CN-122022031-A - Air conditioner load prediction and feedforward energy-saving control method and device integrating airport building mechanism and multi-mode large model

CN122022031ACN 122022031 ACN122022031 ACN 122022031ACN-122022031-A

Abstract

The application provides an air conditioner load prediction and feedforward energy-saving control method and device integrating airport building mechanism and a multi-mode large model, which are applied to the technical field of data processing. The application uses 'demand side prediction and supply side regulation' as a core framework around the energy saving and accurate regulation of an airport air conditioning system. The method comprises the steps of obtaining multisource original data of airport building characteristics, environments, personnel flows and air conditioning working conditions, generating standardized cold energy demand characteristic data through integration and time sequence rule mining, constructing a double-prediction model, performing fusion analysis, outputting accurate cold energy demands of all space-time dimensions, setting equipment cooperative constraint based on the built energy consumption standard model, optimizing operation schemes of equipment such as a water chilling unit, and finally combining a double-layer framework to generate air conditioning system early warning information considering comfort and energy conservation, and realizing intelligent management.

Inventors

WANG KUI
Liao ting
XU HUA
CHEN JINGQIANG
LIU FUWEI
LI YUMING
HE XIU
YAO YE
GUO CHAORUI
YANG JIANZHONG
LIANG LULU
ZHAO XUEYING
LIAO HAITAO
PAN HONG
LIU QIMEI

Assignees

广西桂物节能集团有限公司
广西桂物金岸制冷空调技术有限责任公司
上海交通大学

Dates

Publication Date: 20260512
Application Date: 20260122

Claims (9)

