Large language model-based urban and rural water supply system scheduling decision-making method and system

Abstract

The invention provides an urban and rural water supply system scheduling decision method and system based on a large language model, and the method comprises the steps: obtaining historical water supply operation related monitoring data and water demand data, and building a scheduling decision set with the water demand and the minimum energy consumption of a pump set as targets; the real-time water supply operation related monitoring data and the historical water supply operation related monitoring data are clustered, and a scheduling decision of the historical water supply operation related monitoring data of the same category is found out and serves as a retrieval result; and inputting the retrieval result and the previous stage auxiliary decision into a large language model, and deducing a scheduling auxiliary decision corresponding to the real-time water supply operation related monitoring data. According to the method, real-time water supply operation related monitoring data, predicted water demand data, a retrieval enhancement technology and a large language model are combined, a scheduling strategy with the lowest total energy consumption of a pump set is automatically generated through a non-engineering measure method on the premise that safe and stable water use of a user is guaranteed, and production operation of a multi-stage water supply system is guided.

Description

Technical Field

[0001] The present invention relates to the field of water supply, and more specifically, to a scheduling decision method and system for urban and rural water supply systems based on a large language model. Background Art

[0002] Generally, the coverage of urban-rural integrated water supply is wide, requiring mid-way pressurized water relay supply. In order to make full use of the water distribution network of the original township water supply system, the clear water tank and water delivery pump house of the township water plant will be retained, thus forming a multi-level water supply pattern of urban water plant-midway pressurized pump station-township water plant. Therefore, the water supply pattern is relatively complex, and the water supply energy consumption also faces greater challenges. How to minimize the total energy consumption of the pump group of the water plant and the midway pressurized pump station while ensuring the safety and stable water use of users is a difficult problem faced by urban-rural integrated water supply. Summary of the invention

[0003] In view of the technical problems existing in the prior art, the present invention provides a scheduling decision method and system for urban and rural water supply systems based on a large language model.

[0004] According to a first aspect of the present invention, a scheduling decision method for an urban and rural water supply system based on a large language model is provided, comprising: S10, obtaining real-time water supply operation-related monitoring data and predicted water demand data of urban and rural water supply systems; S20, obtaining historical water supply operation-related monitoring data and water demand data of the urban and rural water supply system, and establishing a scheduling decision set based on the historical water supply operation-related monitoring data and water demand data, with water demand and minimum energy consumption of the pump group as the target, wherein the scheduling decision set includes a mapping relationship between the historical water supply operation-related monitoring data and the scheduling decision; S30, clustering the real-time water supply operation-related monitoring data and the historical water supply operation-related monitoring data using a text embedding model, and finding the historical water supply operation-related monitoring data of the same category as the real-time water supply operation-related monitoring data; querying the scheduling decision set based on the historical water supply operation-related monitoring data of the same category as the real-time water supply operation-related monitoring data, and obtaining the corresponding scheduling decision as the scheduling decision of the real-time water supply operation-related monitoring data as a search result; S40, inputting the search results and the auxiliary decision of the previous stage into the large language model to infer the scheduling auxiliary decision corresponding to the real-time water supply operation related monitoring data; S50: Scheduling the water supply pump group based on the scheduling auxiliary decision.

[0005] Based on the above technical solution, the present invention can also make the following improvements.

[0006] Optionally, the real-time water supply operation-related monitoring data is the water level of the water plant's clear water tank, the water plant's outlet water pressure, the booster station's outlet water pressure, and the water pressure at the water point transmitted back by the on-site water condition equipment via the Internet of Things. The predicted water demand data is the predicted water demand data for urban and rural areas from now to a set time period in the future obtained by statistical methods or AI models; The format and content of the real-time water supply operation-related monitoring data and predicted water demand data are as follows:

[0007] in, To collect relevant data for real-time water supply operation, Representative This type of data is real-time water supply operation related monitoring data or predicted water demand data.

