Big data-based urban water supply demand prediction method

Abstract

This invention discloses a big data-based method for predicting urban water supply demand, belonging to the field of urban water supply system management technology. It includes multi-source data integration, water use behavior profiling modeling, and urban water supply demand prediction. The invention employs a dynamic water use behavior profiling modeling method that combines state space construction and conditional evolution modeling. This method can reflect the formation characteristics of different water use patterns and their evolutionary trends with environmental changes within a unified behavioral state framework, thus providing a clearly structured behavioral foundation for urban water supply demand prediction. Furthermore, the invention utilizes an urban water supply demand prediction method based on the evolution of water use behavior state proportions, achieving a dynamic reflection of changes in urban water use structure. This allows the water supply demand prediction results to adaptively adjust with the evolution of water use behavior structure, improving the accuracy, stability, and applicability of urban water supply demand prediction.

Description

Technical Field

[0001] This invention belongs to the field of urban water supply system management technology, specifically referring to a method for predicting urban water supply demand based on big data. Background Technology

[0002] Urban water demand forecasting based on big data comprehensively utilizes urban water metering data, population and building structure data, as well as external environmental data such as meteorology and holidays. By analyzing and modeling urban water use behavior, it depicts the changing patterns of water demand under different regional and temporal conditions, enabling dynamic forecasting of urban and regional water supply demand. The aim is to improve the accuracy and stability of urban water demand forecasting from the perspective of water use behavior and demand evolution, providing a scientific basis for urban water supply scheduling, optimization of water supply facility operation, rational allocation of resources, and emergency water supply management, thereby ensuring the safe, efficient, and refined operation of the urban water supply system.

[0003] However, in the existing process of urban water demand forecasting, there are technical problems such as water demand being significantly affected by regional differences, time variations, and external factors such as weather and holidays, and water use behavior patterns being prone to change. This makes it difficult for traditional forecasting methods based on historical water consumption to accurately reflect changes in water use behavior structure, and the forecasting stability is insufficient. Furthermore, there are technical problems such as the fact that the overall urban water demand is composed of multiple types of water use behaviors, and the change patterns of different behaviors are significantly different. When the proportion of some behaviors changes, the overall water demand forecast is prone to structural distortion, weakening of peak and valley characteristics, and lagging response to environmental disturbances. Summary of the Invention

[0004] To address the above issues and overcome the shortcomings of existing technologies, this invention provides a big data-based method for predicting urban water supply demand. It creatively employs a dynamic water use behavior profiling modeling method that combines state space construction with conditional evolution modeling. This method can reflect the formation characteristics of different water use patterns and their evolutionary trends with environmental changes within a unified behavioral state framework, thus providing a clearly structured behavioral foundation for urban water supply demand prediction. Furthermore, it creatively adopts an urban water supply demand prediction method based on the evolution of water use behavior state proportions, achieving a dynamic reflection of changes in urban water use structure. This allows the water supply demand prediction results to adaptively adjust with the evolution of water use behavior structure, improving the accuracy, stability, and applicability of urban water supply demand prediction.

[0005] The technical solution adopted by this invention is as follows: The urban water supply demand prediction method based on big data provided by this invention includes the following steps:

[0006] Step S1: Multi-source data integration;

[0007] Step S2: Model water behavior profile;

[0008] Step S3: Urban water supply demand forecasting.

[0009] Furthermore, in step S1, the multi-source data integration specifically involves collecting water metering data, external environmental data, and regional structural attribute parameters, performing missing value processing, outlier removal, and standardization preprocessing on data from different sources, and performing time alignment and spatial mapping according to a unified time scale and regional division rules to form a multi-source integrated dataset.

[0010] Further, in step S2, the water use behavior profile modeling specifically involves obtaining a dynamic water use behavior profile based on a multi-source integrated dataset and using a dynamic water use behavior profile modeling method that combines state space construction and conditional evolution modeling. This includes the following steps:

[0011] Step S21: Decoupling of behavioral features, used to construct interpretable water use behavior features. Specifically, by decoupling the behavioral features of water use metering data, water use intensity features, water use rhythm features, and water use elasticity features are extracted and normalized to construct a multidimensional water use behavior feature vector.

[0012] Step S22: Constructing the state space of water use behavior. Specifically, by setting water use intensity range constraints, rhythm similarity constraints, and water use elasticity range constraints, the water use behavior is divided into states. Water use behaviors that satisfy different combinations of constraints are mapped into K discrete water use behavior states to form the water use behavior state space.

