A decision support method for optimizing building energy-saving transformation
Through the low-order white box simulation engine and multi-objective optimization method based on resistor-capacitance model, the problems of medium and high calculation costs and uncertainty of climate change are solved, automated decision-making support is achieved, and the cost of building renovation is reduced and the cost of building renovation is improved. Transparency in decision making.
Patent Information
- Application Number
- CN202111333416.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-11-11
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2041-11-11
AI Technical Summary
The high calculation cost in the existing research on energy conservation optimization of building has been limited in its application and promotion in actual energy conservation transformation projects, and it is difficult to effectively deal with the uncertainty of future climate change on building energy consumption.
Using a low-order white box building simulation engine based on resistor-capacitance model, combining marginal emission reduction cost analysis and sensitivity analysis, multi-objective optimization is carried out, and decision-making support paths are generated using hierarchical clustering algorithms to reduce calculation costs and consider climate change for buildings. The impact of energy consumption.
A multi-objective optimization solution for automated building renovation in the face of future climate change scenarios has been realized, which significantly reduces computing costs, provides a transparent and intuitive decision-making process, and has the potential to be widely used in energy-saving renovation projects of single buildings.
Smart Images

Figure QLYQS_1 
Figure QLYQS_4 
Figure QLYQS_8
Abstract
Description
Technical Field
[0001] The invention belongs to the field of building energy-saving transformation, and specifically relates to a building energy-saving transformation optimization decision support method. Background Art
[0002] Energy-saving renovation has become an important means for existing buildings to improve building energy efficiency, reduce carbon emissions, and improve the quality of the living environment. With the intensification of climate change, more and more studies have shown that climate change will have a greater impact and uncertainty on future building energy consumption levels. Among them, under future climate conditions, changes in outdoor temperature and humidity have the most significant impact on building energy consumption. This trend of change has put forward new requirements and challenges for existing buildings to adapt to future climate conditions. Therefore, under the trend of climate change, how existing buildings can ensure and achieve the excellent performance of building energy-saving renovation (that is, the climate adaptability of existing building energy-saving renovation) is an issue worthy of in-depth discussion.
[0003] With the rapid development of computer-aided technology, building performance simulation technology and its tools, the powerful computing power of computers can be used to optimize building energy consumption. Li Zhengrong et al. used EnergyPlus to conduct modeling analysis. Through the orthogonal experimental design method, they studied the influence of the heat transfer coefficient of the envelope structure on the heating / cooling load of the building in the hot summer and cold winter areas, and obtained the optimal envelope structure design scheme. Wang Yilin et al. used DeST-H to simulate and analyze building energy consumption, studied the influence of building energy-saving renovation schemes of four insulation materials at different thicknesses on building energy consumption, and combined with economic analysis, provided a reasonable external envelope structure renovation scheme. Yu Zhenyu et al. conducted a near-zero energy building performance optimization design analysis based on genetic algorithms and with the building energy consumption simulation software TRNSYS. However, current building optimization research mostly uses complex dynamic simulation models (such as Energy Plus, etc.) to evaluate energy-saving effects, which leads to huge computational costs in the process of building energy-saving optimization. The high computational cost is not conducive to the application and promotion of such methods in actual energy-saving renovation projects, nor is it conducive to the deduction and implementation of the final renovation plan. Summary of the invention
[0004] In view of the problems existing in the above-mentioned prior art, the present invention proposes a practical automatic calculation scheme, which integrates a building feature selection method based on marginal emission reduction cost analysis and sensitivity analysis (based on variance); a multi-objective optimization problem solved based on non-dominated sorting differential optimization; and a method for providing decision support is proposed. The building energy consumption performance simulation engine uses the low-order white box building simulation engine based on resistance and capacitance developed by the team in the early stage, which has a lower calculation cost compared with the traditional white box model. With the help of data-driven models, the currently verified and applied global climate model is used to consider the uncertainty of the future energy consumption of buildings affected by climate change. Further exploration and use of hierarchical clustering algorithms provide decision support for high-dimensional multi-objective optimization algorithms (Pareto frontiers).