[0008] Optionally, in S20, the scheduling decision set is established based on the historical water supply operation-related monitoring data and water demand data, with water demand and minimum energy consumption of the pump group as the target, including: S21, obtaining historical water supply operation-related monitoring data and water demand data of the urban and rural water supply system, and constructing a historical data set, wherein the historical data set is constructed by constructing a plurality of water supply operation-related monitoring data and water demand data at different historical time points obtained through a historical process, and the format and content of the water supply operation-related monitoring data and water demand data at a single historical time point in the historical data set are consistent with the format and content of the real-time water supply operation-related monitoring data and predicted water demand data, and the historical data set is expressed as:

[0009] in, A collection of historical water supply operation related data. Representative A subset of historical moments, Representative The first historical moment Kind of data; S22, according to the historical water supply operation related monitoring data and water demand data at each time point, the scheduling decision at each time is determined based on the model analysis to achieve the water demand target and the minimum energy consumption of the pump group, and finally a set of optimized scheduling decisions at multiple different historical time points is formed, which is expressed as:

[0010] is the scheduling decision set, represents the scheduling decision subset at the mth moment, Representative The first one in the scheduling decision subset at the time Messages; S23, based on the historical water supply operation related monitoring data and the scheduling decision subset corresponding to the time point, a mapping relationship between the historical water supply operation related monitoring data and the scheduling decision at each historical time point is constructed to obtain the historical water supply operation related monitoring data-scheduling decision set:

[0011] in, It is a set of monitoring data related to historical water supply operation and scheduling decision; Representative A collection of historical water supply operation related data at each moment ; Representative The scheduling decision subset of the pen time , the historical water supply operation related data set and the scheduling decision set are bijective relations .

[0012] Optionally, the step S22, according to the historical water supply operation-related monitoring data and water demand data at each time point, determines the scheduling decision at each time point based on the model analysis, including: Through the hydraulic model, in different preset scheduling decisions, after using the pump group to pressurize at different water outlet pressures and water pressures at water points, it is simulated whether the water demand target and the minimum energy consumption of the pump group can be achieved, and the pump group scheduling decision that achieves the water demand target and the minimum energy consumption of the pump group is obtained.

[0013] Optionally, the number of messages in the scheduling decision subset at each moment varies according to the time point, water level in the water plant's clear water tank, outlet water pressure, water pressure at the water point, pump group performance, pressurization location, and water demand data.

[0014] Optionally, the text embedding model adopts an unsupervised clustering model, including a K-means model, a self-organizing map model, a neighbor propagation algorithm model, a hierarchical clustering model, a mean shift algorithm model, a DBSCAN model or a binary K-means model.

[0015] Optionally, the S30, clustering the real-time water supply operation related monitoring data and the historical water supply operation related monitoring data using a text embedding model, and finding the historical water supply operation related monitoring data of the same category as the real-time water supply operation related monitoring data, includes: S31, collecting the real-time water supply operation related data and the historical water supply operation related data set At the same time, the text embedding model is input for clustering analysis, and the clustering results are expressed as:

[0016] in, is the clustering category set; Represents a collection of real-time water supply operation related data Clustering categories; Represents a collection of historical water supply operation related data The clustering categories of the subsets at different historical time points in the real-time water supply operation related data set , historical water supply operation related data collection and the set of cluster categories It is a bijective relation, ; S32, by clustering category set Find the real-time water supply operation related data set Clustering categories ,pass and subcollections Intersection finds the same cluster category , ∈ , find the same cluster categories Corresponding historical water supply operation related data set .

[0017] Optionally, the previous stage auxiliary decision refers to a scheduling decision that was released before the current time point, and the large language model includes GPT, LaMDA, LLaMA, Gemini, Wenxin Yiyan, Baichuan, Xinghuo, Tongyi Qianwen and Hunyuan.

[0018] According to a second aspect of the present invention, a scheduling decision system for urban and rural water supply systems based on a large language model is provided, comprising: An acquisition module is used to obtain real-time water supply operation-related monitoring data and predicted water demand data of urban and rural water supply systems; Establish a module for obtaining historical water supply operation-related monitoring data and water demand data of urban and rural water supply systems, and establish a scheduling decision set based on the historical water supply operation-related monitoring data and water demand data, with water demand and minimum energy consumption of pump groups as targets, wherein the scheduling decision set includes a mapping relationship between the historical water supply operation-related monitoring data and the scheduling decision; a clustering module, for clustering the real-time water supply operation-related monitoring data and the historical water supply operation-related monitoring data using a text embedding model, and finding the historical water supply operation-related monitoring data of the same category as the real-time water supply operation-related monitoring data; and performing a query in the scheduling decision set based on the historical water supply operation-related monitoring data of the same category as the real-time water supply operation-related monitoring data, and obtaining a corresponding scheduling decision as the scheduling decision of the real-time water supply operation-related monitoring data as a search result; An inference module, used to input the search results and the auxiliary decision of the previous stage into the large language model, and infer the scheduling auxiliary decision corresponding to the real-time water supply operation related monitoring data; The scheduling module is used to schedule the water supply pump group based on the scheduling auxiliary decision.

