in this application, the second sample data of other construction sites is collected. Then, the target influencing factors related to the water consumption are analyzed using the first sample data of the current construction site, and then a statistical model that can reflect the general water consumption law of the current construction site is constructed. The statistical model is used to evaluate the confidence level of the second sample data of other construction sites, so as to retain the sample data that meets the current construction site. The BP neural network is trained in combination with the existing sample data of the current construction site, so as to obtain a prediction model that can accurately predict the construction water supply. This application analyzes the target influencing factors for different construction sites, and screens out suitable sample data from other data sources, so that even when the sample data is insufficient, a prediction model that can accurately predict the water consumption can be trained based on the existing data, and the prediction result is more accurate.

[0166] Figure 4 The structural schematic diagram of the computer system of the electronic device suitable for implementing the embodiments of the present application is shown. It should be noted that, Figure 4 The computer system of the shown electronic device is only an example, and should not bring any limitation to the functions and usage scope of the embodiments of the present application.

[0167] As Figure 4 shown, the computer system includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 402 or a program loaded from a storage section 408 into a random access memory (RAM) 403, such as executing the method in the above embodiments. In the RAM 403, various programs and data required for system operations are also stored. The CPU 401, ROM 402, and RAM 403 are connected to each other via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0168] The following components are connected to the I / O interface 405: an input section 406 including a keyboard, a mouse, etc.; an output section 407 including, for example, a cathode ray tube (CRT), a liquid crystal display (LCD), etc. and a speaker; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to the I / O interface 405 as required. A removable medium 411, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 410 as required so that a computer program read from it can be installed into the storage section 408 as required.

[0169] Specifically, according to the embodiments of the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments of the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication section 409, and / or installed from the removable medium 411. When the computer program is executed by a central processing unit (CPU) 401, various functions defined in the system of the present application are executed.

[0170] It should be noted that the computer-readable medium shown in the embodiments of the present application can be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable computer program. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable medium can be transmitted using any appropriate medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0171] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present application. Among them, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the above module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram or flowchart, as well as the combination of blocks in the block diagram or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0172] The units involved in the embodiments described in this application can be implemented in software or in hardware, and the described units can also be provided in a processor. Among them, the names of these units do not constitute a limitation to the units themselves in some cases.

[0173] Another aspect of this application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor of a computer, the computer is caused to execute the method as described above. The computer-readable storage medium can be included in the electronic device described in the above embodiments, or can exist alone without being assembled into the electronic device.

[0174] Another aspect of this application also provides a computer program product or a computer program. The computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the methods provided in the above various embodiments.

[0175] The above embodiments are only preferred embodiments given to fully illustrate this application, and the protection scope of this application is not limited thereto. Equivalent substitutions or transformations made by those skilled in the art on the basis of this application are within the protection scope of this application.

Claims

1. A method for predicting water consumption in construction engineering, characterized in that: Includes steps: Obtaining first sample data of a current construction site and second sample data of other construction sites, and obtaining a construction plan and a weather forecast for the construction site, wherein the first sample data and the second sample data both include water consumption at multiple time points and values ​​of multiple influencing factors at multiple time points; Based on the first sample data, a correlation analysis is performed between the water consumption of the current construction site and a plurality of influencing factors to obtain a target influencing factor affecting the construction site; Building a statistical model based on the first sample data and the target influencing factor; and using the statistical model to perform confidence evaluation on the second sample data to obtain the confidence of the second sample data, wherein the statistical model includes a value relationship between water consumption and the target influencing factor; Training the BP neural network based on the first sample data and the second sample data whose confidence is higher than a preset confidence threshold to obtain a prediction model; The predicted value of the target influencing factor is extracted from the construction plan of the construction site and the weather forecast, and the water consumption of the current construction site is predicted based on the prediction model and the predicted value of the target influencing factor.

2. A method for predicting water consumption in construction engineering according to claim 1, characterized in that: Based on the first sample data, the water consumption of the current construction site is analyzed for correlation with various influencing factors to obtain target influencing factors affecting the construction site, including: The water consumption at multiple time points in the first sample data is normalized to obtain a water consumption time series {X1, X2, ..., X t }; and normalize the values ​​of multiple influencing factors at multiple time points in the first sample data to obtain the time series of multiple influencing factors {n1, n2, ..., n t }, where X1 is the normalized water consumption at the first time point, X2 is the normalized water consumption at the second time point, and X t is the normalized water consumption at the tth time point, n1 is the normalized value of the influencing factor at one time point, n2 is the normalized value of the influencing factor at the second time point, and n t is the normalized value of the impact factor at the tth time point; The water consumption time series is aligned with the time series of multiple influencing factors in the time dimension; and the difference between the aligned water consumption time series and the time series is calculated to obtain the difference sequence {|X1-n1|,|X2-n2|,...,|X t -n t |}; The correlation R of the impact factor is calculated based on the difference sequence. The mathematical expression of the correlation R is: In the formula, N is the number of multiple time points; The influencing factors with a correlation degree greater than a preset correlation degree threshold are taken as target influencing factors affecting the construction site.

