Human-computer interaction method, device and system of energy storage monitoring system
Through the agent, the user's problems described in natural language in the energy storage monitoring system are received, the prompt word template is determined, the data analysis tools are encapsulated, and the Q&A results are obtained. The complex operation of the energy storage monitoring system is solved, and the user's operation efficiency and learning cost are improved. .
Patent Information
- Application Number
- CN202410913507.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-08
- Publication Date
- 2024-12-03
AI Technical Summary
The existing energy storage monitoring system is complex in operation, and users need to use the system functions in a fixed mechanical manner, resulting in high learning costs and low operation efficiency.
Call the functions of the energy storage monitoring system through natural language description, use the interactive function interface, inference decision module, template library, tool execution module and front-end processing module in the agent to receive user problems, determine the prompt word template, encapsulate data analysis tools, and obtain Question and answer results are provided to users.
It enables users to obtain problem results without manually operating each functional module. It is simple and convenient to operate, reduces user learning costs and improves operation efficiency.
Smart Images

Figure CN119066160A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the technical field of energy storage monitoring systems, and in particular to a human-computer interaction method, device and system for an energy storage monitoring system. Background Art
[0002] The current energy storage monitoring system is built according to traditional IT technology. The product design generally includes functional modules such as large screen, overview of new energy power station, data monitoring, data analysis, equipment monitoring, energy storage battery analysis, energy storage temperature field, alarm management, and system management. The functions and data rendering methods of these functional modules are pre-developed by software engineers, making the energy storage monitoring system a pre-designed, modular big data analysis system. Users need to use the system functions in a fixed and mechanical manner according to the operating manual. The operation is complicated and the user's learning cost is high. Summary of the invention
[0003] In view of this, the purpose of the present disclosure is to provide a human-computer interaction method, device and system for an energy storage monitoring system, which can call the functions of the energy storage monitoring system through natural language description, is simple and convenient to operate, and reduces the user's learning cost.
[0004] In a first aspect, an embodiment of the present disclosure provides a human-computer interaction method for an energy storage monitoring system, which includes receiving a first user question input by a user; determining whether a first prompt word template corresponding to the first user question exists in a preset template library, and obtaining the first prompt word template; wherein the template library is preset with prompt word templates of multiple task types; a data analysis tool in the energy storage monitoring system is encapsulated in the prompt word template; the first user question is filled into the first prompt word template, a first tool is determined from the data analysis tool encapsulated in the first prompt word template, and a first tool configuration parameter corresponding to the first tool is obtained from the first user question; based on the first tool and the first tool configuration parameter, a first question and answer result corresponding to the first user question is obtained from the energy storage monitoring system; and the first question and answer result is provided to the user.
[0005] In a second aspect, an embodiment of the present disclosure provides a human-computer interaction device for an energy storage monitoring system, the device includes an intelligent agent, the intelligent agent includes: an interactive function interface, an inference decision module, a template library, a tool execution module and a front-end processing module; the interactive function interface is used to: receive a first user question input by a user; the inference decision module is used to: determine whether a first prompt word template corresponding to the first user question exists in a preset template library, and obtain the first prompt word template; wherein the template library is preset with prompt word templates of multiple task types; the prompt word template encapsulates a data analysis tool in the energy storage monitoring system; the inference decision module is also used to: fill the first user question into the first prompt word template, determine the first tool from the data analysis tool encapsulated in the first prompt word template, and obtain the first tool configuration parameters corresponding to the first tool from the first user question; the tool execution module is used to: based on the first tool and the first tool configuration parameters, obtain a first question and answer result corresponding to the first user question from the energy storage monitoring system; the front-end processing module is used to: feed back the first question and answer result to the interactive function interface, and provide the first question and answer result to the user through the interactive function interface.
[0006] In a third aspect, an embodiment of the present disclosure provides a human-computer interaction system for an energy storage monitoring system, the system includes an intelligent agent, and the intelligent agent is communicatively connected with the energy storage monitoring system; the intelligent agent includes: an interactive function interface, an inference decision module, a template library, a tool execution module and a front-end processing module; the interactive function interface is used to receive a first user question input by a user; the inference decision module is used to: determine whether there is a first prompt word template corresponding to the first user question in a preset template library, and obtain the first prompt word template; wherein the template library is preset with prompt word templates of multiple task types; the prompt word template encapsulates a data analysis tool in the energy storage monitoring system; the inference decision module is also used to: fill the first user question into the first prompt word template, determine the first tool from the data analysis tool encapsulated in the first prompt word template, and obtain the first tool configuration parameters corresponding to the first tool from the first user question; the tool execution module is used to: based on the first tool and the first tool configuration parameters, obtain a first question and answer result corresponding to the first user question from the energy storage monitoring system; the front-end processing module is used to: feed back the first question and answer result to the interactive function interface, and provide the first question and answer result to the user through the interactive function interface.
[0007] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium, wherein instructions are stored in the computer-readable storage medium, and when the computer-readable storage medium is run on a computer, the computer executes the above-mentioned human-computer interaction method of the energy storage monitoring system.
[0008] The embodiments of the present disclosure bring the following beneficial effects:
[0009] The human-computer interaction method, device and system of the energy storage monitoring system include: receiving a first user question input by a user; determining whether a first prompt word template corresponding to the first user question exists in a preset template library, and obtaining the first prompt word template; wherein the template library presets prompt word templates of multiple task types; the prompt word template encapsulates a data analysis tool in the energy storage monitoring system; filling the first user question into the first prompt word template, determining a first tool from the data analysis tool encapsulated in the first prompt word template, and obtaining a first tool configuration parameter corresponding to the first tool from the first user question; based on the first tool and the first tool configuration parameter, obtaining a first question and answer result corresponding to the first user question from the energy storage monitoring system; and providing the first question and answer result to the user.
[0010] In this method, after receiving the user question input by the user through the interactive function interface provided by the intelligent body, the prompt word template corresponding to the user question is obtained from the template library, the user question is filled into the corresponding prompt word template, and based on the filled prompt word template, the corresponding data analysis tool in the energy storage monitoring system encapsulated by the prompt word template is called, and the question and answer result corresponding to the user question is obtained from the energy storage monitoring system according to the tool configuration parameters corresponding to the data analysis tool, and the result is fed back to the interactive function interface. In this method, the user does not need to mechanically perform manual operations according to the various preset function modules of the energy storage monitoring system to obtain the desired question results. The user only needs to input the user question described in natural language to directly call the corresponding data analysis tool in the energy storage monitoring system and obtain the corresponding result from the energy storage monitoring system. The operation is simple and convenient, which reduces the user's learning cost and improves the operation efficiency.
[0011] Other features and advantages of the present disclosure will be described in the following description, and partly become apparent from the description, or understood by practicing the present disclosure. The purpose and other advantages of the present disclosure are realized and obtained by the structures particularly pointed out in the description, claims and drawings.
[0012] In order to make the above-mentioned objectives, features and advantages of the present disclosure more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0013] In order to more clearly illustrate the specific embodiments of the present disclosure or the technical solutions in the prior art, the drawings required for use in the specific embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without paying any creative work.
[0014] Figure 1A flowchart of a human-computer interaction method for an energy storage monitoring system provided in an embodiment of the present disclosure;
[0015] Figure 2 An operational diagram of a process for fine-tuning a large language model provided by an embodiment of the present disclosure;
[0016] Figure 3 A schematic diagram of a workflow of an energy storage LLM model provided in an embodiment of the present disclosure;
[0017] Figure 4 A schematic diagram of the workflow of another energy storage LLM model provided in an embodiment of the present disclosure;
[0018] Figure 5 A schematic diagram of the architecture of an intelligent agent technology provided by an embodiment of the present disclosure;
[0019] Figure 6 A schematic diagram of the structure of a human-computer interaction device for an energy storage monitoring system provided in an embodiment of the present disclosure;
[0020] Figure 7 A schematic diagram of the structure of a human-computer interaction system of an energy storage monitoring system provided in an embodiment of the present disclosure. DETAILED DESCRIPTION
[0021] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution of the present disclosure will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present disclosure, rather than all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present disclosure.
[0022] The current energy storage monitoring system is built according to traditional IT technology. The product design generally includes functional modules such as large screen, overview of new energy power station, data monitoring, data analysis, equipment monitoring, energy storage battery analysis, energy storage temperature field, alarm management, and system management. The functions and data rendering methods of these functional modules are pre-developed by software engineers. This is a pre-designed, modular big data analysis system. Users can only use the functions of the energy storage monitoring system in a fixed and mechanical manner according to the operating manual. The operation is complicated and the user's learning cost is high.
[0023] Based on this, the human-computer interaction method, device and system of an energy storage monitoring system provided in the embodiments of the present disclosure can be applied to the human-computer interaction scenario of the energy storage monitoring system.
[0024] In order to facilitate the understanding of this embodiment, a human-computer interaction method of an energy storage monitoring system disclosed in an embodiment of the present disclosure is described in detail. Figure 1As shown, the human-computer interaction method of the energy storage monitoring system includes the following steps:
[0025] Step S102, receiving a first user question input by a user;
[0026] The intelligent agent can be understood as a processor that can quickly identify and respond to user questions. The intelligent agent is connected to the energy storage monitoring system. After receiving the user's question, the intelligent agent can use the intelligent agent interaction function to directly call the corresponding data analysis tool in the energy storage monitoring system according to the user's question and obtain the question and answer results matching the user's question from the energy storage monitoring system.