[0019] The present invention provides a scheduling decision method and system for urban and rural water supply systems based on a large language model, which obtains historical water supply operation-related monitoring data and water demand data, establishes a scheduling decision set with water demand and minimum energy consumption of pump groups as the target; clusters the real-time water supply operation-related monitoring data and the historical water supply operation-related monitoring data, and finds the scheduling decisions of the same category of historical water supply operation-related monitoring data as the search results; inputs the search results and the auxiliary decisions of the previous stage into the large language model, and infers the scheduling auxiliary decisions corresponding to the real-time water supply operation-related monitoring data. The present invention combines real-time water supply operation-related monitoring data, predicted water demand data, retrieval enhancement technology, and a large language model, and through a non-engineering measure method, realizes the automatic generation of a scheduling strategy with the lowest total energy consumption of the pump group under the premise of ensuring user safety and stable water use, and guides the production and operation of the multi-level water supply system. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A flow chart of a scheduling decision method for urban and rural water supply systems based on a large language model provided by the present invention; Figure 2 To construct a flowchart for scheduling decisions; Figure 3 Flowchart for retrieving scheduling decisions; Figure 4 A block diagram of a scheduling decision system for urban and rural water supply systems based on a large language model provided by the present invention. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. In addition, the technical features in the various embodiments or single embodiments provided by the present invention can be arbitrarily combined with each other to form a feasible technical solution. This combination is not subject to the constraints of the sequence of steps and / or the structural composition mode, but must be based on the ability of ordinary technicians in this field to achieve. When the combination of technical solutions is contradictory or cannot be achieved, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0022] Based on the defects in the background technology, the present invention proposes to construct a real-time, multivariable, nonlinear near-optimal urban-rural integrated water supply optimization scheduling model to achieve the purpose of urban-rural integrated water supply optimization scheduling and energy conservation and emission reduction. With the maturity and perfection of the large language model, the relevant technology can be used to promote the development of urban-rural integrated water supply scheduling and ensure the stable operation of the multi-stage water supply system. However, the current large language models are mostly general-purpose, and there are certain illusions and deviations in the results of professional scheduling strategy generation. The combination of real-time water supply operation-related monitoring data, predicted water demand data, retrieval enhancement technology, and large language model of the present invention provides some solutions to the above problems. Through a non-engineering measure method, under the premise of ensuring user safety and stable water use, the automatic generation of a scheduling strategy with the lowest total energy consumption of the pump group is achieved to guide the production and operation of the multi-stage water supply system.

[0023] Figure 1 A flow chart of a scheduling decision method for urban and rural water supply systems based on a large language model provided by the present invention, such as Figure 1 As shown, the method includes: S10, obtaining real-time water supply operation-related monitoring data and predicted water demand data of urban and rural water supply systems.

[0024] Among them, the real-time water supply operation-related monitoring data of the urban and rural water supply system is transmitted back through the on-site water condition equipment through the Internet of Things, including but not limited to the water level of the water plant's clear water tank, the water plant's outlet water pressure, the booster station's outlet water pressure, the water pressure at the water point, and other water supply operation-related monitoring data. The predicted water demand data is the predicted water demand data for urban and rural areas from now to the future, obtained through but not limited to statistical methods and AI models, such as urban and rural residents, public, commercial, industrial and other water demand data in the next few hours or 1 day. The real-time water supply operation-related data set obtained in this step For use in step S30.

[0025] S20, obtaining historical water supply operation-related monitoring data and water demand data of the urban and rural water supply system, and establishing a scheduling decision set based on the historical water supply operation-related monitoring data and water demand data, with water demand and minimum energy consumption of the pump group as the target, wherein the scheduling decision set includes a mapping relationship between the historical water supply operation-related monitoring data and the scheduling decision.

