CN-121998387-A - Intelligent park energy-saving cooperative regulation and control system and method based on multi-mode large model

CN121998387ACN 121998387 ACN121998387 ACN 121998387ACN-121998387-A

Abstract

The invention discloses an intelligent park energy-saving cooperative regulation and control system and method based on a multi-mode large model, and belongs to the technical field of intelligent management of the Internet of things. The method comprises the steps of obtaining multi-mode perception data of vision, an internet of things, text and voice, respectively extracting features of the data to obtain multi-mode features, taking the text features and the audio features as result guide features, vectorizing the vision features and the internet of things features, inputting the vectorized vision features and the internet of things features into a fusion function of a multi-mode large model, generating a unified state characterization vector through the fusion function, inputting a regulation and control decision module to generate a collaborative regulation and control instruction, and controlling park equipment according to the instruction. According to the method, through multi-mode feature fusion and unified characterization, cross-subsystem joint optimization and dynamic priority regulation are realized, and the energy utilization efficiency and the intelligent operation and maintenance level of a park are effectively improved.

Inventors

HUANG YUANXIN
DAI LU
MENG JIN
HUANG JIANJUN

Assignees

广州晟能软件科技有限公司
广州晟能电子科技有限公司
晟能数智科技(杭州)有限公司

Dates

Publication Date: 20260508
Application Date: 20260410

Claims (10)

1. An intelligent park energy-saving cooperative regulation and control method based on a multi-mode large model is characterized by comprising the following steps of: S1, acquiring multi-mode sensing data, wherein the multi-mode sensing data comprises visual sensing data, internet of things sensing data, text sensing data and voice sensing data; s2, respectively carrying out feature extraction on the visual perception data, the internet of things perception data, the text perception data and the voice perception data to obtain visual features V_t, internet of things features I_t, text features T_t and audio features A_t, wherein T represents a time point; s3, taking the text feature T_t and the audio feature A_t as result guiding features, converting the visual feature V_t and the Internet of things feature I_t into vector forms through vectorization processing, and inputting the vector forms into a fusion function F in a multi-modal large model; s4, generating a unified state representation vector S_t through a fusion function F, inputting the S_t to a regulation and control decision module in the multi-mode large model, and generating a cooperative regulation and control instruction of park equipment; and S5, controlling the park equipment according to the cooperative regulation and control instruction so as to realize energy-saving regulation and control.
2. The intelligent park energy-saving collaborative regulation and control method based on the multi-mode large model according to claim 1, further comprising: And inputting the unified state characterization vector S_t to a load prediction module in the multi-mode large model, and generating a time-by-time cold load prediction result of a future period.
3. The intelligent park energy-saving collaborative regulation and control method based on the multi-mode large model according to claim 2, further comprising: and inputting the load prediction result to the regulation and control decision module to construct the multi-objective optimization function.
4. The intelligent park energy-saving collaborative regulation and control method based on the multi-mode large model according to claim 3, wherein the multi-objective optimization function construction formula is as follows: min[α·E_total + β·C_carbon + γ·D_comfort]; Wherein E_total represents the total energy consumption index, C_carbon represents the carbon emission index, D_comfort represents the comfort level deviation index, and alpha, beta and gamma are respectively corresponding weight coefficients; The value range of the weight coefficient alpha is 0.3-0.6, the value range of beta is 0.2-0.5, and the value range of gamma is 0.1-0.4.
5. The intelligent park energy-saving collaborative regulation and control method based on a multi-mode big model according to claim 4, wherein the total energy consumption index e_total is calculated by the following method: e_total = Σ (device power x run time) +fixed base energy consumption; the carbon emission index c_carbon is calculated by: c_carbon=Σ (energy consumption x corresponding carbon factor); the comfort level deviation index d_comfort includes at least a temperature deviation degree calculated by: temperature deviation = 。
6. The method of claim 1, wherein S4 further comprises: and quantifying the contribution degree of the factors to the carbon emission through a transducer attention weight, and generating a visual carbon flow map and a natural language interpretation report, wherein the visual carbon flow map and the natural language interpretation report at least comprise one optimization strategy.
7. The intelligent park energy-saving collaborative regulation and control method based on the multi-mode large model according to claim 6, further comprising: And carrying out strategy distribution on the generated cooperative regulation and control instruction and the optimization strategy through a large model decision engine, and sending the cooperative regulation and control instruction to corresponding park equipment or subsystems thereof.
8. Intelligent park energy-saving cooperative regulation and control system based on multi-mode large model, which is characterized by comprising: the multi-modal sensing module is used for acquiring multi-modal sensing data, wherein the multi-modal sensing data comprises visual sensing data, internet of things sensing data, text sensing data and voice sensing data; the feature extraction module is used for carrying out feature extraction on the multi-mode sensing data to obtain a visual feature vector V_t, an internet of things feature vector I_t, a text feature vector T_t and an audio feature vector A_t, wherein T represents a time point; The fusion module is used for taking the text feature T_t and the audio feature A_t as result guiding features, converting the visual feature V_t and the internet of things feature I_t into vector forms through vectorization processing, and inputting the vector forms into a fusion function F in a multi-modal large model; the load prediction module is used for generating a unified state representation vector S_t by the fusion function F, inputting the S_t to the regulation and control decision module in the multi-mode large model and generating a cooperative regulation and control instruction of park equipment; the regulation and control decision module is used for inputting the load prediction result into the regulation and control decision module in the multi-mode large model, constructing a multi-objective optimization function based on an attention mechanism and generating a cooperative regulation and control instruction of at least one park device; and the control execution module is used for controlling at least one park device according to the cooperative regulation and control instruction.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the multi-modal large model smart campus energy conservation collaborative tuning method of any one of claims 1 to 7 when the program is executed by the processor.
10. A non-transitory computer readable storage medium having stored thereon a computer program, wherein the computer program when executed by a processor implements the multi-modal large model smart campus energy conservation collaborative tuning method of any one of claims 1 to 7.