[0027] The agent can run on a local terminal device, cloud or edge device. For example, the agent can run on a server or edge device with a GPU. In actual implementation, the interactive function interface of the agent can be displayed on a mobile terminal, PC (personal computer) or web terminal and other related devices that can be visualized. The user can input the user question into the interactive function interface provided by the agent. The agent receives the user question input by the user through the interactive function interface, responds to the user question, and then feeds back the question and answer results that match the user question to the interactive function interface for display to the user.
[0028] In one approach, the intelligent agent also includes a reasoning and decision-making module, a template library, a tool execution module, and a front-end processing module. The reasoning and decision-making module can be constructed through a large language model (LLM), retrieval augmented generation technology (RAG technology), and intelligent agent technology (AI Agent). It is the decision-making center of the intelligent agent and can be compared to the human brain. The reasoning model is pre-trained with relevant information from the energy storage industry and can understand and process the professional terms and energy storage technology data of the energy storage industry to make effective decisions. Specifically, the construction process of the reasoning and decision-making module is as follows:
[0029] 1. Use an open-source large language model as the base model in the reasoning decision module. Then, collect energy storage-related data to fine-tune the large language model to obtain a large language model with energy storage domain knowledge, namely the energy storage LLM model. The fine-tuning process is as follows: Figure 2As shown, here, energy storage related information includes but is not limited to: energy storage technology agreement, energy storage technology plan, energy storage topology, battery data, PCS (Power Conversion System, energy storage converter) data, energy storage knowledge base (including operation and maintenance knowledge base) and other energy storage documents and data. Among them, battery data includes basic battery information, voltage, temperature, current and other telemetry data, as well as battery fault data. PCS data includes PCS fault and PCS operation and maintenance data.
[0030] 2. Design workflow 1 based on the energy storage LLM model, such as Figure 3 shown.
[0031] In this workflow 1, the starting point is "user questions" and the end is to get "satisfactory answers", which may go through multiple rounds of question-answering optimization in the middle.
[0032] 3. Based on the LLM model of energy storage, RAG technology is introduced.
[0033] In order to constrain the output of the energy storage LLM model and obtain more professional results, the second workflow is introduced based on RAG technology, such as Figure 4 shown.
[0034] In this workflow, the preset knowledge repository is first retrieved according to the user's question. Once the historical data with a high degree of similarity to the user's question is matched, it is filled into the corresponding prompt word template of the user's question together with the user's question. The prompt word template is used to guide the energy storage LLM model to generate a specific output. It can help the energy storage LLM model understand the user's needs by providing contextual information and instructions, thereby generating more accurate and relevant text. Here, the energy storage LLM model is guided to infer the question and answer results according to the filled prompt word template. If the user is satisfied with the feedback of the generated question and answer results, the user question and the question and answer results are stored in the knowledge repository in association with each other. The above-mentioned knowledge repository can also be a relational repository, a non-relational repository, a document repository, or a vector repository, without specific limitation.
[0035] 4. Based on the energy storage LLM model with intelligent agent technology and RAG technology, a reasoning and decision-making module is constructed. The architecture of the intelligent agent technology in the reasoning and decision-making module is as follows: Figure 5 shown.
[0036] Intelligent agent technology can combine the memory module and the planning module to independently decide whether to split the user problem into multiple sub-problems for separate processing based on the complexity of the user problem, and select the corresponding tool in its own tool set for calling. Based on the feedback of the call result, the intelligent agent will reason again to decide the next action. After a series of actions, the result will be finally fed back to the user.
[0037] The reasoning and decision-making module constructed in the above manner has powerful computing and analytical reasoning capabilities, making the reasoning and decision-making module the decision-making center of the intelligent agent.
[0038] The template library in the intelligent agent has preset prompt word templates for various task types. The prompt word templates encapsulate the data analysis tools of the energy storage monitoring system. The prompt word templates of each task type are used to help the reasoning decision module to reason about specific user problems, so that the reasoning decision module can determine the data analysis tools and tool configuration parameters that can be used to solve the user problem from the data analysis tools of the energy storage monitoring system encapsulated by the prompt word template according to the corresponding prompt word template. Other templates can also be stored in the template library in the intelligent agent, for example, chart style templates for meeting various data display requirements, task decomposition templates for decomposing user problems, etc. The tool execution module in the intelligent agent can be used to call the data analysis tool to obtain the question and answer results matching the user problem from the energy storage monitoring system according to the data analysis tools and tool configuration parameters determined by the reasoning decision module. The front-end processing module in the intelligent agent can be used to render the question and answer results and feed them back to the interactive function interface.
[0039] In this step, the user can input the first user question through the interactive function interface provided by the agent, and after receiving the first user question through the interactive function interface, the agent can quickly respond to the first user question. Here, the first user question is a question described by the user in natural language, and the first user question can be input in a variety of ways such as text input or voice input.
[0040] Step S104, determining whether a first prompt word template corresponding to the first user question exists in a preset template library, and obtaining the first prompt word template; wherein the template library presets prompt word templates of multiple task types; and the prompt word template encapsulates a data analysis tool in the energy storage monitoring system;
[0041] The above-mentioned data analysis tools include: functions for implementing the functions of each functional module of the energy storage monitoring system, methods for accessing external APIs and other AI algorithms.
[0042] Here, after the intelligent agent receives the first user question through the interactive function interface, it sends the first user question to the reasoning and decision-making module in the intelligent agent. First, it is necessary to determine through the reasoning and decision-making module in the intelligent agent whether there is a first prompt word template corresponding to the first user question in the preset template library, and obtain the first prompt word template.
[0043] Specifically, after receiving the first user question, the reasoning decision module first searches the knowledge repository preset by the agent to see whether there is a historical question that is the same as the first user question. Here, the knowledge repository stores historical user questions that have been processed by the agent and recognized by the user and the historical question-and-answer results that match them. If a historical user question that is the same as the first user question is retrieved from the knowledge repository, the reasoning decision module can determine that there is a first prompt word template corresponding to the first user question in the template library, and determine the prompt word template corresponding to the historical user question as the first prompt word template. If there is no historical question that is the same as the first user question, it is necessary to determine the first prompt word template corresponding to the first user question through the intention recognition template in the model library. Here, the intention recognition template is used to help the reasoning decision module identify and process the intention of the user question, so as to determine the prompt word template applicable to the first user question. The reasoning decision module fills the first user question into the intention recognition template. The reasoning decision module can identify the prompt word template that can solve the first user question based on the intention recognition template, and determine that there is a first prompt word template corresponding to the first user question in the template library.
[0044] Further, after determining that the first prompt word template corresponding to the first user question exists in the preset template library, the reasoning decision module obtains the first prompt word template from the template library.
[0045] Step S106, filling the first user question into the first prompt word template, and based on the first prompt word template filled with the first user question, determining a first tool from the data analysis tools encapsulated by the first prompt word template, and a first tool configuration parameter matching the first tool;
[0046] The above-mentioned first tool configuration parameters refer to the tool configuration parameters that need to be input into the first tool, which are inferred or directly extracted from the first user question by the reasoning decision module in combination with the parameter attribute requirements of the first tool. For example, if the first tool is a power plant alarm management tool, the first tool configuration parameters may include the power plant ID information and time information obtained from the first user question. For another example: if the first tool is a battery status monitoring tool, the first tool configuration parameters may include the battery ID information, time information, and battery parameter information to be monitored obtained from the first user question.
[0047] Here, after the reasoning and decision module obtains the first prompt word template from the template library, it fills the first user question into the first prompt word template. The reasoning and decision module will analyze and infer the first prompt word template filled with the first user question to obtain the first tool and the first tool configuration parameters corresponding to the first tool, so as to obtain the first question and answer result from the energy storage monitoring system through the first tool and the first tool configuration parameters.
[0048] Specifically, the reasoning and decision module pre-identifies and understands the achievable functions of the data analysis tool encapsulated by the first prompt word template. After the first user question is filled into the first prompt word template, the reasoning and decision module will perform semantic understanding of the first user question to obtain user needs. Based on the user needs, the reasoning and decision module will infer the first tool that can realize the function of the user needs from the data analysis tool encapsulated by the first prompt word template. At the same time, the reasoning and decision module will also obtain the first tool configuration parameters corresponding to the first tool from the first user question.
[0049] For example, the first user question is "Query the number of alarms of the first power station on April 5th". The first user question is filled into the first prompt word template. The reasoning and decision module performs semantic understanding of the first user question to determine that the user needs to query the number of alarms of the power station. It is inferred from the data analysis tool encapsulated in the first prompt word template that the tool that can realize the query of the number of power station alarms is the power station alarm management tool. The reasoning and decision module determines the power station alarm management tool as the first tool. At the same time, combined with the parameter information required by the power station alarm management tool, the reasoning and decision module extracts "first power station" and "April 5th" from the first user question as the first tool configuration parameters.
[0050] Step S108: Based on the first tool and the first tool configuration parameters, obtain a first question-and-answer result corresponding to the first user question from the energy storage monitoring system; and provide the first question-and-answer result to the user.
[0051] The above-mentioned first question and answer result can be presented in the form of reports, links, charts, etc.