3. A method for predicting water consumption in construction engineering according to claim 2, characterized in that: Building a statistical model based on the first sample data and the target impact factor includes: Dividing the first sample data based on multiple water consumption ranges to obtain multiple sample data units; Density clustering is performed on the values ​​of target influencing factors in multiple sample data units at multiple time points to obtain multiple clusters; The cluster with the number of data points greater than the preset threshold is taken as the target cluster, and the average value of the target impact factor in the target cluster is calculated; Eliminate data points in the target cluster whose deviation from the average value is greater than a preset deviation rate threshold to obtain an updated target cluster; Calculate the average value A and standard deviation σ of the target impact factor in the updated target cluster; Based on the average value A and standard deviation σ of the target influencing factor in the target cluster, a first reference value range (A-α1σ, A+α1σ), a second reference value range {(A-α2σ, A-α1σ), (A+α1σ, A+α2σ)} and a third reference value range {(A-α3σ, A-α2σ), (A+α2σ, A+α3σ)} of the target influencing factor under the current water use range are constructed, wherein α1, α2 and α3 are all range adjustment parameters and satisfy α1<α2<α3; A statistical model is constructed based on multiple first reference value ranges, second reference value ranges, and third reference value ranges of the target influencing factor corresponding to multiple water consumption ranges.

4. A method for predicting water consumption in construction engineering according to claim 3, characterized in that: Using the statistical model to perform confidence evaluation on the second sample data to obtain the confidence of the second sample data includes: Extracting the water consumption and the value of the influencing factor at each time point in the second sample data to obtain a plurality of samples to be evaluated; Determine the water consumption range, the first reference value range, the second reference value range and the third reference value range in the statistical model based on the water consumption of the sample to be evaluated; The value of the target impact factor in the sample to be evaluated is compared with the corresponding first reference value range, second reference value range and third reference value range, and the confidence C of the sample to be evaluated is obtained based on the comparison result, wherein the mathematical expression of the confidence C is: In the formula, N1 is the number of target influencing factors in the sample to be evaluated that fall into the first reference value range, C1 is the confidence level when falling into the first reference value range, N2 is the number of target influencing factors in the sample to be evaluated that fall into the second reference value range, C2 is the confidence level when falling into the second reference value range, N3 is the number of target influencing factors in the sample to be evaluated that fall into the third reference value range, and C3 is the confidence level when falling into the third reference value range.

5. A method for predicting water consumption in construction engineering according to claim 1, characterized in that: The BP neural network is trained based on the first sample data and the second sample data whose confidence is higher than a preset confidence threshold to obtain a prediction model, including: The water consumption at multiple time points in the first sample data and the second sample data whose confidence is higher than a preset confidence threshold is used as labels, and the values ​​of the target influencing factors at multiple time points in the first sample data and the second sample data whose confidence is higher than a preset confidence threshold are used as training data to construct a training data set; The BP neural network is trained based on the training data set combined with a particle swarm optimization algorithm, and a prediction model is obtained after the loss function is minimized.

6. A method for predicting water consumption in construction engineering according to claim 5, characterized in that: The mathematical expression of the loss function F(ψ,ω,θ,r) is: In the formula, ψ is the error value in the BP neural network, ω, θ, and r are the input layer, hidden layer, and output layer of the neural network respectively; N1 and M are the weight and threshold node numbers respectively; y t (s) is the expected output of the neural network; is the actual output; t is a node that needs to optimize the connection weight, and s is a node that needs to optimize the threshold.

7. A method for predicting water consumption in construction engineering according to claim 6, characterized in that: The water consumption of the current construction site is predicted based on the prediction model and the predicted value of the target influencing factor, including: The predicted value of the target influencing factor is input into the prediction model to obtain the predicted value of water consumption corresponding to the predicted value time of the target influencing factor.

8. A construction engineering water consumption prediction system, characterized in that: include: An acquisition module, used to acquire first sample data of a current construction site and second sample data of other construction sites, and acquire a construction plan and a weather forecast for the construction site, wherein the first sample data and the second sample data both include water consumption at multiple time points and values ​​of multiple influencing factors at multiple time points; A correlation analysis module, used to perform correlation analysis on the water consumption of the current construction site and multiple influencing factors based on the first sample data, to obtain target influencing factors affecting the construction site; A sample screening module, configured to construct a statistical model based on the first sample data and the target influencing factor; and to use the statistical model to perform confidence evaluation on the second sample data to obtain the confidence of the second sample data, wherein the statistical model includes a value relationship between water consumption and the target influencing factor; A model training module, used to train a BP neural network based on the first sample data and second sample data with a confidence level higher than a preset confidence threshold to obtain a prediction model; The prediction module is used to extract the predicted value of the target influencing factor from the construction plan of the construction site and the weather forecast, and predict the water consumption of the current construction site based on the prediction model and the predicted value of the target influencing factor.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

10. An electronic terminal, characterized in that: include: Processor and memory; The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the terminal executes the method according to any one of claims 1 to 7.

Images