[0026] It is understandable that the historical water supply operation-related monitoring data and water demand data of urban and rural water supply systems are obtained to establish a scheduling decision set. The establishment steps are as follows: Figure 2 , the specific steps are as follows: S21. Obtain historical water supply operation-related monitoring data and water demand data, and generate a historical data set. The historical data set is generated by multiple different historical time points obtained through the historical process. The format and content of each data in the set must be consistent with the real-time water supply operation-related data set in step S10. The historical data set can be expressed as:

[0027] Collection of data related to historical water supply operations; Representative A subset of pen moments; Representative Pen time The historical water supply operation related data set is used in step S23.

[0028] S22. Make scheduling decisions based on historical water supply operation-related monitoring data and water demand data. Multiple-degree decision making is to review the historical water supply operation-related monitoring data and water demand data, and through expert manual experience or model analysis, review the history to see which scheduling decision should be used most correctly at the moment. The model analysis described above refers to the use of a hydraulic model to simulate whether the water demand target and minimum energy consumption can be achieved under different preset scheduling decisions, including but not limited to the use of a pump group to pressurize at different water outlet pressures and water pressures at water points. When the simulation shows that the water demand target and minimum energy consumption can be achieved, the hydraulic model outputs the optimal scheduling decision. Whether the water demand target and minimum energy consumption can be achieved includes but is not limited to trial and error using manual experience; operations research analysis can also be used; optimization model analysis can also be used. The multiple optimized scheduling decisions at different historical time points formed in the end can be expressed as:

[0029] To optimize the scheduling decision set, represents the optimal scheduling subset at the mth moment, Representative The optimal scheduling subset of the time The number of sub-collections at each moment may vary depending on the time, water level in the water plant's clear water tank, outlet water pressure, water pressure at water points, pump performance, pressurization location, and water demand. It may be an empty collection or may contain an indefinite number of collections. For use in step S23.

[0030] S23. By using the historical water supply operation related data set and optimizing the scheduling decision set corresponding to the time point, a mapping set of historical water supply operation related monitoring data and scheduling decisions is produced.

[0031]

[0032] A set of mappings for monitoring data and scheduling decisions; Representative A collection of historical water supply operation related data at each moment ; Representative The optimal scheduling decision set at pen time , the historical water supply operation related data set and the optimization scheduling decision set are bijective relations .

[0033] Historical water supply operation related data collection And monitoring data and scheduling decision sets , for use in step S30.

[0034] S30, clustering the real-time water supply operation related monitoring data and the historical water supply operation related monitoring data by using a text embedding model, and finding the historical water supply operation related monitoring data of the same category as the real-time water supply operation related monitoring data; querying the scheduling decision set based on the historical water supply operation related monitoring data of the same category as the real-time water supply operation related monitoring data, and obtaining the corresponding scheduling decision as the scheduling decision for the real-time water supply operation related monitoring data, as a retrieval result.

[0035] Among them, the text embedding model adopts an unsupervised clustering model, including but not limited to the K-means model, the self-organizing map model, the neighbor propagation algorithm model, the hierarchical clustering model, the mean shift algorithm model, the DBSCAN model, the binary K-means model, etc. The method flow for inferring the retrieval results is as follows Figure 3 , the specific steps are as follows: S31, collect the real-time water supply operation related data of the urban and rural water supply system obtained in step S10 and the historical water supply operation related data set obtained in step S20 At the same time, the text embedding model is input for clustering analysis, and the final clustering results can compare the categories to which each sub-collection in the collection belongs.

[0036]

[0037] in, is the clustering category set; Represents a collection of real-time water supply operation related data Clustering categories; Represents a collection of historical water supply operation related data Cluster categories of each subset in the cluster. The cluster category may be 2 to multiple categories, so the categories in each subset in the cluster category set may be the same or different. Real-time water supply operation related data set , historical water supply operation related data collection and the set of cluster categories It is a bijective relation, . Cluster category set For use in step S32.