[0052] After the reasoning and decision module obtains the first tool and the first tool configuration parameters, the tool execution module will call the first tool, obtain the corresponding data from the energy storage monitoring system according to the first tool configuration parameters, and use the obtained data as the question and answer result of the first user question. For example, the reasoning and decision module determines that the first tool is a power station alarm management tool, and the first tool configuration parameters are: "First Power Station", "April 5th". After that, the reasoning and decision module sends the first tool and the first tool configuration parameters to the tool execution module in the intelligent agent. The tool execution module can call the power station alarm management tool to query the number of alarms of the first power station on April 5th according to the first tool configuration parameters, and directly use it as the first question and answer result. In another way, the tool execution module calls the power plant alarm management tool to query the number of alarms of the first power plant on April 5 according to the first tool configuration parameters, which can only be used as an initial question and answer result. The reasoning and decision module can also fill this initial question and answer result into the first prompt word template after the first user question is filled in, and obtain the data summary result corresponding to the initial question and answer result. This data summary result has richer content than the initial question and answer result. Finally, the integration module in the intelligent body integrates the data summary result and the initial question and answer result to obtain the first question and answer result corresponding to the first user question. The first question and answer result obtained in this way is richer in content.
[0053] Finally, the front-end processing module in the intelligent agent feeds back the first question and answer result to the interactive function interface, and presents it in the visualization interface provided by the visualization device for the user to view.
[0054] The human-computer interaction method of the energy storage monitoring system described above receives a first user question input by a user; determines whether a first prompt word template corresponding to the first user question exists in a preset template library, and obtains the first prompt word template; wherein the template library presets prompt word templates of multiple task types; the prompt word template encapsulates a data analysis tool in the energy storage monitoring system; fills the first user question into the first prompt word template, determines a first tool from the data analysis tool encapsulated in the first prompt word template, and obtains a first tool configuration parameter corresponding to the first tool from the first user question; based on the first tool and the first tool configuration parameter, obtains a first question and answer result corresponding to the first user question from the energy storage monitoring system; and provides the first question and answer result to the user.
[0055] In this method, after receiving the user question input by the user, the prompt word template corresponding to the user question is obtained from the template library, the user question is filled into the corresponding prompt word template, based on the filled prompt word template, the corresponding data analysis tool in the energy storage monitoring system encapsulated by the prompt word template is called, and according to the tool configuration parameters obtained from the user question, the question and answer result corresponding to the user question is obtained from the energy storage monitoring system, and the result is provided to the user. In this method, the user does not need to mechanically perform manual operations according to the various preset functional modules of the energy storage monitoring system to obtain the desired question results. The user only needs to enter the user question described in natural language to directly call the corresponding data analysis tool in the energy storage monitoring system and obtain the corresponding result from the energy storage monitoring system. The operation is simple and convenient, which reduces the user's learning cost and improves the operation efficiency.
[0056] In one method, a search is performed from a preset knowledge repository to determine whether there is first historical data; wherein the first historical data includes: a first historical question identical to a first user question and a first historical question-and-answer result associated with the first historical question; the preset knowledge repository stores mutually associated historical questions and historical question-and-answer results, and the historical questions and historical question-and-answer results match each other; if the first historical data exists, it is determined that there is a first prompt word template corresponding to the first user question in the template library; wherein the first prompt word template is a prompt word template corresponding to the first historical question. If the first historical data does not exist, it is determined that there is a first prompt word template in the template library through an intent recognition template in the template library; wherein the intent recognition template contains functional descriptions of preset prompt word templates of multiple task types.
[0057] The above-mentioned knowledge repository stores multiple groups of QA pairs, and the QA pair refers to a pair of interrelated historical questions and historical question and answer results. The QA pair has been recognized by the user, and the historical questions and historical question and answer results in the QA pair match each other. The above-mentioned first historical question is a historical user question that is the same as the first user question and is stored in the preset knowledge repository. The above-mentioned intent recognition template contains functional descriptions of preset prompt word templates of multiple task types. The functions in the functional description of each prompt word template correspond to the functions that can be realized by the data analysis tool of the energy storage monitoring system encapsulated by the prompt word template.
[0058] That is, after the agent receives the first user question, the reasoning and decision module in the agent first searches the knowledge repository preset by the agent to see whether there is a historical question that is the same as the first user question. If the historical question that is the same as the first user question is retrieved from the preset knowledge repository, the reasoning and decision module can determine that there is a first prompt word template corresponding to the first user question in the template library. The first prompt word template is the prompt word template used when solving the first historical question that is the same as the first user question.
[0059] If no historical question identical to the first user question is retrieved from the preset knowledge repository, the reasoning decision module may also determine the presence of the first prompt word template in the template library through the intention recognition template in the template library.
[0060] In actual implementation, the reasoning decision module will fill the first user question into the "user input" in the intention recognition template. The reasoning decision module will perform reasoning based on the functional descriptions of the prompt word templates of multiple task types contained in the intention recognition template to determine whether the data analysis tools of the energy storage monitoring system encapsulated by these prompt word templates can solve the first user problem. If it is determined that the data analysis tools in the energy storage monitoring system encapsulated by the prompt word template of a certain task type can solve the first user problem, the reasoning decision module determines the prompt word template of this task type as the first prompt word template.
[0061] In one method, in the process of filling the first user question into the first prompt word template, historical questions similar to the first user question can also be found, and the historical questions and the question and answer results corresponding to the historical questions that are recognized by the user are filled into the first prompt word template, helping the reasoning decision module to find a more accurate first question and answer result that better meets the user's needs.
[0062] Specifically, whether there is second historical data is retrieved from a preset knowledge repository; wherein the second historical data includes: a second historical question whose degree of similarity to the first user question is greater than a preset degree threshold, and a second historical question and answer result corresponding to the second historical question; if there is second historical data, based on the degree of similarity corresponding to the second historical question in the second historical data, a preset number of recommended data is obtained from the second historical data; and the recommended data and the first user question are filled into the first prompt word template.
[0063] The similarity between the second historical question and the first user question can be obtained by matching the semantic similarity through the semantic model, and all historical questions with a similarity greater than a preset threshold are determined as second historical questions. Then, the second historical question and answer results associated with the second historical question are obtained from the knowledge repository, and each second historical question and its associated second historical question and answer results are used as a second historical data. The second historical data are arranged in descending order according to the similarity between the second historical question and the first user question. The recommended data are the first few second historical data, and the number of recommended data can be set in advance as needed.
[0064] In actual implementation, after determining the first prompt word template, the reasoning decision module will search from the preset knowledge repository to see whether there is a second historical question whose similarity to the first user question is greater than a preset threshold. If so, it means that there is second historical data. The second historical question and answer results associated with the second historical question are obtained from the knowledge repository, and each second historical question and its associated second historical question and answer results are used as a second historical data. Then, a preset number of recommended data are obtained from the second historical data, and the recommended data and the first user question are filled into the first prompt word template together. The reasoning decision module can provide a reference when reasoning about the first prompt word template after the first user question is filled in based on the recommended data, thereby improving the accuracy of the reasoning of the reasoning decision module, and then accurately determining the first tool in the data analysis tool encapsulated in the first prompt word template.
[0065] In one manner, based on the first tool and the first tool configuration parameters, the first question-and-answer result corresponding to the first user question can be obtained from the energy storage monitoring system in the following manner:
[0066] Specifically, based on the first tool and the first tool configuration parameters, an initial question and answer result is obtained from the energy storage monitoring system, the initial question and answer result is filled into the first prompt word template after the first user question is filled, and a data summary result corresponding to the initial question and answer result is obtained. The data summary result and the initial question and answer result are integrated and processed to obtain a first question and answer result corresponding to the first user question.
[0067] Here, after the reasoning and decision module determines the first tool and the first tool configuration parameters, the tool execution module calls the first tool and obtains a question and answer result from the energy storage monitoring system according to the first tool configuration parameters. The question and answer result is used as an initial question and answer result. The initial question and answer result can be understood as a question and answer result that simply lists data information. The information content displayed by the initial question and answer result may not be rich enough. Therefore, after obtaining the initial question and answer result, it is necessary to continue to fill the initial question and answer result into the first prompt word template filled with the first user question. At this time, the reasoning and decision module can use the initial question and answer result as a reference, and further reason according to the first prompt word template filled with the initial question and answer result and the first user question to obtain a richer data summary result. The integration module in the intelligent agent integrates the data summary result and the initial question and answer result as the first question and answer result corresponding to the first user question.
[0068] For example, the first user question is "What are the faults of the first power station in the energy storage monitoring system on May 3?" The tool execution module calls the first tool and obtains the initial question and answer result from the energy storage monitoring system according to the configuration parameters of the first tool, which is "The faults of the first power station in the energy storage monitoring system on May 3 are fault 1, fault 2, fault 3, fault picture 1, fault picture 2, fault picture 3". After obtaining the initial question and answer result, the reasoning decision module continues to fill the initial question and answer result into the first prompt word template filled with the first user question, and performs reasoning again to obtain the data summary result "The faults of the first power station in the energy storage monitoring system on May 3 are fault 1, fault 2, fault 3; fault picture 1, fault picture 2, fault picture 3; fault 1 is a serious level fault; fault 2 is a normal level fault, and fault 3 is a normal level fault." Finally, the data summary results and the initial question and answer results are integrated through the integration module in the intelligent agent to obtain the first question and answer result with smooth semantics. In this embodiment, the first question and answer result is "The faults of the first power station in the energy storage monitoring system on May 3 were Fault 1, Fault 2, and Fault 3, among which the fault levels are as follows: Fault 1 is a serious fault; Fault 2 and Fault 3 are ordinary faults; the fault pictures are summarized as follows: Fault Picture 1, Fault Picture 2, Fault Picture 3".
