Research On Stochastic Optimization Design Of Building Energy System Under Uncertain Boundaries

Along with the supply-side reform of the energy system deepens,renewable energy will play an important role in future energy supply system.With the promotion of the energy demand-side reform,the role of buildings will change and move from the "energy-consumption era" to the "empowerment era".In the context of energy supply and demand reform,the energy production-energy conversion-energy demand should be considered in the design of building energy systems.In the near future,the complexity of building energy systems will become the design dilemma that designers need to address first.In addition,the building energy system design process will face uncertainties of renewable energy production and energy demand.With the reform of energy supply and demand,the problem of uncertainty will be more prominent,and the traditional deterministic optimization design method may face economic and reliability risks.In view of the above problems,this paper studies the design method of building energy systems under the uncertain design boundaries.The main work is as follows:(1)In order to deal with the complexity of building energy system design,three types of building energy system optimization design models with different design complexity were developed.The design and operation conditions are integrated in these models and then solved using an accurate optimization algorithm.By comparing with the original design scheme,the effectiveness and superiority of the optimal design model in the optimization design of the building energy system are verified.(2)To generate a complete set of uncertain boundary scenarios,the Monte Carlo simulation method combined with information entropy theory is proposed.A Monte Carlo simulation framework for generating random boundary scenarios is first established.Then,relative entropy is used as a Monte Carlo simulation convergence index,and Monte Carlo simulation convergence is used to determine the completeness of uncertain information.Based on the complete set of uncertain boundary scenarios,the influence of design boundary uncertainty on the optimal design goals and design schemes was explored.The study found that uncertainty will have a significant impact on optimal objectives and configurations of building energy system.There is a risk that the building energy system which optimized based on a single deterministic scenario will be unreliable.Therefore,the uncertainty of the design boundary conditions should be considered during the design phase of the building energy system to improve the economy and reliability of the energy system.(3)Facing the complexity of building energy system design and load uncertainty,a stochastic programming model considering one-dimensional uncertainty on the load side is proposed.In order to consider complete uncertainty information of the load scenario set in the stochastic programming model and ensure the model's solvability and efficient solvability,this paper proposed a time-frequency domain data reduction method for the random load scenario set-named Bin method.The advantage of the Bin method is that the lossless of uncertainty information can be guaranteed during the dimensionality reduction from time domain to frequency domain,that is,the dimensionality reduction is effective,and the data dimension after dimensionality reduction has nothing to do with the number of random scenarios,that is,the dimensionality reduction is efficient;Further consider the source-side uncertainty of the building energy system caused by renewable energy output,this paper proposes a stochastic programming method that takes into account the dual source-load uncertainties and the time-series correlation.Based on the Bin method,the Bi-Bin method is further developed to reduce the dimension of the source / load uncertainty scenarios.Compared with the Bin method,the Bi-Bin method can consider the timecorrelation between the source and the load,and ensure the validity of the timefrequency domain data reduction process.Through an example,the validity and efficiency of the above-mentioned stochastic programming method are verified.By comparing with the deterministic optimization design scheme,the advantages of the stochastic programming method in realizing the economic and reliability of the building energy system design scheme are verified.(4)Energy storage can improve the flexibility of the energy system.Meanwhile,the randomness of the energy storage strategy increases the uncertainty of the energy system design process.Facing the multi-dimensional uncertainty of source,load and energy storage,this paper proposes a stochastic programming model for building energy systems with multi-dimensional uncertainties.The double-scene reduction technology for typical years and typical days is constructed to ensure the solvability of the stochastic programming model while taking into account the uncertainty of the design boundary conditions.Because the stochastic programming model considers the uncertainty of the boundary conditions,it can maximize the life-cycle economic benefits of the building energy system under the uncertain design conditions.This paper further discusses the ability of energy storage to handle the uncertainties of source and load.The results show that the energy storage flexibility is able to resist a certain degree of load forecasting bias.Therefore,in the face of the uncertainty of the future load and the output of renewable energy,it is necessary to consider the planning of the energy storage system to improve the flexibility of the building energy system,and increase the system's ability to combat uncertainty.