[0038] S32. Clustering by category set Find the scheduling decisions of the same cluster category as the retrieval result. The calculation process is as follows:

[0039] Because real-time water supply operation related data collection and the set of cluster categories It is a bijective relationship, so we can find the data set related to real-time water supply operation Clustering categories .pass and subcollections Intersection finds the same cluster category Because the historical water supply operation data collection and the set of cluster categories It is a bijective relationship, so we can find the same clustering category A collection of historical water supply operation related data Because the historical water supply operation related data set and the optimized scheduling decision set are bijective, the mapping set of monitoring data and scheduling decisions in step S20 is , we can find the optimal scheduling decision subset of the same cluster category as the retrieval result , for use in step S40.

[0040] S40, inputting the search results and the auxiliary decision of the previous stage into the large language model to infer the scheduling auxiliary decision corresponding to the real-time water supply operation related monitoring data.

[0041] The previous stage auxiliary decision refers to the scheduling decision that was released before the current time point. The large language model includes but is not limited to GPT, LaMDA, LLaMA, Gemini, Wenxin Yiyan, Baichuan, Xinghuo, Tongyi Qianwen, and Hunyuan. The scheduling auxiliary decision is inferred by retrieving the results in step S30. It is input into the large language model at the same time as the auxiliary decision in the previous stage, and finally the scheduling auxiliary decision can be inferred through the analysis of the large language model.

[0042] S50: Scheduling the water supply pump group based on the scheduling auxiliary decision.

[0043] Among them, an example of scheduling auxiliary decision inferred in S40 is that the water level in a water plant's clear water tank is expected to be 4.5 meters in the next 15 minutes; the water pressure of the water plant's outlet water needs to be increased by 0.08Mpa, and the water pressure at the water point needs to be increased by 0.05MPa; the No. 1 pump group of the water plant and the No. 1 pump group of the intermediate booster pump station need to be pressurized immediately; in order to meet the water demand and water pressure of 6 cubic meters per second in the next 15 minutes.

[0044] The pump group is dispatched according to the inferred dispatch auxiliary decision.

[0045] See also Figure 4 , provides a scheduling decision system for urban and rural water supply systems based on a large language model of the present invention, comprising: Acquisition module 401, used to acquire real-time water supply operation-related monitoring data and predicted water demand data of urban and rural water supply systems; Establishing module 402, for obtaining historical water supply operation-related monitoring data and water demand data of the urban and rural water supply system, and establishing a scheduling decision set based on the historical water supply operation-related monitoring data and water demand data, with water demand and minimum energy consumption of the pump group as the target, wherein the scheduling decision set includes a mapping relationship between the historical water supply operation-related monitoring data and the scheduling decision; The clustering module 403 is used to cluster the real-time water supply operation related monitoring data and the historical water supply operation related monitoring data using a text embedding model, and find the historical water supply operation related monitoring data of the same category as the real-time water supply operation related monitoring data; and query the scheduling decision set according to the historical water supply operation related monitoring data of the same category as the real-time water supply operation related monitoring data, and obtain the corresponding scheduling decision as the scheduling decision of the real-time water supply operation related monitoring data as a search result; The inference module 404 is used to input the search results and the auxiliary decision of the previous stage into the large language model to infer the scheduling auxiliary decision corresponding to the real-time water supply operation related monitoring data; The scheduling module 405 is used to schedule the water supply pump group based on the scheduling auxiliary decision.

[0046] It can be understood that the scheduling decision system for urban and rural water supply systems based on a large language model provided by the present invention corresponds to the scheduling decision method for urban and rural water supply systems based on a large language model provided in the aforementioned embodiments. The relevant technical features of the scheduling decision system for urban and rural water supply systems based on a large language model can refer to the relevant technical features of the scheduling decision method for urban and rural water supply systems based on a large language model, which will not be repeated here.

[0047] The scheduling decision method and system of the urban and rural water supply system based on the large language model provided by the embodiment of the present invention have the following beneficial effects: (1) The semantic retrieval of the existing general large language model mainly stays at the matching level, and cannot efficiently achieve semantic understanding and question-answer summarization. The generated results have certain illusions and deviations. The present invention combines the retrieval augmented generation technology (RAG) with the large language model to further improve the efficiency of knowledge acquisition and knowledge utilization in the decision-making assistance of highly professional water supply scheduling conditions. In addition, the technology has the advantages of moderate cost, good effect and low technical difficulty.