[0069] In this method, the reasoning decision module calls the first tool through the tool execution module, and continues to fill the initial question and answer result obtained from the energy storage monitoring system according to the configuration parameters of the first tool into the first prompt word template filled with the first user question, so that the reasoning decision module can perform reasoning again according to the initial question and answer result and the first prompt word template filled with the first user question to obtain a data summary result with richer content. Finally, after integrating the data summary result and the initial question and answer result, a first question and answer result with rich content and smooth semantics is obtained, so as to feed back a question and answer result with rich content and smooth semantics to the interactive function interface and display it to the user.
[0070] In one method, after providing the first question and answer result to the user, feedback information from the user regarding the first question and answer result can also be received; in response to receiving positive feedback information, the first question and answer result and the first user question are associated and saved in a preset knowledge repository; the positive feedback information indicates that the first question and answer result matches the first user question.
[0071] Here, after displaying the first question and answer result, the interactive function interface will also provide a feedback interface for obtaining the user's evaluation of the first question and answer result. The above positive feedback information is positive feedback, indicating that the first question and answer result matches the first user question. For example: the feedback interface can be a dialog box, which contains dialog box information and two controls displayed in association with the dialog box information. The dialog box information can be "Are you satisfied with the result?" When the user clicks the "Yes" control, the reasoning and decision module in the intelligent agent determines that the positive feedback information is received, and the first question and answer result and the first user question are associated and saved in the preset knowledge repository, so that when a user question similar to the first user question is received later, the first question and answer result can be used to help the reasoning and decision module find a more accurate question and answer result that meets the user's needs more quickly.
[0072] In one method, if the data analysis tool of the energy storage monitoring system cannot directly solve the user question after the interactive function interface receives the user question input by the user, that is, the prompt word template corresponding to the user question does not exist in the preset template library, the question and answer result can also be obtained in the following way:
[0073] Specifically, a second user question input by a user is received; it is determined that a prompt word template corresponding to the second user question does not exist in a preset template library; the second user question is filled into a task decomposition template in the template library for decomposition, and the second user question is decomposed into multiple sub-questions; it is determined whether there is a prompt word template corresponding to the sub-question in the preset template library, based on the determination result, a template corresponding to the sub-question is determined from the template library, a tool corresponding to the sub-question is determined based on the template corresponding to the sub-question, the tool corresponding to the sub-question is called to obtain a question and answer result corresponding to the sub-question, and the question and answer results corresponding to the multiple sub-questions obtained by decomposition are used as the second question and answer result corresponding to the second user question; the second question and answer result is provided to the user; wherein the type of template corresponding to the sub-question includes a structured query language template or a prompt word template; the tool corresponding to the sub-question includes: a data analysis tool in an energy storage monitoring system, or a structured query tool; the structured query language template includes context information of a preset structured query statement.
[0074] The above-mentioned task decomposition template is used to help the reasoning decision module decompose the second user question into multiple sub-questions, so that the reasoning decision module can obtain the corresponding template for each sub-question respectively, so as to infer the question and answer results corresponding to each sub-question. The database of the above-mentioned energy storage monitoring system stores all data in the energy storage monitoring system. The above-mentioned structured query language template contains the context information of the preset structured query statement. After the sub-question is filled into the structured query language template, the structured query statement corresponding to the sub-question can be generated. According to the structured query statement corresponding to the sub-question, the corresponding data is directly obtained from the database of the energy storage monitoring system. The data is the result corresponding to the sub-question. The tools corresponding to the above-mentioned sub-questions include: data analysis tools in the energy storage monitoring system, or structured query tools.
[0075] Here, after the interactive function interface receives the second user question input by the user, the reasoning decision module retrieves from the knowledge repository preset by the intelligent agent and determines that there is no historical question identical to the second user question. At the same time, the reasoning decision module performs reasoning based on the functional description of the prompt word templates of multiple task types contained in the intention recognition template, and determines that the data analysis tools in the energy storage monitoring system encapsulated by these prompt word templates cannot solve the second user question. At this time, the reasoning decision module determines that there is no prompt word template corresponding to the second user question in the preset template library. In this case, the reasoning decision module will fill the second user question into the task decomposition template in the template library, perform task reasoning planning according to the task decomposition template, and decompose the second user question into multiple sub-questions. Then, the reasoning decision module determines whether there is a prompt word template corresponding to the sub-question in the preset template library, and according to the judgment result, determines the template corresponding to the sub-question from the template library, determines the tool corresponding to the sub-question according to the template corresponding to the sub-question, calls the tool corresponding to the sub-question to obtain the question and answer result corresponding to the sub-question, and summarizes the question and answer result corresponding to each sub-question as the question and answer result of the second user question and feeds it back to the interactive function interface.
[0076] For example, the task decomposition template includes a task description, a task decomposition requirement, and an output subtask list, wherein the user question is filled in the task description, the number of task decomposition requirements is at least 1, and the decomposed subtasks are obtained in the output subtask list. For example, the second user question is "Help me find the alarm information of the first device, the temperature statistics of the battery cell, and the maximum battery cell voltage during the charging process", and the reasoning decision module retrieves from the knowledge repository preset by the intelligent agent to determine that there is no historical question that is the same as the second user question. At the same time, the reasoning decision module performs reasoning based on the functional description of the prompt word templates of multiple task types contained in the intention recognition template, and determines that the data analysis tools in the energy storage monitoring system encapsulated by these prompt word templates cannot solve the second user problem, that is, the reasoning decision module determines that there is no prompt word template corresponding to the second user problem in the preset template library. In this case, the reasoning decision module will fill the second user question into the task decomposition template, and the reasoning decision module will perform task reasoning planning based on the task decomposition template, and decompose the second user question into three sub-questions: "Help me find the alarm information of the first device", "Help me find the temperature statistics of the battery cell", and "Help me find the maximum battery cell voltage during the charging process". Then, the reasoning decision module determines whether there are prompt word templates corresponding to these sub-questions in the preset template library. According to the judgment results, the templates corresponding to each sub-question are determined from the template library, and then the tools corresponding to each sub-question are determined. Then, the tools corresponding to each sub-question are called respectively to obtain the question and answer results corresponding to each sub-question. These question and answer results are summarized as the question and answer results of the second user question and fed back to the interactive function interface, which is provided to the user by the interactive function interface.
[0077] In one method, determining whether there is a prompt word template corresponding to a sub-question in a preset template library can be performed in the following manner: filling the sub-question into the intent recognition template, and based on the functional description of the prompt word template contained in the intent recognition template, determining whether there is a prompt word template applicable to the sub-question.
[0078] The reasoning and decision module fills the sub-problem into the intent recognition template, and performs reasoning based on the functional descriptions of the prompt word templates of multiple task types contained in the intent recognition template to determine whether the data analysis tools of the energy storage monitoring system encapsulated by these prompt word templates can solve the sub-problem. If it is determined that the data analysis tool encapsulated by the prompt word template of a certain task type can solve the sub-problem, the reasoning and decision module determines that there is a prompt word template corresponding to the sub-problem, and determines the prompt word template of the task type as the prompt word template corresponding to the sub-problem. If the data analysis tool in the energy storage monitoring system encapsulated by the prompt word template of any task type in the template library cannot solve the sub-problem, the reasoning and decision module determines that there is no prompt word template corresponding to the sub-problem, and matches the sub-problem to the structured query language template.
[0079] Further, according to the judgment result, the template corresponding to the sub-problem is determined from the template library, the tool corresponding to the sub-problem is determined based on the template corresponding to the sub-problem, and the tool corresponding to the sub-problem is called to obtain the question and answer result corresponding to the sub-problem. Specifically, if there is a prompt word template corresponding to the first question in the sub-problem in the template library; fill the first question into the prompt word template corresponding to the first question; determine the second tool from the data analysis tool in the energy storage monitoring system encapsulated by the prompt word template corresponding to the first question; obtain the second tool configuration parameters corresponding to the second tool from the first question, and obtain the question and answer result corresponding to the first question from the energy storage monitoring system based on the second tool and the second tool configuration parameters; if there is no prompt word template corresponding to the second question in the sub-problem in the template library, fill the second question into the structured query language template in the template library, and generate the structured query statement corresponding to the second question based on the filled structured query language template; call the structured query tool according to the structured query statement corresponding to the second question, and obtain the question and answer result corresponding to the second question from the database of the energy storage monitoring system; use the question and answer result corresponding to the first question and the question and answer result corresponding to the second question as the question and answer result corresponding to the sub-problem.
[0080] That is to say, if the reasoning decision module determines that there is a prompt word template corresponding to the first question in the sub-problem, the first question is filled into the prompt word template corresponding to the first question, and the reasoning decision module analyzes and infers the prompt word template filled with the first question to obtain the second tool and the second tool configuration parameters corresponding to the second tool. The tool execution module calls the second tool and obtains the corresponding data from the energy storage monitoring system according to the second tool configuration parameters as the question and answer result corresponding to the first question. If the prompt word template corresponding to the second question in the sub-problem does not exist in the template library, and the second question is matched to the structured query language template after reasoning by the reasoning decision module, the second question is filled into the structured query language template, and a structured query statement corresponding to the second question is generated according to the filled structured query language template. The structured query tool is called according to the structured query statement to directly obtain data from the database of the energy storage monitoring system, and the data is used as the question and answer result corresponding to the second question.
