Search

CN-121995771-A - Building energy consumption optimization management method and device based on digital twin

CN121995771ACN 121995771 ACN121995771 ACN 121995771ACN-121995771-A

Abstract

The embodiment of the application provides a building energy consumption optimization management method and device based on digital twinning, which realize accurate analysis of states through twinning modeling and load prediction. An optimization mechanism is built, and a reliable control strategy is built by combining multi-objective optimization and rolling solution. And introducing self-adaptive adjustment, and ensuring continuous optimization of the model through error compensation and parameter identification. The method effectively solves the defects of the traditional technology in the aspects of state prediction, strategy optimization, model adjustment and the like, and provides technical support for building energy consumption optimization.

Inventors

  • WANG WEI

Assignees

  • 浙江微风智能科技有限公司

Dates

Publication Date
20260508
Application Date
20260408

Claims (10)

  1. 1. The digital twinning-based building energy consumption optimization management method is characterized by comprising the following steps of: Acquiring temperature and humidity data, energy consumption metering data and personnel density data uploaded by a sensor in a building to obtain a multi-source sensing data stream, carrying out anomaly detection and data alignment on the multi-source sensing data stream to obtain a synchronous data set, constructing a digital twin model based on the synchronous data set to obtain a twin body state vector, carrying out time sequence fusion on the twin body state vector and preset forecast data to generate a forecast input matrix, and carrying out load forecast calculation on the forecast input matrix to obtain a future period load sequence; Constructing a multi-objective optimization function based on the future period load sequence to obtain a target equation set, applying model constraint and physical limit constraint to the target equation set to generate a constraint condition set, inputting the constraint condition set into a rolling optimization solver to calculate to obtain a multi-scene control strategy, performing weight distribution on the multi-scene control strategy to obtain a fusion control instruction sequence, extracting control parameters of a first time step from the fusion control instruction sequence to generate an execution instruction, and issuing the execution instruction to building execution equipment; And acquiring an actual load value and an operation parameter of the execution equipment to obtain a feedback data set, carrying out error calculation on the feedback data set and a predicted load value to obtain a predicted deviation value, constructing an error compensation model based on the predicted deviation value to complete prediction correction, carrying out online identification on key parameters of the digital twin model according to the predicted deviation value to generate an updated parameter set, and writing the updated parameter set into the digital twin model to complete self-adaptive adjustment.
  2. 2. The method for optimizing and managing building energy consumption based on digital twin according to claim 1, wherein the steps of acquiring temperature and humidity data, energy consumption metering data and personnel density data uploaded by sensors in a building to obtain a multi-source sensing data stream, performing anomaly detection and data alignment on the multi-source sensing data stream to obtain a synchronous data set comprise: Reading an analog signal output by a sensor according to a preset sampling frequency through a data acquisition gateway to obtain an original signal stream, performing signal separation on the original signal stream to generate a channel data set, performing signal quality detection on the channel data set to obtain a data reliability index, and preprocessing each channel signal based on the data reliability index to generate a multi-source sensing data stream; Calculating signal autocorrelation values according to sliding windows based on the multi-source sensing data flow, inputting the correlation characteristic sets into an anomaly detector to generate an anomaly marking matrix, removing anomaly data points according to the anomaly marking matrix to obtain an effective data set, calculating time delay deviation among multi-channel signals through a time stamp alignment algorithm, and performing time delay compensation on the effective data set to generate a synchronous data set.
  3. 3. The method for optimizing and managing building energy consumption based on digital twin according to claim 1, wherein the constructing a digital twin model based on the synchronous data set to obtain a twin body state vector, performing time sequence fusion on the twin body state vector and preset forecast data to generate a prediction input matrix, and performing load prediction calculation on the prediction input matrix to obtain a future period load sequence comprises the following steps: The method comprises the steps of carrying out data grouping on a synchronous data set according to a building equipment topological structure to obtain an equipment state table, extracting physical characteristic parameters from the equipment state table to obtain a model parameter set, constructing a device performance model based on the model parameter set to obtain a component model set, carrying out modularized assembly on the component model set and a building physical model to generate a twin body state vector, and calculating steady state response of a system according to the twin body state vector to obtain a model initial state; And carrying out data standardization processing on the initial state of the model and preset forecast data to obtain a normalized feature group, generating a state sequence matrix by a time sequence feature extractor, carrying out sliding window segmentation on the state sequence matrix to obtain a sample sequence group, inputting the sample sequence group into a depth forecast network to calculate to obtain a load forecast value, and carrying out time step expansion on the load forecast value to generate a future period load sequence.
  4. 4. The method for optimizing and managing building energy consumption based on digital twin according to claim 1, wherein constructing a multi-objective optimization function based on the future period load sequence to obtain a set of objective equations, applying model constraints and physical limit constraints to the set of objective equations to generate a constraint condition set, inputting the constraint condition set into a rolling optimization solver to calculate a multi-scene control strategy, and the method comprises: performing time scale decomposition on a future period load sequence to obtain a multi-period load matrix, constructing a power grid electricity purchasing cost function based on the multi-period load matrix to obtain a cost item, mapping equipment adjustment amplitude into an energy consumption function to obtain a loss item, constructing a comfort evaluation function according to indoor environment parameters to obtain a comfort item, setting self-adaptive weight coefficients for the cost item, the loss item and the comfort item to generate a target equation set, and extracting equipment operation boundary conditions based on the target equation set to obtain a constraint variable set; And performing parameter matching on the constraint variable set and the equipment dynamic model to obtain a model constraint group, applying equipment physical limit value to the model constraint group to generate a hard constraint condition, constructing environment comfort constraint according to a preset rule to obtain a soft constraint condition, performing constraint fusion on the hard constraint condition and the soft constraint condition to generate a constraint condition set, and performing feasibility verification on the constraint condition set to obtain the multi-scene control strategy.