[0013] Step S23: Water use behavior evolution modeling, specifically, based on the water use behavior state space, establish the water use behavior state sequence corresponding to each region, and introduce external environmental conditions as the modulating factors of state transition, construct a conditional state transition model to describe the transition relationship between water use behavior states under different environmental conditions, and obtain the water use behavior evolution model.

[0014] Step S24: Constructing a water use behavior profile, specifically by comprehensively modeling regional water use behavior using the state space of water use behavior, the evolution model of water use behavior, and regional structural attribute parameters, and constructing a dynamic water use behavior profile.

[0015] Further, in step S3, the urban water supply demand forecasting specifically involves using a method based on the evolution of water use behavior status proportions, based on a multi-source integrated dataset and dynamic water use behavior profiles, to obtain an urban water supply demand forecasting report, including the following steps:

[0016] Step S31: Behavioral state-level demand prediction, which is used to predict the future water demand corresponding to different water use behavior states based on the dynamic water use behavior profile. Specifically, it involves constructing a water demand prediction model under state conditions for each type of water use behavior state, and predicting the state-level water demand of each water use behavior state at future time to obtain the state-level water demand prediction result.

[0017] Step S32: Water use behavior state proportion inference, which is used to reflect the structural evolution trend of different water use behavior patterns. Specifically, by obtaining the proportion vector of each water use behavior state in the urban water use structure at the current time step, and based on the water use behavior evolution model constructed in step S2, the proportion of each water use behavior state at future time is predicted to obtain the water use behavior state proportion prediction result.

[0018] Step S33: Urban water supply demand synthesis, specifically by using the predicted results of the proportion of water use behavior status as weights to perform weighted calculations on the state-level water demand prediction results to obtain the overall urban water supply demand prediction results.

[0019] Step S34: Scale consistency correction, specifically, based on the water use behavior status of each region, the regional water supply demand prediction result is calculated through the water demand prediction model under the stated status conditions. Then, water supply demand balance constraints are introduced to perform consistency correction on the regional water supply demand prediction result, and the regional water supply demand prediction correction result is obtained.

[0020] Step S35: Output the prediction results. Specifically, by executing steps S31 to S34, the prediction results of the overall urban water supply demand, the regional water supply demand prediction correction results, and the prediction results of the proportion of water use behavior status are integrated to generate an urban water supply demand prediction report.

[0021] The beneficial effects achieved by the present invention using the above solution are as follows:

[0022] (1) In the process of predicting urban water supply demand, water demand is significantly affected by external factors such as regional differences, time changes, weather and holidays, and water use behavior patterns are prone to change. This makes it difficult for traditional prediction methods based on historical water consumption to accurately reflect changes in water use behavior structure and has insufficient prediction stability. This solution creatively adopts a dynamic water use behavior profile modeling method that combines state space construction and condition evolution modeling. It can reflect the formation characteristics of different water use patterns and their evolution trend with environmental changes under a unified behavioral state framework, thereby providing a clear behavioral foundation for predicting urban water supply demand.

[0023] (2) In the existing process of urban water demand forecasting, the overall urban water demand is composed of multiple types of water use behaviors, and the change patterns of different behaviors are significantly different. When the proportion of some behaviors changes, the overall water demand forecast is prone to structural distortion, weakening of peak and valley characteristics, and lag in response to environmental disturbances. This solution creatively adopts an urban water demand forecasting method based on the evolution of the proportion of water use behavior states. This realizes the dynamic reflection of the change process of urban water use structure, so that the water demand forecasting results can be adaptively adjusted with the evolution of water use behavior structure, thereby improving the accuracy, stability and applicability of urban water demand forecasting. Attached Figure Description

[0024] Figure 1 A flowchart illustrating the urban water supply demand forecasting method based on big data provided by this invention;

[0025] Figure 2 A flowchart illustrating the process of modeling a water behavior profile for step S2;

[0026] Figure 3 This is a flowchart illustrating the process of predicting urban water supply demand in step S3.

[0027] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. Detailed Implementation

[0028] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0029] In the description of this invention, it should be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0030] Example 1, see Figure 1 The present invention provides a method for predicting urban water supply demand based on big data, which includes the following steps:

[0031] Step S1: Multi-source data integration;

[0032] Step S2: Model water behavior profile;

[0033] Step S3: Urban water supply demand forecasting.

[0034] Example 2, see Figure 1 This embodiment is based on the above embodiment. In step S1, the multi-source data integration specifically involves collecting water metering data, external environmental data, and regional structural attribute parameters, performing missing value processing, outlier removal, and standardization preprocessing on data from different sources, and performing time alignment and spatial mapping according to a unified time scale and regional division rules to form a multi-source integrated dataset.