[0081] It should be noted that if the user question still cannot be matched to any existing template in the template library after a limited number of reasonings, the task is terminated, and the user is given a friendly prompt through the interactive function interface, or the user is asked to provide further information and restart the user question.
[0082] The question-and-answer results corresponding to all the sub-questions obtained after decomposing the second user's question are obtained in the above manner, and the question-and-answer results corresponding to all the sub-questions are fed back to the interactive function interface.
[0083] In this method, when the existing functions of the energy storage monitoring system cannot directly solve the user problem, the reasoning and decision-making module in the intelligent agent can decompose the second user problem into multiple sub-problems through the task decomposition template, and search for the corresponding template for each sub-problem. If the existing functions of the energy storage monitoring system still cannot solve the sub-problem, the structured query language template can also be used to generate a structured query statement to directly obtain the corresponding data from the database of the energy storage monitoring system to obtain the question and answer results.
[0084] In one method, if the second user question contains result rendering information, a first chart style template corresponding to the result rendering information is obtained from a preset template library, the second question and answer result is filled into the first chart style template, the second question and answer result is rendered for the first time based on the rendering data in the first chart style template, and the second question and answer result after the first rendering is provided to the user.
[0085] The above result rendering information is information about chart style and layout, such as words such as list, bar chart, dashboard, etc. The above chart style template contains rendering data for generating forms, reports, charts or dashboards. After the data is filled into the corresponding chart style template, the corresponding form, report, chart or dashboard displaying data information can be generated according to the filled data information and rendering data.
[0086] If the inference decision module recognizes that the user question contains result rendering information, for example, the user question is "the report displays the battery status data and the alarm information of the first device", then after obtaining the battery status data and the alarm information of the first device, the inference decision module obtains a first chart style template from a preset template library, and the first chart style template contains rendering data of the report, fills the battery status data and the alarm information of the first device into the first chart style template, performs a first rendering on the battery status data according to the rendering data of the report in the first chart style template, feeds back the rendered battery status data to the interactive function interface, and provides it to the user through the interactive function interface.
[0087] In another method, if the second user question does not contain result rendering information, the question and answer results corresponding to the sub-question are filled into the preset chart style template, and the question and answer results corresponding to the sub-question are rendered for the second time based on the rendering data in the preset chart style template, and the question and answer results corresponding to the sub-question after the second rendering are provided to the user.
[0088] The above-mentioned preset chart style templates can be pre-classified and set according to the data format of the question and answer results corresponding to the sub-questions. For example: if the data format of the question and answer results corresponding to the sub-questions is a simple combination of text and numbers, the preset chart style template can include rendering data for generating reports.
[0089] That is to say, if the reasoning decision module recognizes that the second user question does not contain result rendering information, the second question and answer result will be filled into the preset chart style template, the second question and answer result will be rendered according to the rendering data in the preset chart style template, and the rendered second question and answer result will be provided to the user.
[0090] In one method, after obtaining the question-and-answer results, you can also communicate with the intelligent agent to continuously fine-tune the data rendering method.
[0091] Specifically, a third user question input by a user is received; the third user question is used to update the second question and answer result; if the third user question contains rendering update information for the first specified sub-question, a second chart style template corresponding to the rendering update information is obtained from a preset template library, the question and answer result corresponding to the first specified sub-question is determined from the second question and answer result, the question and answer result corresponding to the first specified sub-question is filled into the second chart style template, a third rendering is performed on the question and answer result corresponding to the first specified sub-question based on the rendering data in the second chart style template, and the third rendering result is provided to the user.
[0092] The third user question is used to update the second question and answer result. The rendering update information refers to information that changes the layout style or data presentation form of the second question and answer result, such as information containing words such as "bar chart changed to line chart", "table changed to report", etc. The first designated sub-question is the question parameter mentioned in the third user question that requires changing the layout style or data presentation form. The second user question is: "The report displays the battery status data and the alarm information of the first device"; after obtaining the result of the report presentation, the user can enter the third user question: "Change the presented battery status data from a bar chart to a line chart". At this time, the reasoning decision module can recognize that the rendering update information is the change from a bar chart to a line chart, and the first designated sub-question is the battery status data.
[0093] Here, the reasoning decision module can identify the rendering update information in the third user question and the first designated sub-question corresponding to the rendering update information. Then, the reasoning decision module will obtain the second chart style template corresponding to the rendering update information from the preset template library according to the rendering update information. For example, the rendering update information is "bar chart changed to line chart". At this time, the reasoning decision module will obtain the second chart style template corresponding to the line chart from the preset template library. The second chart style template contains the rendering data of the line chart. Then, the reasoning decision module determines the question and answer result of the first designated sub-question from the second question and answer result, and renders the question and answer result of the first designated sub-question according to the rendering data of the line chart in the second chart style template, generates the question and answer result of the first designated sub-question in the form of a line chart, and feeds it back to the interactive function interface and provides it to the user.
[0094] In one method, it is recognized that the third user question contains question editing information for a second designated sub-question, an updated second designated sub-question is determined, a template corresponding to the updated second designated sub-question is determined from a template library, and a question-and-answer result of the updated second designated sub-question is obtained; wherein the type of template corresponding to the updated second designated sub-question includes a structured query language template or a prompt word template; the second question-and-answer result is updated according to the question-and-answer result of the updated second designated sub-question, and provided to the user; wherein the type of template corresponding to the updated second designated sub-question includes a structured query language template or a prompt word template.
[0095] The above-mentioned second designated sub-problem may be a sub-problem after the second user problem is decomposed by the task decomposition template, and the types of templates corresponding to the second designated sub-problem include structured query language templates or prompt word templates. The question editing information of the above-mentioned second designated sub-problem refers to information related to editing operations such as replacement and deletion of the second designated sub-problem. For example, the second user question is: "Display the battery status data and the alarm information of the first device"; after seeing the question and answer results, the user wants to modify the user question, and can directly enter the third user question: "Please replace the battery status data with the battery cell temperature statistics". At this time, the reasoning decision module can recognize that the second designated sub-problem is the "battery status data" in the second user question, and the question editing information is "Replace with battery cell temperature statistics".
[0096] After the reasoning and decision module identifies the question editing information for the second designated sub-question contained in the third user question, the updated second designated sub-question is determined according to the question editing information. For example, if the third user question is "Please replace the battery status data with the battery cell temperature statistics", the reasoning and decision module determines "battery cell temperature statistics" as the updated second designated sub-question. Then, the reasoning and decision module obtains the template corresponding to the updated second designated sub-question from the template library. The template corresponding to the updated second designated sub-question can be a prompt word template or a structured query language template. The reasoning and decision module can continue to infer the updated question and answer result of the second designated sub-question according to the template corresponding to the updated second designated sub-question. Then, the question and answer result of the original second designated sub-question in the second question and answer result is replaced with the question and answer result of the updated second designated sub-question to obtain the updated second question and answer result. The updated second question and answer result is fed back to the interactive function interface and provided to the user. In this way, according to the initially generated question and answer result, the user can continue to communicate with the intelligent agent and continuously fine-tune the data rendering method of the question and answer result to meet various data display requirements.
[0097] Corresponding to the above method embodiment, see Figure 6A human-computer interaction device for an energy storage monitoring system is shown. The human-computer interaction device for the energy storage monitoring system can implement the human-computer interaction method for the energy storage monitoring system through an intelligent agent. The device includes an intelligent agent, and the intelligent agent includes: an interactive function interface, a reasoning decision module, a template library, a tool execution module, and a front-end processing module;
[0098] The above-mentioned interactive function interface is used to receive a first user question input by a user; the above-mentioned reasoning decision module is used to determine whether a first prompt word template corresponding to the first user question exists in a preset template library, and obtain the first prompt word template; wherein the template library presets prompt word templates of multiple task types; the prompt word template encapsulates a data analysis tool in the energy storage monitoring system;
[0099] The above-mentioned reasoning and decision module is further used to fill the first user question into the first prompt word template, determine the first tool from the data analysis tools encapsulated in the first prompt word template, and obtain the first tool configuration parameters corresponding to the first tool from the first user question;
[0100] The tool execution module is configured to obtain a first question-and-answer result corresponding to a first user question from the energy storage monitoring system based on the first tool and the first tool configuration parameters;
[0101] The above-mentioned front-end processing module is used to: feed back the first question and answer result to the interactive function interface, and provide the first question and answer result to the user through the interactive function interface.
[0102] The human-computer interaction device of the energy storage monitoring system provided by the embodiment of the present disclosure includes an intelligent agent. The device does not require the user to mechanically perform manual operations according to various preset functional modules of the energy storage monitoring system to obtain the desired problem results. The user only needs to input the user problem described in natural language through the interactive function interface of the intelligent agent to directly call the corresponding data analysis tool in the energy storage monitoring system and obtain the corresponding result from the energy storage monitoring system. The operation is simple and convenient, which reduces the user's learning cost and improves the operation efficiency.