[0005] The present invention is achieved through the following technical solutions:
[0006] A building energy-saving transformation optimization decision support method comprises the following steps:
[0007] S1, building energy consumption model based on resistance-capacitance model theory;
[0008] S2, predicting future hourly meteorological data, inputting the future hourly meteorological data into the building energy consumption model, obtaining future hourly energy consumption, and training the future hourly energy consumption simulation results to generate a data-driven model;
[0009] S3, select pre-optimization decision variables, calculate the future hourly energy consumption and net present value of the life cycle of future hourly meteorological data under each pre-optimization decision variable; analyze and determine the optimal decision variables based on the net present value;
[0010] S4, determine the multi-objective optimization problem, use the differential evolution algorithm of non-dominated solution sorting to perform multi-objective optimization on the selected decision variables;
[0011] S5, a hierarchical clustering algorithm is used to automatically cluster the generated multi-objective optimization results, and a visual decision support path is generated based on the parallel coordinate diagram.
[0012] As one of the preferred solutions of the present invention, in step S1, a low-level white box modeling tool based on a resistor-capacitor analog simulation framework is used.
[0013] As one of the preferred solutions of the present invention, in step S3, pre-optimization decision variables are selected based on energy-saving measures, and the pre-optimization decision variables include thermal performance of the building envelope, air tightness, natural ventilation, set temperatures for heating and cooling, and set temperature fallback during vacancy time.
[0014] As one of the preferred schemes of the present invention, in step S3, the preferred decision variables are analyzed and determined, specifically: S31, using the marginal emission reduction cost analysis method to analyze the impact of each pre-optimization decision variable on the selected objective function; S32, using the variance-based sensitivity analysis method to analyze the impact of each pre-optimization decision variable on the system.
[0015] As one of the preferred solutions of the present invention, in step S4, the multi-objective optimization includes four objectives, namely, energy saving, economic benefit, investment cost, and the sum of the absolute values of PMV in each partition of the building at each hour during the entire life cycle of the renovated building.
[0016] As one of the preferred solutions of the present invention, in step S4, the calculation formulas for energy saving, economic benefit and investment cost are respectively:
[0017] Energy saving:
[0018] Where: E pre,k and E post,k are the annual energy consumption of the unrenovated building and the renovated building in year k, respectively; L represents the entire life cycle of the renovated building;
[0019] Economic benefits:
[0020] Where: τ u is the increase ratio of u-shaped energy cost; C u post,k and C u pre,k are the costs of u-type energy in the kth year of the life cycle of the retrofitted building and the unretrofitted building, respectively;
[0021] Investment cost:
[0022] Where: τ m is the annual maintenance cost growth rate (%); r is the discount rate (%), I o is the initial investment cost.
[0023] As one of the preferred solutions of the present invention, step S5 is specifically:
[0024] S51, obtain the Pareto front and cluster the solution set based on the hierarchical clustering algorithm;
[0025] S52, plotting parallel coordinate axes for the optimal decision variables and multiple objectives to obtain a parallel coordinate graph;
[0026] S53, establish a balanced building renovation decision-making thinking model to simulate the decision-making process, and select the corresponding cluster according to the thinking model to cluster the next level solution set, and so on to obtain the final solution set;
[0027] S54, visualize the simulation results and generate a decision support path.
[0028] The beneficial effects of the present invention are:
[0029] The present invention proposes an automated multi-objective optimization scheme for building renovation facing future climate change scenarios. The scheme includes three main stages: pre-optimization feature selection, multi-objective optimization, and post-optimization process and decision-making based on the Pareto front. The core engine of building simulation is a lightweight simulation tool developed by the team in the early stage based on the resistor-capacitor modeling scheme, which can greatly speed up the speed of building simulation. Although the decision-making process usually involves subjectivity, the scheme can visualize, retain and restore the entire decision path, making the decision-making process of multi-objective solutions more transparent and intuitive. The scheme has the potential to be widely used in single building energy-saving renovation projects, and at the same time has practical significance and application value for the rapid optimization and decision-making of actual building energy-saving renovation projects. BRIEF DESCRIPTION OF THE DRAWINGS
[0030] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0031] Figure 1 This is a real picture of the LRSM building described in the embodiment;
[0032] Figure 2 A top view of the area where the LRSM building described in the embodiment is located;
[0033] Figure 3 is the marginal abatement cost curve of LRSM;
[0034] Figure 4 The results of the Sobol sensitivity analysis for LRSM - the first-order Sobol index;
[0035] Figure 5 The results of Sobol sensitivity analysis for LRSM - the second-order Sobol index;
[0036] Figure 6 This is the LRSM decision path described in the embodiment. DETAILED DESCRIPTION
[0037] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments.