[0035] The water metering data includes, but is not limited to, historical water consumption data of the region, water consumption records of different time periods, and cumulative water consumption data, which are used to reflect the actual water consumption of different regions in different time periods.

[0036] The external environmental data includes, but is not limited to, temperature data, rainfall data, and holiday type data. Calendar data is used to identify weekdays, weekends, and holiday types to characterize the impact of external environmental and time factors on water use behavior.

[0037] The regional structural attribute parameters include, but are not limited to, the number of people served in the region, population density, building type, and regional functional attributes, which are used to characterize the differences in water use structure among different regions.

[0038] The time scale can specifically adopt an hourly or daily time granularity;

[0039] The specific rules for dividing the area can be set based on water supply zones, administrative divisions, or the service area of ​​the pipeline network.

[0040] Example 3, see Figure 1 and Figure 2 This embodiment is based on the above embodiment. In step S2, the water use behavior profiling modeling is used to characterize the water use behavior patterns and their evolution laws of various regions under different time and external environmental conditions. Specifically, based on a multi-source integrated dataset, a dynamic water use behavior profiling modeling method combining state space construction and conditional evolution modeling is adopted to obtain a dynamic water use behavior profile, including the following steps:

[0041] Step S21: Decoupling of behavioral features, used to construct interpretable water use behavior features. Specifically, by decoupling the behavioral features of water use metering data, water use intensity features, water use rhythm features, and water use elasticity features are extracted and normalized to construct a multidimensional water use behavior feature vector.

[0042] The formula for calculating the water intensity characteristic is as follows:

[0043] ;

[0044] In the formula, This represents the water intensity characteristics of region u. It is region u at time point The water consumption, where u is the region index. It is a time point index, N u It refers to the number of people served in the region;

[0045] The water use rhythm characteristics are used to characterize the periodic changes in water use behavior in different regions, and the calculation formula is as follows:

[0046] ;

[0047] ;

[0048] In the formula, R u (d) represents the water use rhythm characteristics of region u. This represents the average water consumption of region u on dates of type d, where d is the date type, including weekdays, weekends, and public holidays, h is the hourly index, and D... h It is the set of time points belonging to the h-th hour;

[0049] The water elasticity characteristic is used to measure the sensitivity of regional water consumption to changes in external environmental factors, and the calculation formula is as follows:

[0050] ;

[0051] In the formula, E u It is the water elasticity characteristic of region u. It is a temperature variable;

[0052] Step S22: Construction of the water use behavior state space, used to map continuous water use behavior characteristics into discrete water use behavior states with engineering interpretability. Specifically, by setting water use intensity interval constraints, rhythm similarity constraints, and water use elasticity interval constraints, the water use behavior is divided into states. Water use behaviors satisfying different combinations of constraints are mapped into K discrete water use behavior states, forming the water use behavior state space. The calculation formula is as follows:

[0053] ;

[0054] ;

[0055] In the formula, S is the state space of water use behavior, s1 is the first type of water use behavior state, s2 is the second type of water use behavior state, and s K This is the Kth type of water use behavior state, where K is the number of water use behavior states, and s kThis represents the k-th water use behavior state, where k is the first index of the water use behavior state, and I is the water use intensity feature. Let sim(·) be the water intensity interval corresponding to the k-th type of water use behavior state, sim(·) be the water use rhythm similarity metric function, and R be the water use rhythm feature. k It is a typical water use rhythm template corresponding to the k-th type of water use behavior state. It is the rhythmic similarity threshold corresponding to the k-th type of water use behavior state, and E is the water use elasticity feature. It is the water use elasticity interval corresponding to the kth type of water use behavior state;

[0056] Step S23: Water use behavior evolution modeling, used to characterize the dynamic evolution law of water use behavior state in the time dimension. Specifically, based on the water use behavior state space, establish the water use behavior state sequence corresponding to each region, and introduce external environmental conditions as the modulating factor of state transition to construct a conditional state transition model to describe the transition relationship between water use behavior states under different environmental conditions, and obtain the water use behavior evolution model.

[0057] The external environmental conditions include, but are not limited to, holiday types, temperature changes, and rainfall.

[0058] The conditional state transition model is specifically a conditional Markov state transition model, which is used to describe the transition trend and intensity between water use behavior states under different external conditions.

[0059] Step S24: Water use behavior profile construction, used to form a behavior profile that can characterize the dynamic changes of water use behavior in various regions under different time and external environmental conditions. Specifically, by integrating the water use behavior state space, water use behavior evolution model, and regional structural attribute parameters, a unified model of regional water use behavior is constructed to build a dynamic water use behavior profile. The calculation formula is as follows:

[0060] ;

[0061] In the formula, It is a dynamic water use behavior profile of region u, P u (t) is the water use behavior evolution model for region u, where t is the time step index. These are the region structure attribute parameters.