[0103] Corresponding to the above method embodiment, see Figure 7A structural schematic diagram of a human-computer interaction system of an energy storage monitoring system is shown, the system includes an intelligent agent; the intelligent agent is communicatively connected with the energy storage monitoring system, the intelligent agent includes: an interactive function interface, a reasoning and decision module, a template library, a tool execution module and a front-end processing module; the interactive function interface is used to receive a first user question input by a user; the reasoning and decision module is used to: determine whether there is a first prompt word template corresponding to the first user question in the preset template library, and obtain the first prompt word template; wherein the template library is preset with prompt word templates of multiple task types; the prompt word template encapsulates a data analysis tool in the energy storage monitoring system; the reasoning and decision module is also used to: fill the first user question into the first prompt word template, determine the first tool from the data analysis tool encapsulated in the first prompt word template, and obtain the first tool configuration parameter corresponding to the first tool from the first user question; the tool execution module is used to: based on the first tool and the first tool configuration parameter, obtain the first question and answer result corresponding to the first user question from the energy storage monitoring system; the front-end processing module is used to: feed back the first question and answer result to the interactive function interface, and provide the first question and answer result to the user through the interactive function interface.
[0104] After receiving the user's question, the intelligent agent can use the intelligent agent interaction function to directly call the corresponding data analysis tool in the energy storage monitoring system according to the user's question, and obtain the question and answer results matching the user's question from the energy storage monitoring system. The intelligent agent at least includes an interactive function interface, a reasoning and decision module, a template library, a tool execution module, and a front-end processing module. Among them, the reasoning and decision module is constructed through a large language model, retrieval enhancement generation technology, and intelligent agent technology, and is the decision center of the intelligent agent.
[0105] The user can input the first user question through the interactive function interface provided by the intelligent agent. After the interactive function interface receives the first user question, the reasoning and decision module in the intelligent agent first searches the knowledge repository preset by the intelligent agent to see whether there is a historical question that is the same as the first user question. Here, the knowledge repository stores historical user questions that have been processed by the intelligent agent and recognized by the user and the historical question-and-answer results that match them. If a historical user question that is the same as the first user question is retrieved from the knowledge repository, the reasoning and decision module can determine that there is a first prompt word template corresponding to the first user question in the template library, and determine the prompt word template corresponding to the historical user question as the first prompt word template. If there is no historical question that is the same as the first user question, the reasoning and decision module can identify the prompt word template that can solve the first user question based on the intention recognition template, use it as the prompt word template applicable to the first user question, and determine that there is a first prompt word template corresponding to the first user question in the template library.
[0106] After the reasoning and decision module obtains the first prompt word template from the template library, it fills the first user question into the first prompt word template. The reasoning and decision module will semantically understand the first user question to obtain the user's needs. Based on the user's needs, the first tool that can realize the function of the user's needs is inferred from the data analysis tool encapsulated in the first prompt word template. At the same time, the reasoning and decision module obtains the first tool configuration parameters corresponding to the first tool from the first user question, and sends the first tool and the first tool configuration parameters to the tool execution module in the intelligent body. The tool execution module calls the first tool and obtains the corresponding data from the energy storage monitoring system according to the first tool configuration parameters, and uses the obtained data as the question and answer result of the first user question. Finally, the front-end processing module in the intelligent body feeds back the first question and answer result to the interactive function interface, and presents it in the visual interface provided by the visual display device for the user to view.
[0107] The human-computer interaction system of the energy storage monitoring system includes an intelligent agent, which is connected to the energy storage monitoring system in communication. The intelligent agent includes an interactive function interface, a reasoning decision module, a template library, a tool execution module, and a front-end processing module. After the system receives the user question input by the user through the interactive function interface, the reasoning decision module obtains the prompt word template corresponding to the user question from the template library, fills the user question into the corresponding prompt word template, and obtains the first tool and the first tool configuration parameters based on the filled prompt word template. The tool execution module calls the first tool in the energy storage monitoring system encapsulated by the prompt word template, obtains the question and answer result corresponding to the user question from the energy storage monitoring system according to the first tool configuration parameters, and feeds the result back to the interactive function interface through the front-end processing module, and is provided to the user by the interactive function interface. In this way, the user does not need to mechanically perform manual operations according to the preset function modules of the energy storage monitoring system to obtain the desired question results. The user only needs to input the user question described in natural language to directly call the corresponding data analysis tool in the energy storage monitoring system and obtain the corresponding result from the energy storage monitoring system. The operation is simple and convenient, which reduces the user's learning cost and improves the operation efficiency.
[0108] The human-computer interaction system of the energy storage monitoring system can implement the human-computer interaction method of the energy storage monitoring system. The intelligent agent in the system also includes: a knowledge repository; the reasoning decision module is also used to: retrieve whether there is a first historical data from the preset knowledge repository; wherein the first historical data includes: a first historical question that is the same as the first user question and a first historical question and answer result associated with the first historical question; the preset knowledge repository stores mutually related historical questions and historical question and answer results, and the historical questions and historical question and answer results match each other; if the first historical data exists, determine whether there is a first prompt word template corresponding to the first user question in the template library; wherein the first prompt word template is a prompt word template corresponding to the first historical question. If the first historical data does not exist, determine whether there is a first prompt word template in the template library through the intention recognition template in the template library; wherein the intention recognition template contains the functional description of the preset prompt word templates of multiple task types.
[0109] That is to say, the reasoning and decision module in the intelligent agent retrieves whether the first historical data exists from the knowledge repository preset by the intelligent agent. If the first historical data is retrieved from the preset knowledge repository, the reasoning and decision module can determine whether the first prompt word template corresponding to the first user problem exists in the template library. If the first historical data is not retrieved from the preset knowledge repository, the reasoning and decision module will infer based on the functional descriptions of the prompt word templates of multiple task types contained in the intention recognition template to determine whether the data analysis tools of the energy storage monitoring system encapsulated by these prompt word templates can solve the first user problem. If it is determined that the data analysis tools in the energy storage monitoring system encapsulated by the prompt word template of a certain task type can solve the first user problem, the reasoning and decision module determines the prompt word template of the task type as the first prompt word template.
[0110] In one embodiment, the intelligent agent in the above system also includes: a knowledge repository; the above reasoning decision module is also used to: retrieve whether there is second historical data from the preset knowledge repository; wherein the second historical data includes: a second historical question whose similarity to the first user question is greater than a preset degree threshold, and a second historical question and answer result corresponding to the second historical question; if the second historical data exists, based on the similarity corresponding to the second historical question in the second historical data, obtain a preset number of recommended data from the second historical data; fill the recommended data and the first user question into the first prompt word template.
[0111] The reasoning decision module will retrieve whether there is second historical data from a preset knowledge repository. If so, the reasoning decision module obtains a preset number of recommended data from the second historical data, and fills the recommended data and the first user question into the first prompt word template. The reasoning decision module can provide a reference when reasoning about the first prompt word template after filling in the first user question based on the recommended data, thereby improving the accuracy of the reasoning of the reasoning decision module, and then accurately determining the first tool in the data analysis tool encapsulated in the first prompt word template.
[0112] In one embodiment, the intelligent agent in the system also includes: an integration module; the tool execution module is also used to obtain an initial question and answer result from the energy storage monitoring system based on the first tool and the first tool configuration parameters; the reasoning decision module is used to fill the initial question and answer result into the first prompt word template filled with the first user question, and obtain the data summary result corresponding to the initial question and answer result; the integration module is used to integrate the data summary result and the initial question and answer result to obtain the first question and answer result corresponding to the first user question.
[0113] Here, after the reasoning and decision module determines the first tool and the first tool configuration parameters, the tool execution module calls the first tool and obtains the initial question and answer result from the energy storage monitoring system according to the first tool configuration parameters. After that, the reasoning and decision module needs to continue to fill the initial question and answer result into the first prompt word template filled with the first user question. At this time, the reasoning and decision module can use the initial question and answer result as a reference to further reason on the first prompt word template filled with the initial question and answer result and the first user question to obtain a richer data summary result. The integration module in the intelligent agent integrates the data summary result and the initial question and answer result as the first question and answer result corresponding to the first user question.
[0114] In one embodiment, the intelligent agent in the above system also includes: a knowledge repository; the above reasoning and decision module is also used to: receive user feedback information on the first question and answer result through an interactive function interface; in response to receiving positive feedback information, associate the first question and answer result and the first user question and save them in a preset knowledge repository; the positive feedback information indicates that: the first question and answer result matches the first user question.
[0115] After displaying the first question-and-answer result, the interactive function interface also provides a feedback interface for obtaining the user's evaluation of the first question-and-answer result. The reasoning and decision-making module in the intelligent agent determines that positive feedback information has been received, and associates the first question-and-answer result with the first user question and saves it in a preset knowledge repository, so that when a user question similar to the first user question is received later, the first question-and-answer result can be used to help the reasoning and decision-making module find a more accurate question-and-answer result that better meets the user's needs.
[0116] In one embodiment, the interactive function interface is used to: receive a second user question input by a user; the reasoning decision module is used to: determine whether there is a prompt word template corresponding to the second user question in the preset template library; fill the second user question into the task decomposition template in the template library for decomposition, and decompose the second user question into multiple sub-questions; the reasoning decision module is also used to: determine whether there is a prompt word template corresponding to the sub-question in the preset template library, determine the template corresponding to the sub-question from the template library based on the judgment result, and determine the tool corresponding to the sub-question based on the template corresponding to the sub-question; the tool execution module is used to: call the tool corresponding to the sub-question to obtain the question and answer result corresponding to the sub-question, and use the question and answer results corresponding to the multiple sub-questions as the second question and answer result corresponding to the second user question; the front-end processing module is used to: control the second question and answer result to be fed back to the interactive function interface. Among them, the types of templates corresponding to the sub-questions include structured query language templates or prompt word templates; the tools corresponding to the sub-questions include: data analysis tools in the energy storage monitoring system, or structured query tools; the structured query language template contains the context information of the preset structured query statement.