1. An air conditioner load prediction and feedforward energy-saving control method integrating airport building mechanism and multi-mode large model is characterized by comprising the following steps: acquiring airport multisource original data comprising airport building characteristics, environment parameters, personnel flows and operation conditions of an air conditioning system; processing airport multisource original data, and generating standardized cold energy demand related characteristic data by integrating outdoor weather, indoor environment, passenger track and multi-dimensional historical data of equipment operation and adopting a cyclic neural network technology to mine time sequence data rules; Processing the standardized cold energy demand related characteristic data, constructing a multi-region room temperature neural network large-lag prediction model and a central air conditioning system dynamic energy consumption trend prediction model, and generating accurate cold energy demand prediction results of each space-time dimension through multi-mode large model fusion analysis; processing the accurate cold demand prediction result, constructing an air conditioner energy consumption standard model which covers peak-valley period division, regional load grade grading and equipment energy efficiency threshold setting, and generating equipment group collaborative optimization control constraint conditions based on a strategy of dynamic scheduling after pre-determined evaluation standards; Processing the equipment group collaborative optimization control constraint condition, and carrying out collaborative optimization calculation on the air conditioning equipment group comprising the water chilling unit, the fan coil and the cooling tower by combining the cold energy demand prediction data to generate a target equipment operation scheme; And processing the operation scheme of the target equipment and related data, and generating airport air conditioning system early warning information considering comfort and energy saving targets by combining a double-layer architecture of accurate prediction of a demand side and intelligent regulation of a supply side.
2. The method of claim 1, wherein the processing the airport multisource raw data to generate standardized cold demand related characteristic data by integrating outdoor weather, indoor environment, passenger trajectory and multi-dimensional historical data of equipment operation and mining time series data rules by using a cyclic neural network technology, comprises: Aiming at the heterogeneous characteristics of outdoor weather, indoor environment, passenger track and equipment operation data, establishing an integration mechanism of space-time dimension alignment and data quality grading to realize space-time matching of multi-source data, and then intelligently completing through outlier rejection and missing values to form a complete airport multi-source data set covering all elements of building-environment-personnel-equipment; The method comprises the steps of constructing an improved circulating neural network model according to the characteristic that the air conditioner load of a docking airport is obviously influenced by time sequence factors of flight peak and season alternation, adding an airport operation characteristic adaptation layer in a model structure, and deeply mining and accurately extracting the time sequence rule of multi-dimensional historical data; and constructing a load demand-data characteristic mapping model, converting unstructured time sequence rules into structured characteristic indexes, introducing airport building mechanism correction factors to conduct scene calibration on the characteristic indexes, and conducting format unification and range normalization processing on the calibrated characteristic indexes in combination with data standardization standards to generate standardized cold demand related characteristic data with data standardization and airport scene suitability.
3. The method of claim 1, wherein the processing of the normalized cold demand related characteristic data to construct a multi-region room temperature neural network large hysteresis prediction model and a central air conditioning system dynamic energy consumption trend prediction model, and the generating of accurate cold demand prediction results of each space-time dimension through multi-mode large model fusion analysis comprises: The method comprises the steps of carrying out dimension splitting on standardized cold energy demand related characteristic data, extracting building mechanism related characteristics comprising regional space parameters and heat transfer characteristics, and dynamic demand characteristics related to time load fluctuation and people flow density to form a characteristic subset adapting to dual-model training; Combining the multi-equipment coupling operation characteristics of a central air conditioning system, constructing a dynamic energy consumption trend prediction model, taking the system operation characteristics of equipment load rate, energy efficiency parameters and a cold energy conveying path as core input, and capturing a dynamic matching rule of cold energy supply and demand through time sequence association analysis; And introducing a multi-mode large model fusion mechanism, iteratively adjusting the weight ratio based on the prediction accuracy performance of the double models under different scenes, and simultaneously, fusing the building structure characteristics and the multi-mode auxiliary information of the real-time environment parameters to correct the prediction deviation so as to generate the accurate cold energy demand prediction result of each space-time dimension covering the whole area and the whole time period.
4. The method of claim 3, wherein processing the accurate cooling demand prediction results, constructing an air conditioner energy consumption standard model covering peak-valley period division, regional load level classification and equipment energy efficiency threshold setting, generating equipment group collaborative optimization control constraint conditions based on a strategy of dynamic scheduling after pre-determined evaluation criteria, and comprising: Extracting and processing space-time distribution characteristics, load fluctuation amplitude and peak demand period data in the accurate cold demand prediction result, and determining peak-valley period division basis, regional load grade judgment standard and equipment energy efficiency threshold interval by combining airport operation peak-valley rules, regional function differences and equipment operation limit parameters; constructing an air conditioner energy consumption standard model which covers peak-valley period division, regional load grade grading and equipment energy efficiency threshold setting based on the extracted core data and the judgment standard, and defining the quantization boundary of each dimension evaluation index; and (3) following a core strategy of dynamic scheduling after a pre-determined evaluation standard, carrying out association integration on peak-valley period rules, load levels corresponding to energy consumption upper limits and equipment energy efficiency constraint conditions in an energy consumption standard model, and generating a constraint system for adapting to the cooperative regulation of equipment groups, wherein the constraint system comprises equipment group cooperative optimization control constraint conditions of the upper limit of cold energy supply, the equipment operation energy efficiency standard reaching requirement and the load dynamic allocation threshold of each area in each period.