[0048] (2) Traditional water supply scheduling decisions are issued according to agreed specifications, but the relevant specification information is outdated or too little information, resulting in poor handling efficiency after the decision is made. The present invention can review the obsolescence problem of existing solutions in real time by optimizing the decision set. Through expert manual experience or model analysis, it is possible to review the history and determine which decision should be issued most correctly at the moment. In the face of changes in the vast urban and rural water supply systems, suburban urbanization, and population migration, the latest water supply scheduling decisions can be used.

[0049] (3) The method of auxiliary decision-making for water supply scheduling conditions that couples retrieval enhancement generation technology and large language models can achieve three advantages. The first is automated auxiliary decision-making, which has immediate and rapid response. The second is the function of advance auxiliary decision-making, which enables the disposal work to be advanced and reduces the occurrence of insufficient water pressure or water outage for the most disadvantaged users at the end of the water plant. The third is the advanced technical advantage that can optimize the problems caused by auxiliary decision-making errors in various situations.

[0050] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0051] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0052] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0053] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0054] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0055] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0056] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A scheduling decision method for urban and rural water supply systems based on a large language model, characterized in that: include: S10, obtaining real-time water supply operation-related monitoring data and predicted water demand data of urban and rural water supply systems; S20, obtaining historical water supply operation-related monitoring data and water demand data of the urban and rural water supply system, and establishing a scheduling decision set based on the historical water supply operation-related monitoring data and water demand data, with water demand and minimum energy consumption of the pump group as the target, wherein the scheduling decision set includes a mapping relationship between the historical water supply operation-related monitoring data and the scheduling decision; S30, clustering the real-time water supply operation-related monitoring data and the historical water supply operation-related monitoring data using a text embedding model, and finding the historical water supply operation-related monitoring data of the same category as the real-time water supply operation-related monitoring data; querying the scheduling decision set based on the historical water supply operation-related monitoring data of the same category as the real-time water supply operation-related monitoring data, and obtaining the corresponding scheduling decision as the scheduling decision of the real-time water supply operation-related monitoring data as a search result; S40, inputting the search results and the auxiliary decision of the previous stage into the large language model to infer the scheduling auxiliary decision corresponding to the real-time water supply operation related monitoring data; S50: Scheduling the water supply pump group based on the scheduling auxiliary decision.

2. The scheduling decision method for urban and rural water supply systems based on a large language model according to claim 1 is characterized in that: The real-time water supply operation-related monitoring data refers to the water level of the water plant's clear water tank, the water plant's outlet water pressure, the booster station's outlet water pressure, and the water pressure at the water point, which are transmitted back by the on-site water condition equipment through the Internet of Things. The predicted water demand data refers to the predicted water demand data for urban and rural areas from now to a set time period in the future obtained through statistical methods or AI models; The format and content of the real-time water supply operation-related monitoring data and predicted water demand data are as follows: in, To collect relevant data for real-time water supply operation, Representative This type of data is real-time water supply operation related monitoring data or predicted water demand data.

3. The scheduling decision method for urban and rural water supply systems based on a large language model according to claim 2 is characterized in that: In S20, based on the historical water supply operation-related monitoring data and water demand data, a scheduling decision set is established with water demand and minimum energy consumption of the pump group as the target, including: S21, obtaining historical water supply operation-related monitoring data and water demand data of the urban and rural water supply system, and constructing a historical data set, wherein the historical data set is constructed by constructing a plurality of water supply operation-related monitoring data and water demand data at different historical time points obtained through a historical process, and the format and content of the water supply operation-related monitoring data and water demand data at a single historical time point in the historical data set are consistent with the format and content of the real-time water supply operation-related monitoring data and predicted water demand data, and the historical data set is expressed as: in, A collection of historical water supply operation related data. Representative A subset of historical moments, Representative The first historical moment Kind of data; S22, according to the historical water supply operation related monitoring data and water demand data at each time point, the scheduling decision at each time is determined based on the model analysis to achieve the water demand target and the minimum energy consumption of the pump group, and finally a set of optimized scheduling decisions at multiple different historical time points is formed, which is expressed as: is the scheduling decision set, represents the scheduling decision subset at the mth moment, Representative The first one in the scheduling decision subset at the time Messages; S23, based on the historical water supply operation related monitoring data and the scheduling decision subset corresponding to the time point, a mapping relationship between the historical water supply operation related monitoring data and the scheduling decision at each historical time point is constructed to obtain the historical water supply operation related monitoring data-scheduling decision set: in, It is a set of monitoring data related to historical water supply operation and scheduling decision; Representative A collection of historical water supply operation related data at each moment ; Representative The scheduling decision subset of the pen time , the historical water supply operation related data set and the scheduling decision set are bijective relations .