[0117] Here, after the interactive function interface receives the second user question input by the user, the reasoning decision module determines that there is no prompt word template corresponding to the second user question in the preset template library. The reasoning decision module will fill the second user question into the task decomposition template in the template library, perform task reasoning planning according to the task decomposition template, and decompose the second user question into multiple sub-questions. Then, the reasoning decision module determines whether there is a prompt word template corresponding to the sub-question in the preset template library. According to the judgment result, the template corresponding to the sub-question is determined from the template library, and the tool corresponding to the sub-question is determined according to the template corresponding to the sub-question. The tool execution module calls the tool corresponding to the sub-question to obtain the question and answer result corresponding to the sub-question, and the question and answer result corresponding to each sub-question is summarized as the question and answer result of the second user question. The front-end processing module feeds back the second question and answer result to the interactive function interface.
[0118] In one embodiment, the reasoning decision module is used to fill the sub-question into the intent recognition template, and determine whether there is a prompt word template applicable to the sub-question based on the functional description of the prompt word template included in the intent recognition template.
[0119] The reasoning decision module fills the sub-questions into the intent recognition template, and performs reasoning based on the functional descriptions of the prompt word templates of multiple task types contained in the intent recognition template to determine whether there is a prompt word template that is applicable to the sub-questions.
[0120] In one embodiment, the above-mentioned reasoning and decision module is used to: if there is a prompt word template corresponding to the first question in the sub-problem in the template library; fill the first question into the prompt word template corresponding to the first question; determine the second tool from the data analysis tool in the energy storage monitoring system encapsulated by the prompt word template corresponding to the first question, and obtain the second tool configuration parameters corresponding to the second tool from the first question; the tool execution module is used to: based on the second tool and the second tool configuration parameters, obtain the question and answer result corresponding to the first question from the energy storage monitoring system; the reasoning and decision module is also used to: if there is no prompt word template corresponding to the second question in the sub-problem in the template library, fill the second question into the structured query language template in the template library, and generate a structured query statement corresponding to the second question based on the filled structured query language template; the tool execution module is also used to: call the structured query tool according to the structured query statement corresponding to the second question, and obtain the question and answer result corresponding to the second question from the database of the energy storage monitoring system.
[0121] The reasoning decision module fills the sub-question into the intent recognition template. If the reasoning decision module determines that there is a prompt word template corresponding to the first question in the sub-question after reasoning based on the intent recognition template, the first question is filled into the prompt word template corresponding to the first question. The reasoning decision module will analyze and reason the prompt word template filled with the first question to obtain the second tool and the second tool configuration parameters corresponding to the second tool. Then, the tool execution module calls the second tool and obtains the corresponding data from the energy storage monitoring system according to the second tool configuration parameters as the question and answer result corresponding to the first question. If the sub-question matches the structured query language template after reasoning, the reasoning decision module fills the sub-question into the structured query language template, generates a structured query statement corresponding to the sub-question according to the filled structured query language template, and the tool execution module calls the structured query tool according to the structured query statement, directly obtains data from the database of the energy storage monitoring system, and uses the data as the question and answer result corresponding to the sub-question. The question and answer results corresponding to all sub-questions decomposed from the second user question are obtained in the above manner, and the question and answer results corresponding to all sub-questions are fed back to the interactive function interface.
[0122] In one embodiment, the above-mentioned reasoning and decision module is also used for: if the second user question contains result rendering information, obtaining the first chart style template corresponding to the result rendering information from the preset template library, and filling the second question and answer result into the first chart style template; the front-end processing module is used for: performing a first rendering on the second question and answer result based on the rendering data in the first chart style template; controlling the second question and answer result after the first rendering to be fed back to the interactive function interface; the reasoning and decision module is also used for: if the second user question does not contain result rendering information, filling the second question and answer result into the preset chart style template; the front-end processing module is used for: performing a second rendering on the second question and answer result based on the rendering data in the preset chart style template, and controlling the second question and answer result after the second rendering to be fed back to the interactive function interface.
[0123] If the reasoning decision module recognizes that the user question contains result rendering information, it obtains the first chart style template corresponding to the result rendering information from the preset template library, fills the second question and answer result into the first chart style template, and the front-end processing module performs a first rendering on the battery status data according to the rendering data of the report in the first chart style template, and feeds back the rendered battery status data to the interactive function interface. If the reasoning decision module recognizes that the second user question does not contain result rendering information, it fills the second question and answer result into the preset chart style template, and the front-end processing module renders the second question and answer result according to the rendering data in the preset chart style template, and feeds back the rendered second question and answer result to the interactive function interface.
[0124] In one embodiment, the above-mentioned interactive function interface is also used to: receive a third user question input by a user; the third user question is used to update the second question and answer result; the reasoning decision module is also used to: recognize that the third user question contains rendering update information for the first specified sub-question, obtain the second chart style template corresponding to the rendering update information from the preset template library, determine the question and answer result corresponding to the first specified sub-question from the second question and answer result, and fill the question and answer result corresponding to the first specified sub-question into the second chart style template; the above-mentioned front-end processing module is used to: perform a third rendering on the question and answer result corresponding to the first specified sub-question based on the rendering data in the second chart style template, and feed back the third rendering result to the interactive function interface.
[0125] Here, a third user question input by the user is received through the interactive function interface. The reasoning and decision module can identify the rendering update information in the third user question and the first designated sub-question corresponding to the rendering update information. Then, the reasoning and decision module will obtain the second chart style template corresponding to the rendering update information from the preset template library according to the rendering update information. The reasoning and decision module determines the question and answer result of the first designated sub-question from the second question and answer result, and the front-end processing module renders the question and answer result of the first designated sub-question according to the rendering data of the line chart in the second chart style template, and feeds it back to the interactive function interface.
[0126] In one embodiment, the above-mentioned reasoning and decision module is also used to: recognize that the third user question contains question editing information for the second designated sub-question, determine the updated second designated sub-question, and determine the template corresponding to the updated second designated sub-question from the template library; wherein the type of template corresponding to the updated second designated sub-question includes a structured query language template or a prompt word template; the tool execution module is also used to: obtain the question and answer result of the updated second designated sub-question; update the second question and answer result according to the updated question and answer result of the second designated sub-question, and obtain the updated second question and answer result; the front-end processing module is used to: feed back the updated second question and answer result to the interactive function interface.
[0127] After the reasoning and decision module identifies the question editing information for the second designated sub-question contained in the third user question, it determines the updated second designated sub-question according to the question editing information. Then, the reasoning and decision module obtains the template corresponding to the updated second designated sub-question from the template library. The template corresponding to the updated second designated sub-question can be a prompt word template or a structured query language template. The reasoning and decision module can continue to reason according to the template corresponding to the updated second designated sub-question to obtain the corresponding execution tool and filling parameters. The tool execution module calls the execution tool and filling parameters to obtain the updated question and answer result of the second designated sub-question, replaces the original question and answer result of the second designated sub-question in the second question and answer result with the updated question and answer result of the second designated sub-question, and obtains the updated second question and answer result. The front-end processing module feeds back the updated second question and answer result to the interactive function interface. In this way, according to the initially generated question and answer result, the user can continue to communicate with the intelligent agent and continuously fine-tune the data rendering method of the question and answer result to meet various data display requirements.
[0128] The present disclosure also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions. When the instructions are executed on a computer, the computer executes the steps of the human-computer interaction method of the energy storage monitoring system. The steps specifically include: receiving a first user question input by a user through an interactive function interface provided by an intelligent agent; determining whether a first prompt word template corresponding to the first user question exists in a preset template library, and obtaining the first prompt word template; wherein the template library is preset with prompt word templates of multiple task types; the prompt word template encapsulates a data analysis tool in the energy storage monitoring system; filling the first user question into the first prompt word template, determining a first tool from the data analysis tool encapsulated in the first prompt word template, and obtaining a first tool configuration parameter corresponding to the first tool from the first user question; obtaining a first question and answer result corresponding to the first user question from the energy storage monitoring system based on the first tool and the first tool configuration parameter; feeding back the first question and answer result to the interactive function interface, and providing the first question and answer result to the user through the interactive function interface.
[0129] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and modules described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.
[0130] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present disclosure, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present disclosure. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.
[0131] Finally, the present disclosure should explain that: the above embodiments are only specific implementation methods of the present disclosure, which are used to illustrate the technical solutions of the present disclosure, rather than to limit them. The protection scope of the present disclosure is not limited thereto. Although the present disclosure is described in detail with reference to the above embodiments, those skilled in the art should understand that: any technician familiar with the technical field can still modify the technical solutions recorded in the above embodiments within the technical scope disclosed in the present disclosure, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present disclosure, and should be included in the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be based on the protection scope of the claims.
Claims
1. A human-computer interaction method for an energy storage monitoring system, characterized in that: The method comprises: receiving a first user question input by a user; Determine whether a first prompt word template corresponding to the first user question exists in a preset template library, and obtain the first prompt word template; wherein the template library presets prompt word templates of multiple task types; and the prompt word template encapsulates a data analysis tool in the energy storage monitoring system; Filling the first user question into the first prompt word template, determining a first tool from the data analysis tools encapsulated in the first prompt word template, and acquiring a first tool configuration parameter corresponding to the first tool from the first user question; Based on the first tool and the first tool configuration parameters, a first question and answer result corresponding to the first user question is obtained from the energy storage monitoring system; and the first question and answer result is provided to the user.