  5. 5. The method for optimizing and managing building energy consumption based on digital twin according to claim 1, wherein the step of performing weight distribution on the multi-scene control strategy to obtain a fused control instruction sequence, extracting control parameters of a first time step from the fused control instruction sequence to generate an execution instruction, and issuing the execution instruction to building execution equipment comprises the steps of: Performing target value evaluation on a multi-scene control strategy to obtain a performance index matrix, calculating scene priority based on the performance index matrix to obtain a weight distribution diagram, constructing a self-adaptive weight function according to the weight distribution diagram to generate a fusion coefficient group, performing weighted combination operation on the fusion coefficient group and the multi-scene control strategy to obtain a fusion control instruction sequence, and performing time sequence consistency check on the fusion control instruction sequence to obtain a check result table; and extracting first time step data from the fusion control instruction sequence based on the verification result table to obtain a control parameter set, performing equipment constraint verification on the control parameter set to generate an execution instruction, converting the execution instruction according to a control protocol to obtain an equipment control message, and distributing the equipment control message to corresponding execution equipment according to an equipment communication address to finish instruction issuing.
  6. 6. The method for optimizing and managing building energy consumption based on digital twin according to claim 1, wherein the acquiring actual load values and operation parameters of the executing device to obtain a feedback data set, and performing error calculation on the feedback data set and the predicted load values to obtain a predicted deviation value, includes: reading the real-time running state of an execution device through a device monitoring interface to obtain a state data stream, carrying out data analysis on the state data stream to obtain a load data set and a parameter data set, carrying out data cleaning on the load data set according to a preset rule to generate an actual load value, extracting the running characteristic of the device from the parameter data set to obtain a running parameter table, and carrying out data merging on the actual load value and the running parameter table to generate a feedback data set; And carrying out time window alignment on the actual load value in the feedback data set to obtain an actual measurement sequence, mapping the actual measurement sequence and the predicted load value according to a time stamp to generate a data pair, calculating a multidimensional error index based on the data pair to obtain an error feature vector, and carrying out weighted fusion operation on the error feature vector to generate a predicted deviation value.
  7. 7. The method for optimizing and managing building energy consumption based on digital twin according to claim 1, wherein the constructing an error compensation model based on the prediction bias value to complete prediction correction, performing on-line identification on key parameters of the digital twin model according to the prediction bias value to generate an updated parameter set, writing the updated parameter set into the digital twin model to complete adaptive adjustment comprises: Performing time sequence analysis on the predicted deviation value to obtain a deviation feature set, inputting the deviation feature set into a recurrent neural network for training to obtain an error compensation model, generating a predicted correction amount based on the error compensation model to obtain a compensation parameter table, performing threshold screening on the compensation parameter table to obtain an effective compensation item, performing correction operation on the effective compensation item and an original predicted value to generate a corrected predicted value, and performing parameter sensitivity analysis on a digital twin model according to the corrected predicted value to obtain a sensitivity matrix; And carrying out importance sorting on key parameters of the digital twin model based on the sensitivity matrix to obtain a parameter priority table, mapping a predicted deviation value guided by the parameter priority table to a model parameter space to generate a parameter deviation vector, carrying out parameter identification on the parameter deviation vector through a recursive least square method to obtain an updated parameter set, and writing the updated parameter set into a corresponding position of the digital twin model according to model structure mapping to complete self-adaptive adjustment.
  8. 8. A digital twinning-based building energy consumption optimization management device, the device comprising: The building data acquisition module is used for acquiring temperature and humidity data, energy consumption metering data and personnel density data uploaded by sensors in a building to obtain a multi-source sensing data stream, carrying out anomaly detection and data alignment on the multi-source sensing data stream to obtain a synchronous data set, constructing a digital twin model based on the synchronous data set to obtain a twin body state vector, carrying out time sequence fusion on the twin body state vector and preset forecast data to generate a forecast input matrix, and carrying out load forecast calculation on the forecast input matrix to obtain a future period load sequence; The scene strategy formulation module is used for constructing a multi-objective optimization function based on the future period load sequence to obtain a target equation set, applying model constraint and physical limit constraint to the target equation set to generate a constraint condition set, inputting the constraint condition set into the rolling optimization solver to calculate to obtain a multi-scene control strategy, carrying out weight distribution on the multi-scene control strategy to obtain a fusion control instruction sequence, extracting control parameters of a first time step from the fusion control instruction sequence to generate an execution instruction, and sending the execution instruction to building execution equipment; the energy consumption optimization management module is used for acquiring an actual load value and an operation parameter of the execution equipment to obtain a feedback data set, carrying out error calculation on the feedback data set and a predicted load value to obtain a predicted deviation value, constructing an error compensation model based on the predicted deviation value to complete prediction correction, carrying out online identification on key parameters of the digital twin model according to the predicted deviation value to generate an updated parameter set, and writing the updated parameter set into the digital twin model to complete self-adaptive adjustment.
  9. 9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the digital twinning based building energy consumption optimization management method of any one of claims 1 to 7 when the program is executed by the processor.
  10. 10. A computer readable storage medium having stored thereon a computer program, characterized in that the computer program when executed by a processor implements the steps of the digital twinning based building energy consumption optimization management method of any one of claims 1 to 7.