[0038] The present invention provides a building energy-saving transformation optimization decision support method, which includes the following core steps:
[0039] S1, building energy consumption model based on resistance-capacitance model theory;
[0040] S2, predicting future hourly meteorological data, inputting the future hourly meteorological data into the building energy consumption model, obtaining future hourly energy consumption, and training the future hourly energy consumption simulation results to generate a data-driven model;
[0041] S3, select pre-optimization decision variables, calculate the future hourly energy consumption and net present value of the life cycle of future hourly meteorological data under each pre-optimization decision variable; analyze and determine the optimal decision variables based on the net present value;
[0042] S4, determine the multi-objective optimization problem, use the differential evolution algorithm of non-dominated solution sorting to perform multi-objective optimization on the selected decision variables;
[0043] S5, a hierarchical clustering algorithm is used to automatically cluster the generated multi-objective optimization results, and a visual decision support path is generated based on the parallel coordinate diagram.
[0044] Specifically, in step S1, the building energy consumption modeling and simulation method adopts a low-order white box modeling tool based on a resistor-capacitor analog simulation framework developed by the inventor in the early stage. The specific method and steps of the tool are shown in the paper - Shen P, Braham W, Yi Y. Development of a lightweight building simulation tool using simplified zone thermal coupling for fast parametric study. Applied energy, 2018, 223: 188-214, which will not be repeated here.
[0045] In step S2, future hourly meteorological data are obtained through existing tools, the random forest algorithm is used to predict future hourly energy consumption, and the radio frequency algorithm is used to train the hourly energy consumption simulation results of extreme meteorological years to generate a data-driven model.
[0046] In step S3, a pre-optimization decision variable (ECM) is selected based on energy-saving measures, most of which are passive energy-saving measures, such as improving the thermal performance of the building envelope, air tightness, natural ventilation, set temperatures for heating and cooling, and set temperature setbacks during vacancy time, etc. Non-passive measures include improving lighting efficiency.
[0047] The replacement costs of different types of windows are shown in Table 1 , the ECMs that require monetary investment are shown in Table 2 , and the ECMs that do not require any cost are shown in Table 3 .
[0048] Table 1 Attributes and costs of window replacement
[0049]
[0050] Table 2 ECM parameters and required costs
[0051]
[0052] Table 3 Cost-free ECM parameters
[0053]
[0054] In step S3, the optimal decision variables are analyzed and determined, specifically:
[0055] S31, the marginal abatement cost analysis method is used to analyze the impact of each pre-optimization decision variable on the selected objective function. Before the sensitivity analysis of feature selection, in order to prevent those input variables that may reduce the performance of the building from entering the optimization procedure, a pre-measure analysis is performed. Marginal abatement cost analysis (MAC) has been widely used to weigh the trade-off between environmental effects and relative costs. The MAC curve is obtained by evaluating the abatement measures individually.
[0056] S32, variance-based sensitivity analysis (Sobol) method, analyzes the impact of each pre-optimized decision variable on the system. In this embodiment, Saltelli's extension of the Sobol method is used to generate samples. The solution proposed by Saltelli extends the sequence of the Sobol method and can reduce the error rate in the resulting sensitivity index calculation.
[0057] In step S4, the multi-objective optimization includes four objectives, namely, energy saving, economic benefits, investment costs, and the sum of the absolute values of PMV in each partition of the building during the entire life cycle of the renovated building. Each ECM (input variable) is preprocessed into a continuous sequence that obeys a uniform distribution. In order to make different objectives comparable, the conversion factor F is used to convert the total energy saving (Es, GJ) into initial energy consumption.