2. The method according to claim 1, characterized in that The step of determining whether a first prompt word template corresponding to the first user question exists in a preset template library also includes: Retrieve from a preset knowledge repository whether there is first historical data; wherein the first historical data includes: a first historical question that is the same as the first user question and a first historical question-and-answer result associated with the first historical question; the preset knowledge repository stores mutually associated historical questions and historical question-and-answer results, and the historical questions and the historical question-and-answer results match each other; If the first historical data exists, determining that a first prompt word template corresponding to the first user question exists in the template library; wherein the first prompt word template is a prompt word template corresponding to the first historical question; If the first historical data does not exist, determine whether the first prompt word template exists in the template library through the intention recognition template in the template library; wherein the intention recognition template includes the functional description of the preset prompt word templates of multiple task types.
3. The method according to claim 1, characterized in that The step of filling the first user question into the first prompt word template also includes: Retrieve from a preset knowledge repository whether there is second historical data; wherein the second historical data includes: a second historical question whose similarity to the first user question is greater than a preset threshold, and a second historical question-and-answer result corresponding to the second historical question; If the second historical data exists, based on the degree of similarity corresponding to the second historical question in the second historical data, obtaining a preset number of recommended data from the second historical data; The recommendation data and the first user question are filled into the first prompt word template.
4. The method according to claim 1, characterized in that: The step of obtaining a first question-and-answer result corresponding to the first user question from the energy storage monitoring system based on the first tool and the first tool configuration parameters further includes: Based on the first tool and the first tool configuration parameters, an initial question and answer result is obtained from the energy storage monitoring system, the initial question and answer result is filled into the first prompt word template after the first user question is filled, and a data summary result corresponding to the initial question and answer result is obtained, and the data summary result and the first question and answer result are integrated to obtain a first question and answer result corresponding to the first user question.
5. The method according to claim 1, characterized in that After the step of providing the first question-and-answer result to the user, the method further includes: Receiving feedback information from the user regarding the first question-and-answer result; In response to receiving positive feedback information, the first question-and-answer result and the first user question are associated and saved in a preset knowledge repository; the positive feedback information indicates that the first question-and-answer result matches the first user question.
6. The method according to claim 1, characterized in that The method further comprises: receiving a second user question input by the user; Determining that there is no prompt word template corresponding to the second user question in a preset template library; filling the second user question into a task decomposition template in the template library for decomposition, and decomposing the second user question into a plurality of sub-questions; Determine whether there is a prompt word template corresponding to the sub-question in a preset template library, determine the template corresponding to the sub-question from the template library based on the determination result, determine the tool corresponding to the sub-question based on the template corresponding to the sub-question, call the tool corresponding to the sub-question to obtain the question and answer result corresponding to the sub-question, use the question and answer results corresponding to the multiple sub-questions as the second question and answer result corresponding to the second user question; and provide the second question and answer result to the user; Among them, the types of templates corresponding to the sub-questions include structured query language templates or prompt word templates; the tools corresponding to the sub-questions include: data analysis tools in energy storage monitoring systems, or structured query tools; the structured query language templates contain context information of preset structured query statements.
7. The method according to claim 6, characterized in that The step of determining whether there is a prompt word template corresponding to the sub-question in the preset template library comprises: The sub-question is filled into the intention recognition template, and based on the functional description of the prompt word template included in the intention recognition template, it is determined whether there is a prompt word template applicable to the sub-question.
8. The method according to claim 6, characterized in that The step of determining the template corresponding to the sub-question from the template library based on the judgment result, determining the tool corresponding to the sub-question based on the template corresponding to the sub-question, and calling the tool corresponding to the sub-question to obtain the question-answering result corresponding to the sub-question includes: If the prompt word template corresponding to the first question in the sub-question exists in the template library; fill the first question into the prompt word template corresponding to the first question; determine a second tool from the data analysis tool in the energy storage monitoring system encapsulated by the prompt word template corresponding to the first question, and obtain a second tool configuration parameter corresponding to the second tool from the first question; based on the second tool and the second tool configuration parameter, obtain the question and answer result corresponding to the first question from the energy storage monitoring system; If the prompt word template corresponding to the second question in the sub-question does not exist in the template library, the second question is filled into the structured query language template in the template library, and a structured query statement corresponding to the second question is generated based on the filled structured query language template; according to the structured query statement corresponding to the second question, a structured query tool is called to obtain a question and answer result corresponding to the second question from the database of the energy storage monitoring system.
9. The method according to claim 6, characterized in that The step of providing the second question-and-answer result to the user comprises: If the second user question includes result rendering information, a first chart style template corresponding to the result rendering information is obtained from a preset template library, the second question and answer result is filled into the first chart style template, the second question and answer result is first rendered based on the rendering data in the first chart style template, and the second question and answer result after the first rendering is provided to the user; If the second user question does not contain result rendering information, the second question and answer result is filled into a preset chart style template, the second question and answer result is rendered for the second time based on the rendering data in the preset chart style template, and the second question and answer result after the second rendering is provided to the user.
10. The method according to claim 6, characterized in that After the step of providing the second question-and-answer result to the user, the method further includes: receiving a third user question input by a user; the third user question is used to update the second question-and-answer result; Recognize that the third user question contains rendering update information for the first specified sub-question, obtain a second chart style template corresponding to the rendering update information from a preset template library, determine the question and answer result corresponding to the first specified sub-question from the second question and answer result, and fill the question and answer result corresponding to the first specified sub-question into the second chart style template; wherein the first specified sub-question includes: the sub-question corresponding to the second user question; Perform a third rendering on the question and answer result corresponding to the first designated sub-question based on the rendering data in the second chart style template, and provide the third rendering result to the user.
11. The method according to claim 10, characterized in that After the step of receiving the third user question input by the user, the method further includes: Recognizing that the third user question contains question editing information for the second designated sub-question, determining an updated second designated sub-question, determining a template corresponding to the updated second designated sub-question from the template library, and obtaining a question-and-answer result for the updated second designated sub-question; wherein the type of the template corresponding to the updated second designated sub-question includes a structured query language template or a prompt word template; The second question and answer result is updated according to the updated question and answer result of the second designated sub-question to obtain an updated second question and answer result, and the updated second question and answer result is provided to the user.
12. A human-computer interaction device for an energy storage monitoring system, characterized in that: The device includes an intelligent agent, which includes: an interactive function interface, a reasoning decision module, a template library, a tool execution module and a front-end processing module; The interactive function interface is used to: receive a first user question input by a user; The reasoning decision module is used to: determine whether a first prompt word template corresponding to the first user question exists in a preset template library, and obtain the first prompt word template; wherein the template library presets prompt word templates of multiple task types; the prompt word template encapsulates a data analysis tool in the energy storage monitoring system; The reasoning decision module is further used to: fill the first user question into the first prompt word template, determine a first tool from the data analysis tools encapsulated by the first prompt word template, and obtain a first tool configuration parameter corresponding to the first tool from the first user question; The tool execution module is used to: obtain a first question-and-answer result corresponding to the first user question from the energy storage monitoring system based on the first tool and the first tool configuration parameters; The front-end processing module is used to: feed back the first question and answer result to the interactive function interface, and provide the first question and answer result to the user through the interactive function interface.
13. A human-computer interaction system for an energy storage monitoring system, characterized in that: The system includes an intelligent agent, which is in communication with the energy storage monitoring system; the intelligent agent includes: an interactive function interface, a reasoning decision module, a template library, a tool execution module and a front-end processing module; The interactive function interface is used to: receive a first user question input by a user; The reasoning decision module is used to: determine whether a first prompt word template corresponding to the first user question exists in a preset template library, and obtain the first prompt word template; wherein the template library presets prompt word templates of multiple task types; the prompt word template encapsulates a data analysis tool in the energy storage monitoring system; The reasoning decision module is further used to: fill the first user question into the first prompt word template, determine a first tool from the data analysis tools encapsulated by the first prompt word template, and obtain a first tool configuration parameter corresponding to the first tool from the first user question; The tool execution module is used to: obtain a first question-and-answer result corresponding to the first user question from the energy storage monitoring system based on the first tool and the first tool configuration parameters; The front-end processing module is used to: feed back the first question and answer result to the interactive function interface, and provide the first question and answer result to the user through the interactive function interface.
14. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed by the processor, the human-computer interaction method of the energy storage monitoring system as described in any one of claims 1 to 11 is implemented.
Citation Information
Patent Citations
Financial knowledge intelligent question and answer method and system, terminal equipment and storage medium
CN117875292A
Substation operation and maintenance intelligent question answering method based on large model, related method and device
CN117891831A
Intelligent question-answering system, question-answering interaction method, computer equipment and storage medium
CN118035392A
Cited By
Equipment intelligent operation and maintenance system and method based on RAG model
CN119293199A
A RAG model-based equipment intelligent operation and maintenance system and method
CN119293199B
Interaction method based on artificial intelligence equipment, server and artificial intelligence equipment
CN119360317A
Charging station management system intelligent question and answer method, electronic equipment and storage medium
CN120448603A