Description

Building energy consumption optimization management method and device based on digital twin Technical Field The application relates to the field of data processing, in particular to a digital twin-based building energy consumption optimization management method and device. Background The existing building energy consumption optimization management method has obvious defects. The traditional system has poor performance in the aspects of multi-source data processing and twin modeling, and cannot effectively realize accurate prediction of states, so that the optimization effect is affected. Furthermore, the prior art has bottlenecks in optimizing control and policy fusion. Most systems lack perfect constraint handling mechanisms and weight distribution strategies, resulting in less than ideal control results. Existing systems have technology stubs in terms of model tuning. Lack of deep analysis of prediction bias, it is difficult to realize efficient parameter update through online identification, and model accuracy is affected. The solution of the problems has important significance for improving the energy consumption optimizing capability of the building. Disclosure of Invention Aiming at the problems in the prior art, the application provides a digital twin-based building energy consumption optimization management method and device, which can effectively solve the defects of the traditional technology in the aspects of state prediction, strategy optimization, model adjustment and the like and provide technical support for building energy consumption optimization. In order to solve at least one of the problems, the application provides the following technical scheme: in a first aspect, the present application provides a building energy consumption optimization management method based on digital twinning, including: Acquiring temperature and humidity data, energy consumption metering data and personnel density data uploaded by a sensor in a building to obtain a multi-source sensing data stream, carrying out anomaly detection and data alignment on the multi-source sensing data stream to obtain a synchronous data set, constructing a digital twin model based on the synchronous data set to obtain a twin body state vector, carrying out time sequence fusion on the twin body state vector and preset forecast data to generate a forecast input matrix, and carrying out load forecast calculation on the forecast input matrix to obtain a future period load sequence; Constructing a multi-objective optimization function based on the future period load sequence to obtain a target equation set, applying model constraint and physical limit constraint to the target equation set to generate a constraint condition set, inputting the constraint condition set into a rolling optimization solver to calculate to obtain a multi-scene control strategy, performing weight distribution on the multi-scene control strategy to obtain a fusion control instruction sequence, extracting control parameters of a first time step from the fusion control instruction sequence to generate an execution instruction, and issuing the execution instruction to building execution equipment; And acquiring an actual load value and an operation parameter of the execution equipment to obtain a feedback data set, carrying out error calculation on the feedback data set and a predicted load value to obtain a predicted deviation value, constructing an error compensation model based on the predicted deviation value to complete prediction correction, carrying out online identification on key parameters of the digital twin model according to the predicted deviation value to generate an updated parameter set, and writing the updated parameter set into the digital twin model to complete self-adaptive adjustment. Further, the method further comprises the steps of reading analog signals output by a sensor according to a preset sampling frequency through a data acquisition gateway to obtain an original signal stream, carrying out signal separation on the original signal stream to generate a channel data set, carrying out signal quality detection on the channel data set to obtain a data reliability index, and carrying out preprocessing on each channel signal based on the data reliability index to generate a multi-source sensing data stream; Calculating signal autocorrelation values according to sliding windows based on the multi-source sensing data flow, inputting the correlation characteristic sets into an anomaly detector to generate an anomaly marking matrix, removing anomaly data points according to the anomaly marking matrix to obtain an effective data set, calculating time delay deviation among multi-channel signals through a time stamp alignment algorithm, and performing time delay compensation on the effective data set to generate a synchronous data set. Further, the method comprises the steps of carrying out data grouping on a synchronous data set according to a build