[0058] Energy saving:
[0059]
[0060] Where: E pre,k and E post,k are the annual energy consumption of the unrenovated building and the renovated building in year k, respectively; L represents the entire life cycle of the renovated building;
[0061] The economic benefit (S,$) is calculated by the difference in energy efficiency cost of the existing building before and after renovation over the entire life cycle, using the following formula:
[0062]
[0063] Where: τ u is the increase ratio of u-shaped energy cost; C u post,k and C u pre,k are the costs of u-type energy in the kth year of the life cycle of the retrofitted building and the unretrofitted building, respectively;
[0064] The calculation of the investment cost (I,$) also takes into account the maintenance costs that occur every five years, which are assumed to be 15% of the total investment cost. Therefore, the investment cost can be calculated as:
[0065]
[0066] Where: τ m is the annual maintenance cost growth rate (%); r is the discount rate (%), I o is the initial investment cost. Assumed to be proportional to the initial investment of each ECM. Assumed discount rate is 4%.
[0067] Step S5 is specifically as follows:
[0068] S51, obtain the Pareto front and cluster the solution set based on the hierarchical clustering algorithm;
[0069] S52, plotting parallel coordinate axes for the optimal decision variables and multiple objectives to obtain a parallel coordinate graph;
[0070] S53, establish a balanced building renovation decision-making thinking model to simulate the decision-making process, and select the corresponding cluster according to the thinking model to cluster the next level solution set, and so on to obtain the final solution set;
[0071] S54, visualize the simulation results and generate a decision support path.
[0072] The building energy-saving renovation optimization decision support method of the present invention is described in detail below with reference to specific cases.
[0073] The case study building is the Laboratory for Research on the Structure of Matter (LRSM) at the University of Pennsylvania, which has a floor area of 7781.4m 2 ,like Figure 1 , Figure 2 shown.
[0074] First, the building was modeled using a low-level white-box modeling tool based on a resistor-capacitor analogy simulation framework, and the model was validated using the collected monthly building energy consumption inventory for 2015. Since all buildings on campus use condensed water and steam for cooling and heating, the building energy consumption model was validated using metered condensed water and steam.
[0075] Secondly, the future hourly meteorological data of the building area is obtained and input into the building energy consumption model, the random forest algorithm is used to predict the future hourly energy consumption, and the radio frequency algorithm is used to train the hourly energy consumption simulation results of extreme meteorological years to generate a data-driven model. In this embodiment, the life cycle of the renovated building is assumed to be 20 years.
[0076] Again, perform feature selection and sensitivity analysis:
[0077] MAC analysis: Simulate all ECMs and parameters individually to calculate the marginal abatement cost. There are 84 parameters in this analysis, including the installation area and inclination of the PV system. The calculation results are as follows: Figure 3 As shown in the figure. The x-axis is the total energy saving for the whole year (E s ,GJ), and the y-axis is "cost" - NPV ($ / GJ). Each ECM is ranked according to its unit cost performance (from left to right). It can be seen that if the building does not use more energy in the future, those ECMs with no monetary investment will have positive marginal abatement costs. At the same time, ECMs with negative marginal abatement costs will be eliminated in this stage. The ECM ranking of LRSM is natural ventilation, daylighting, lighting system upgrade, cooling set temperature, cooling supply air temperature, vacancy time set temperature fallback, heating set temperature, shading, window replacement, air infiltration, photovoltaic panel tilt, exterior wall insulation and photovoltaic panel area.
[0078] Analysis of Sobol results:
[0079] In the MAC analysis, it was found that the optimal inclination angle of PV was 30°, so in the Sobol analysis stage, in addition to the ECM-related parameters excluded after MAC analysis, the PV angle was also excluded. Compared with other ECM parameters that may affect each other, the parameters of the PV system are a relatively independent factor, so the PV angle and PV area are combined. Finally, a variance-based sensitivity analysis was performed on the 12 ECMs automatically generated by MAC analysis involved in LRSM. The analysis results are shown in Figure 2. Figure 4 , Figure 5 Table 4 lists the top 7 ECMs.
[0080] Table 4 Seven ECMs finally selected for LRSM
[0081]
[0082] Finally, the decision based on the clustering method:
[0083] After obtaining the Pareto front, a parallel coordinate diagram is further drawn to visualize the results to support decision-making. To analyze the distribution of the Pareto front, the parameters of the selected ECM and the four sub-goals are drawn in parallel in the parallel coordinate diagram. A balanced building renovation decision-making thinking model is established to simulate the decision-making process. This thinking model refers to selecting building renovation plans in pursuit of higher cost-effectiveness and a more comfortable indoor thermal environment under limited fiscal policies.