5. The method of claim 1, wherein processing the plant group collaborative optimization control constraint, and performing collaborative optimization calculations on an air conditioning plant group including a chiller, a fan coil, and a cooling tower in combination with the refrigeration demand prediction data, to generate a target plant operating scenario comprises: Deep disassembly is carried out on the upper limit of cold energy supply, the energy efficiency meeting the requirements and the load distribution threshold value in the equipment group collaborative optimization control constraint condition through an equipment characteristic analysis algorithm, and an equipment operation characteristic baseline rule and parameter adaptation threshold value interval of an adaptation water chilling unit, a fan coil and a cooling tower are generated; abutting the air conditioning equipment group operation efficiency database and the cold energy demand prediction data, constructing equipment cooperative regulation and control reference, determining operation parameter optimization rules of different equipment through supply and demand matching logic, introducing a multi-equipment coupling cooperative model, combining an intelligent optimization algorithm with an airport air conditioning system operation scene image, and establishing a demand-constraint-optimization cooperative dynamic calculation mechanism; Taking the equipment type as a coverage dimension, fusing the operating parameter range and the cooling capacity demand distribution data corresponding to the water chilling unit, the fan coil and the cooling tower, and combining the equipment operating characteristic baseline rule and the parameter adaptation threshold; And (3) integrating constraint conditions and prediction data by associating pretreatment mechanisms with parameter normalization and supply and demand data, strengthening the accuracy of optimization calculation by combining a multi-equipment coupling cooperative model, and generating a target equipment operation scheme with equipment operation parameters, cooperative logic description and energy-saving benefit estimation by integrating the characteristics of a double-layer architecture of the dynamic adjustment of the demand side prediction and the cooperative response of the supply side equipment.
6. The method of claim 5, wherein processing the target device operating scheme and related data, in combination with a dual-layer architecture for demand side accurate prediction and supply side intelligent regulation, generates airport air conditioning system early warning information that compromises comfort and energy saving objectives, comprising: based on a preset comfort level and energy-saving double-target verification rule, introducing parameter execution deviation, energy consumption standard reaching rate and room temperature fluctuation range in a target equipment operation scheme, performing feature recognition on equipment operation data and double-layer framework suitability, and generating an early warning abnormal feature set; Comparing the related data of the operation scheme of the target equipment with a preset optimal operation parameter library and an energy consumption reference value library of the equipment, verifying different operation scenes of an airport, determining the deviation degree of the operation state of the equipment and the double-target requirement, and generating an operation deviation degree assessment result; Performing association analysis on the operation deviation assessment result and the early warning abnormal feature set, and performing weighting processing on the deviation assessment result by combining the accuracy feedback of the demand side cold prediction and the cooperative response efficiency of the supply side equipment to generate an early warning optimization vector fused with the prediction-regulation cooperative logic; And carrying out normalization and comprehensive operation on the operation deviation evaluation result and the early warning optimization vector, and generating early warning information of the airport air conditioning system, which comprises early warning level, abnormal parameter identification and energy-saving comfort balance suggestion, by combining with the real-time regulation response requirement of the airport air conditioning system.
7. An air conditioner load prediction and feedforward energy-saving control device integrating airport building mechanism and multi-mode large model, which is characterized by comprising: The acquisition module is used for acquiring airport multisource original data comprising airport building characteristics, environment parameters, personnel flows and operation conditions of an air conditioning system; The system comprises a processing module, a target device operation scheme, an air conditioning system operation scheme and a target energy saving system, wherein the processing module is used for processing airport multisource original data, generating standardized cold demand related characteristic data by integrating multi-dimensional historical data of outdoor weather, indoor environment, passenger track and device operation and adopting a cyclic neural network technology mining time sequence data rule, processing the standardized cold demand related characteristic data, constructing a multi-region room temperature neural network large hysteresis prediction model and a central air conditioning system dynamic energy consumption trend prediction model, generating accurate cold demand prediction results of each space-time dimension through multi-mode large model fusion analysis, processing the accurate cold demand prediction results, constructing an air conditioning energy consumption standard model which covers peak-valley time interval division, regional load grade grading and device energy efficiency threshold setting, generating a device group collaborative optimization control constraint condition based on a strategy which is set by a pre-determined evaluation standard, processing the device group collaborative optimization control constraint condition, combining the cold demand prediction data, performing collaborative optimization calculation on an air conditioning device group comprising a cold machine set, a fan coil and a cooling tower, generating the target device operation scheme, processing the target device operation scheme and related data, combining accurate demand side prediction and intelligent dual-layer air conditioning system comfort level and energy saving architecture.
8. An electronic device, comprising: and a memory for storing executable instructions of the first processor; Wherein the first processor is configured to execute the air conditioning load prediction and feed-forward energy saving control method of fusing airport building mechanisms with multi-modal large models of any one of claims 1-6 via execution of the executable instructions.
9. A computer readable storage medium having stored thereon a computer program, wherein the computer program when executed by a second processor implements the air conditioning load prediction and feed forward energy saving control method of merging airport building mechanisms with multi-modal large models according to any one of claims 1-6.