[0062] By performing the above operations, this solution addresses the technical problem that in the existing urban water demand forecasting process, water demand is significantly affected by external factors such as regional differences, time variations, weather, and holidays, and water use behavior patterns are prone to change. This makes it difficult for traditional forecasting methods based on historical water consumption to accurately reflect changes in water use behavior structure and results in insufficient forecast stability. This solution creatively adopts a dynamic water use behavior profiling modeling method that combines state space construction and conditional evolution modeling. It can reflect the formation characteristics of different water use patterns and their evolutionary trends with environmental changes within a unified behavioral state framework, thereby providing a clear behavioral foundation for urban water demand forecasting.

[0063] Example 4, see Figure 1 and Figure 3 This embodiment is based on the above embodiment. In step S3, the urban water supply demand forecasting specifically involves using a city water supply demand forecasting method based on the evolution of the proportion of water use behavior states, based on a multi-source integrated dataset and dynamic water use behavior profiles, to obtain an urban water supply demand forecasting report, including the following steps:

[0064] Step S31: Behavioral state-level demand prediction, which is used to predict the future water demand corresponding to different water use behavior states based on the dynamic water use behavior profile. Specifically, it involves constructing a water demand prediction model under state conditions for each type of water use behavior state, and predicting the state-level water demand of each water use behavior state at future time to obtain the state-level water demand prediction result, thereby avoiding prediction bias caused by mixed modeling of different behavioral data.

[0065] The water demand prediction model under the aforementioned state conditions is specifically constructed by using sample data of the same water use behavior state as training samples, selecting the average water use intensity and external environmental conditions corresponding to the water use behavior state as model inputs, and using a time series-based prediction model for model training to obtain the water demand prediction model under the state conditions.

[0066] Step S32: Deducing the proportion of water use behavior states to reflect the structural evolution trend of different water use behavior patterns. Specifically, this involves obtaining the proportion vector of each water use behavior state in the urban water use structure at the current time step, and based on the water use behavior evolution model constructed in Step S2, predicting the proportion of each water use behavior state at future times to obtain the predicted proportion of water use behavior states. The calculation formula is as follows:

[0067] ;

[0068] In the formula, is the predicted proportion of the k-th type of water use behavior in the urban water use structure, and i is the second index of water use behavior. P represents the proportion of the current time step i-th type of water use behavior in the urban water use structure. ik (t) is the state transition probability of the i-th type of water use behavior state to the k-th type of water use behavior state at the current time step;

[0069] Step S33: Urban water supply demand synthesis, specifically, involves using the predicted proportion of water use behavior states as weights to perform a weighted calculation on the state-level water demand prediction results to obtain the overall urban water supply demand prediction result. The calculation formula is as follows:

[0070] ;

[0071] In the formula, Q city (t+1) is the predicted result of the city's overall water supply demand, Q k (t+1) is the state-level water demand forecast value of the k-th type of water use behavior state;

[0072] Step S34: Scale consistency correction, specifically, based on the water use behavior status of each region, the regional water supply demand prediction result is calculated through the water demand prediction model under the stated status conditions. Then, water supply demand balance constraints are introduced to perform consistency correction on the regional water supply demand prediction result, thereby eliminating regional prediction bias and improving the stability of the prediction result.

[0073] The formula for calculating the water supply demand balance constraint is as follows:

[0074] ;

[0075] In the formula, min is the minimum value. This is a regional-level water supply demand forecast correction value. This is the regional water supply demand forecast, where U represents the number of regions.

[0076] Step S35: Output the prediction results. Specifically, by executing steps S31 to S34, the prediction results of the overall urban water supply demand, the regional water supply demand prediction correction results, and the prediction results of the proportion of water use behavior status are integrated to generate an urban water supply demand prediction report. This report is used to support urban water supply scheduling, resource optimization, emergency management, and water supply strategy decision-making, thereby achieving refined management and scientific regulation of the urban water system.

[0077] By performing the above operations, this solution addresses the technical problems in existing urban water demand forecasting processes, where the overall urban water demand is composed of multiple types of water use behaviors with significantly different patterns of change. When the proportion of some behaviors changes, the overall water demand forecast is prone to structural distortion, weakened peak-valley characteristics, and delayed response to environmental disturbances. This solution creatively adopts an urban water demand forecasting method based on the evolution of the proportion of water use behavior states. This achieves a dynamic reflection of the changes in the urban water use structure, enabling the water demand forecasting results to adaptively adjust with the evolution of the water use behavior structure, thereby improving the accuracy, stability, and applicability of urban water demand forecasting.