[0084] The LRSM decision path production process under the neutral building renovation thinking mode includes multiple layers of clustering, and the preferred clusters in each layer of clustering may be different. For example, 4 clusters and 3 layers of clustering are set. In the first layer of clustering, cluster 4 will be the first choice due to lower economic investment and optimal indoor thermal comfort. At the same time, considering the same reasons, clusters 2 and 3 are selected in the second and third layers of clustering, respectively. After comprehensive consideration, the path corresponding to cluster 3 can be finally selected as the final decision path, such as Figure 6 shown.
[0085] Final decision: Table 5 lists the top five solutions for energy saving payback for each building. The selected ECMs include common parameter selections, such as cooling supply air temperature of 17°C, cooling set temperature of 25°C, and no vacancy time set temperature fallback.
[0086] Table 5 shows the top 5 ECM combinations selected by LRSM
[0087]
[0088] The embodiments described above are merely descriptions of preferred implementations of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope of the present invention.
Claims
1. A building energy-saving transformation optimization decision support method, characterized in that: The following steps are involved: S1, building energy consumption model based on resistance-capacitance model theory; S2, predicting future hourly meteorological data, inputting the future hourly meteorological data into the building energy consumption model, obtaining future hourly energy consumption, and training the future hourly energy consumption simulation results to generate a data-driven model; S3, select pre-optimization decision variables, calculate the future hourly energy consumption and net present value of the life cycle of future hourly meteorological data under each pre-optimization decision variable; analyze and determine the optimal decision variables based on the net present value; S4, determine the multi-objective optimization problem, use the differential evolution algorithm of non-dominated solution sorting to perform multi-objective optimization on the selected decision variables; S5, using a hierarchical clustering algorithm to automatically cluster the generated multi-objective optimization results, and generating a visual decision support path based on a parallel coordinate diagram; In step S3, the optimal decision variables are analyzed and determined, specifically: S31, using the marginal abatement cost analysis method to analyze the impact of each pre-optimization decision variable on the selected objective function; S32, variance-based sensitivity analysis method, analyzes the impact of each pre-optimization decision variable on the system; Step S5 is specifically as follows: S51, obtain the Pareto front and cluster the solution set based on the hierarchical clustering algorithm; S52, plotting parallel coordinate axes for the optimal decision variables and multiple objectives to obtain a parallel coordinate graph; S53, establish a balanced building renovation decision-making thinking model to simulate the decision-making process, and select the corresponding cluster according to the thinking model to cluster the next level solution set, and so on to obtain the final solution set; S54, visualize the simulation results and generate a decision support path.
2. The building energy-saving transformation optimization decision support method according to claim 1 is characterized in that: In step S1, a low-level white box modeling tool based on a resistor-capacitor analog simulation framework is used.
3. The building energy-saving transformation optimization decision support method according to claim 2 is characterized in that: In step S3, pre-optimization decision variables are selected based on energy-saving measures, and the pre-optimization decision variables include thermal performance of the building envelope, air tightness, natural ventilation, set temperatures for heating and cooling, and set temperature fallback during vacancy time.
4. The building energy-saving transformation optimization decision support method according to claim 1 is characterized in that: In step S4, the multi-objective optimization includes four objectives, namely, energy saving, economic benefit, investment cost, and the sum of the absolute values of PMV in each partition of the building at each hour during the entire life cycle of the renovated building.
5. The building energy-saving transformation optimization decision support method according to claim 4 is characterized in that: In step S4, the calculation formulas for energy saving, economic benefit and investment cost are: Energy saving: (1) Where: and They are k Annual energy consumption of unrenovated buildings and renovated buildings; L Represents the entire life cycle of a renovated building; Economic benefits: (2) Where: yes u The proportion of increase in energy costs; and They are respectively the first k years u the cost of renewable energy; Investment cost: (3) Where: is the annual growth rate of maintenance costs (%); r is the discount rate (%), I o is the initial investment cost.