Description

Intelligent park energy-saving cooperative regulation and control system and method based on multi-mode large model Technical Field The invention relates to the field of intelligent management of the Internet of things, in particular to an intelligent park energy-saving collaborative regulation and control method, system, equipment and storage medium based on a multi-mode large model. Background Along with the rapid development of the internet of things technology, the traditional internet of things system can realize interconnection and intercommunication of physical equipment and data acquisition and transmission through various sensors, communication modules and network protocols. The prior art generally adopts a centralized or distributed data platform to gather collected data, and performs data processing, analysis and early warning notification through a preset rule or a traditional machine learning model (such as a regression model and a simple neural network). For example, in the energy management field of the intelligent park, the system can monitor the operation parameters and environment information of subsystems such as heating ventilation and air conditioning, illumination, micro-grid and the like in real time, notify related personnel in a mode such as short messages, mails or platform alarms when data are abnormal, and part of the system can also carry out simple automatic adjustment based on preset rules. However, such systems still have significant limitations in practical applications. First, the decision-making closed loop of the system is highly dependent on manual intervention. And each subsystem (heating ventilation, illumination and micro-grid) is independently optimized, global coordination is lacked, a chimney-type optimization pattern is formed, and when strategy conflict occurs across the systems, personnel are difficult to coordinate in a short time, so that the overall energy-saving effect and the operation reliability of the system are affected. Furthermore, the fault handling and decision making process relies heavily on the personal experience of the operator. Due to the lack of systematic summaries and efficient precipitation of historical process cases and expert experience, it is difficult to form a standardized decision support knowledge base. When facing complex or long-term dynamic conditions such as equipment performance attenuation, use habit change and the like, personnel often can only judge by means of self experience, the treatment efficiency is low, and decision deviation is easily caused by insufficient experience. In addition, the perception capability of the system is single, and the comprehensive research and judgment are difficult to be carried out by fusing the multi-mode data. The prior art can only process structured numerical sensor data, and lacks semantic fusion capability for multi-source heterogeneous data such as video streams, natural language work orders, equipment drawings, weather reports and the like. Therefore, the current internet of things technology framework breaks through the bottleneck of relying on manual decision in the prior art on the basis of realizing data interconnection and intercommunication, and realizes global collaborative optimization across systems, fusion perception of multi-mode data and interpretability of decision processes by introducing more intelligent data analysis, decision support and automatic execution mechanisms, so that autonomous response capability, processing efficiency and reliability of the system are improved, and cost reduction, efficiency improvement and intelligent operation and maintenance of an intelligent park are truly realized. Disclosure of Invention The invention aims to overcome at least one defect of the prior art, and provides an intelligent park energy-saving collaborative regulation and control system and method based on a multi-mode large model, which are used for solving the technical problems of single system perception capability and dependence on manual decision in the prior art. The invention provides an intelligent park energy-saving cooperative regulation and control method based on a multi-mode large model, which comprises the following steps: S1, acquiring multi-mode sensing data, wherein the multi-mode sensing data comprises visual sensing data, internet of things sensing data, text sensing data and voice sensing data; s2, respectively carrying out feature extraction on the visual perception data, the internet of things perception data, the text perception data and the voice perception data to obtain visual features V_t, internet of things features I_t, text features T_t and audio features A_t, wherein T represents a time point; s3, taking the text feature T_t and the audio feature A_t as result guiding features, converting the visual feature V_t and the Internet of things feature I_t into vector forms through vectorization processing, and inputting the vector forms into a fusio