[0078] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0079] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.

[0080] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A method for predicting urban water supply demand based on big data, characterized in that: The method includes the following steps: Step S1: Multi-source data integration; Step S2: Water use behavior profile modeling, specifically based on a multi-source integrated dataset, adopts a dynamic water use behavior profile modeling method that combines state space construction and conditional evolution modeling to obtain a dynamic water use behavior profile, including the following steps: Step S21 Decoupling of behavioral features, Step S22 Construction of water use behavior state space, Step S23 Evolutionary modeling of water use behavior, and Step S24 Construction of water use behavior profile. In step S23, the water use behavior evolution modeling specifically involves establishing a water use behavior state sequence corresponding to each region based on the water use behavior state space, introducing external environmental conditions as modulation factors for state transition, constructing a conditional state transition model to describe the transition relationship between water use behavior states under different environmental conditions, and obtaining a water use behavior evolution model. Step S3: Urban water supply demand forecasting. Specifically, based on the multi-source integrated dataset and dynamic water use behavior profile, an urban water supply demand forecasting method based on the evolution of the proportion of water use behavior status is adopted to obtain an urban water supply demand forecasting report, including the following steps: Step S31 Behavioral status level demand forecasting, Step S32 Water use behavior status proportion deduction, Step S33 Urban water supply demand synthesis, Step S34 Scale consistency correction, and Step S35 Forecast result output. In step S32, the water use behavior state proportion inference is used to reflect the structural evolution trend of different water use behavior patterns. Specifically, it involves obtaining the proportion vector of each water use behavior state in the urban water use structure at the current time step, and predicting the proportion of each water use behavior state at future times based on the water use behavior evolution model constructed in step S2, thereby obtaining the water use behavior state proportion prediction result.

2. The urban water supply demand forecasting method based on big data according to claim 1, characterized in that: In step S21, the behavior feature decoupling is used to construct interpretable water use behavior features. Specifically, by decoupling the behavior features of water use metering data, water use intensity features, water use rhythm features, and water use elasticity features are extracted respectively, and normalized to construct a multi-dimensional water use behavior feature vector. In step S22, the water use behavior state space is constructed by setting water use intensity range constraints, rhythm similarity constraints, and water use elasticity range constraints to divide water use behavior into states, and mapping water use behavior that satisfies different combinations of constraints into K discrete water use behavior states to form a water use behavior state space.

3. The urban water supply demand forecasting method based on big data according to claim 2, characterized in that: In step S24, the construction of the water use behavior profile specifically involves modeling the regional water use behavior in a unified manner by integrating the water use behavior state space, the water use behavior evolution model, and regional structural attribute parameters, thereby constructing a dynamic water use behavior profile.

4. The urban water supply demand forecasting method based on big data according to claim 3, characterized in that: In step S31, the behavior state-level demand prediction is used to predict the future water demand corresponding to different water use behavior states based on the dynamic water use behavior profile. Specifically, it involves constructing a water demand prediction model under state conditions for each type of water use behavior state, and predicting the state-level water demand of each water use behavior state at future times to obtain the state-level water demand prediction result.

5. The urban water supply demand forecasting method based on big data according to claim 4, characterized in that: In step S33, the synthesis of urban water supply demand specifically involves weighting the state-level water demand prediction results by using the water use behavior state proportion prediction results as weights to obtain the overall urban water supply demand prediction results. In step S34, the scale consistency correction specifically involves calculating the regional water supply demand prediction result based on the water use behavior status of each region using the water demand prediction model under the stated status conditions, and then introducing water supply demand balance constraints to perform consistency correction on the regional water supply demand prediction result to obtain the regional water supply demand prediction correction result.

6. The urban water supply demand forecasting method based on big data according to claim 5, characterized in that: In step S35, the prediction result is output specifically by integrating the overall urban water supply demand prediction result, the regional water supply demand prediction correction result, and the water use behavior status proportion prediction result through steps S31 to S34 to generate an urban water supply demand prediction report.

7. The urban water supply demand forecasting method based on big data according to claim 6, characterized in that: In step S1, the multi-source data integration specifically involves collecting water metering data, external environmental data, and regional structural attribute parameters, performing missing value processing, outlier removal, and standardization preprocessing on data from different sources, and then performing time alignment and spatial mapping according to a unified time scale and regional division rules to form a multi-source integrated dataset.

Images