Description

Air conditioner load prediction and feedforward energy-saving control method and device integrating airport building mechanism and multi-mode large model Technical Field The invention relates to the technical field of data processing, in particular to an air conditioner load prediction and feedforward energy-saving control method and device integrating airport building mechanism and a multi-mode large model. Background In the airport operation process, an air conditioning system is used as a core infrastructure to bear the key tasks of maintaining proper temperature and humidity in a terminal building and guaranteeing comfort level of passengers and staff. However, airport buildings have the particularity which is obviously different from common buildings, on one hand, the airport buildings are of high-volume space structures, the heat transfer characteristics of building enclosures (such as large-area glass curtain walls) are complex, the cold energy loss paths are multiple, and on the other hand, the airport people flow and the traffic flow are influenced by a flight take-off and landing plan to present strong fluctuation, the personnel density difference between peak time and flat peak time is obvious, and the air conditioner cold energy requirement has severe and irregular fluctuation in space-time dimension. In addition, the air conditioning system comprises a plurality of equipment units such as a water chilling unit, a fan coil, a cooling tower and the like, and the equipment units have a strong coupling operation relationship, so that the complexity of system regulation and control is further increased. The following problems generally exist in the conventional regulation and control mode of an airport air conditioning system: The traditional prediction method is dependent on a single data dimension or a general time sequence model, airport building mechanisms (such as space structures and building envelope thermal resistances) and airport operation characteristics (such as flight peaks and passenger tracks) are not fully fused, space-time fluctuation rules and hysteresis characteristics of cold energy demands are difficult to capture, the deviation between a prediction result and actual demands is large, and reliable basis cannot be provided for accurate regulation. The existing regulation and control mode mostly adopts 'single-point control' or 'fixed parameter operation', a device group cooperative optimization mechanism is not established, the coupling relation and operation energy efficiency threshold constraint among devices are ignored, and the conditions of 'super-demand supply' or 'insufficient supply' often occur, so that energy waste is caused, and indoor comfort level is possibly influenced. The energy conservation and comfort level balance is unbalanced, a systematic regulation and control framework which is compatible with the energy conservation target and the comfort level requirement is lacking, the traditional strategy is often focused on single target optimization (such as pursuing energy consumption reduction or excessively guaranteeing comfort level), and a dynamic balance mechanism is not established, so that the requirements of the two are difficult to meet under complex scenes such as flight peak, season alternation and the like. The early warning response mechanism is lagged, the existing early warning is mostly based on the passive triggering of equipment faults or parameter overrun, the early warning is not combined with the prediction data of the demand side and the regulation effect of the supply side to conduct early judgment, the active recognition and early warning capability of the 'prediction-regulation' cooperative deviation is lacking, and the risk of exceeding energy consumption or reducing comfort level cannot be avoided timely. It should be noted that the information disclosed in the foregoing background section is only for enhancement of understanding of the background of the disclosure and does not constitute prior art information that is already known to those of ordinary skill in the art. Disclosure of Invention According to one aspect of the application, an air conditioner load prediction and feedforward energy-saving control method integrating an airport building mechanism and a multi-mode large model is provided, and the method comprises the steps of obtaining airport multi-source original data comprising airport building characteristics, environment parameters, personnel flows and operation working conditions of an air conditioning system, processing the airport multi-source original data, mining time sequence data rules by adopting a cyclic neural network technology through integrating multi-dimensional historical data of outdoor weather, indoor environments, passenger tracks and equipment operation, generating standardized cold demand related characteristic data, processing the standardized cold demand related characteristic data, constructing a multi