4. The scheduling decision method for urban and rural water supply systems based on a large language model according to claim 3 is characterized in that: S22, according to the historical water supply operation related monitoring data and water demand data at each time point, determines the scheduling decision at each time based on model analysis, including: Through the hydraulic model, in different preset scheduling decisions, after using the pump group to pressurize at different water outlet pressures and water pressures at water points, it is simulated whether the water demand target and the minimum energy consumption of the pump group can be achieved, and the pump group scheduling decision that achieves the water demand target and the minimum energy consumption of the pump group is obtained.

5. The scheduling decision method for urban and rural water supply systems based on a large language model according to claim 3 is characterized in that: The number of messages in the scheduling decision subset at each moment varies according to the time point, water level in the water plant's clear water tank, outlet water pressure, water pressure at the water point, pump group performance, pressurization location, and water demand data.

6. The scheduling decision method for urban and rural water supply systems based on a large language model according to claim 1 is characterized in that: The text embedding model adopts an unsupervised clustering model, including a K-means model, a self-organizing map model, a neighbor propagation algorithm model, a hierarchical clustering model, a mean shift algorithm model, a DBSCAN model or a binary K-means model.

7. The scheduling decision method for urban and rural water supply systems based on a large language model according to claim 3 is characterized in that: The S30, clustering the real-time water supply operation related monitoring data and the historical water supply operation related monitoring data using a text embedding model, and finding the historical water supply operation related monitoring data of the same category as the real-time water supply operation related monitoring data, includes: S31, collecting the real-time water supply operation related data and the historical water supply operation related data set At the same time, the text embedding model is input for clustering analysis, and the clustering results are expressed as: in, is the clustering category set; Represents a collection of real-time water supply operation related data Clustering categories; Represents a collection of historical water supply operation related data The clustering categories of the subsets at different historical time points in the real-time water supply operation related data set , historical water supply operation related data collection and the set of cluster categories It is a bijective relation, ; S32, by clustering category set Find the real-time water supply operation related data set Clustering categories ,pass and subcollections Intersection finds the same cluster category , ∈ , find the same cluster categories Corresponding historical water supply operation related data set .

8. The scheduling decision method for urban and rural water supply systems based on a large language model according to claim 1 is characterized in that: The previous stage auxiliary decision refers to the scheduling decision that has been released before the current time point, and the large language models include GPT, LaMDA, LLaMA, Gemini, Wenxin Yiyan, Baichuan, Xinghuo, Tongyi Qianwen and Hunyuan.

9. A scheduling decision system for urban and rural water supply systems based on a large language model, characterized in that: include: An acquisition module is used to obtain real-time water supply operation-related monitoring data and predicted water demand data of urban and rural water supply systems; Establish a module for obtaining historical water supply operation-related monitoring data and water demand data of urban and rural water supply systems, and establish a scheduling decision set based on the historical water supply operation-related monitoring data and water demand data, with water demand and minimum energy consumption of pump groups as targets, wherein the scheduling decision set includes a mapping relationship between the historical water supply operation-related monitoring data and the scheduling decision; A clustering module, used to cluster the real-time water supply operation related monitoring data and the historical water supply operation related monitoring data using a text embedding model, and find the historical water supply operation related monitoring data of the same category as the real-time water supply operation related monitoring data; A retrieval module, configured to query the scheduling decision set according to the historical water supply operation related monitoring data of the same category as the real-time water supply operation related monitoring data, and obtain the corresponding scheduling decision as the scheduling decision of the real-time water supply operation related monitoring data as the retrieval result; An inference module, used to input the search results and the auxiliary decision of the previous stage into the large language model, and infer the scheduling auxiliary decision corresponding to the real-time water supply operation related monitoring data; The scheduling module is used to schedule the water supply pump group based on the scheduling auxiliary decision.

Images