6.4 KiB
| title | chunk | source | category | tags | date_saved | instance |
|---|---|---|---|---|---|---|
| Building performance simulation | 2/3 | https://en.wikipedia.org/wiki/Building_performance_simulation | reference | science, encyclopedia | 2026-05-05T13:59:10.370223+00:00 | kb-cron |
== History == The history of BPS is approximately as long as that of computers. The very early developments in this direction started in the late 1950s and early 1960s in the United States and Sweden. During this period, several methods had been introduced for analyzing single system components (e.g. gas boiler) using steady state calculations. The very first reported simulation tool for buildings was BRIS, introduced in 1963 by the Royal Institute of Technology in Stockholm. Until the late 1960s, several models with hourly resolution had been developed focusing on energy assessments and heating/cooling load calculations. This effort resulted in more powerful simulation engines released in the early 1970s, among those were BLAST, DOE-2, ESP-r, HVACSIM+ and TRNSYS. In the United States, the 1970s energy crisis intensified these efforts, as reducing the energy consumption of buildings became an urgent domestic policy interest. The energy crisis also initiated development of U.S. building energy standards, beginning with ASHRAE 90-75. The development of building simulation represents a combined effort between academia, governmental institutions, industry, and professional organizations. Over the past decades the building simulation discipline has matured into a field that offers unique expertise, methods and tools for building performance evaluation. Several review papers and state of the art analysis were carried out during that time giving an overview about the development. In the 1980s, a discussion about future directions for BPS among a group of leading building simulation specialists started. There was a consensus that most of the tools, that had been developed until then, were too rigid in their structure to be able to accommodate the improvements and flexibility that would be called for in the future. Around this time, the very first equation-based building simulation environment ENET was developed, which provided the foundation of SPARK. In 1989, Sahlin and Sowell presented a Neutral Model Format (NMF) for building simulation models, which is used today in the commercial software IDA ICE. Four years later, Klein introduced the Engineering Equation Solver (EES) and in 1997, Mattsson and Elmqvist reported on an international effort to design Modelica. BPS still presents challenges relating to problem representation, support for performance appraisal, enabling operational application, and delivering user education, training, and accreditation. Clarke (2015) describes a future vision of BPS with the following, most important tasks which should be addressed by the global BPS community.
Better concept promotion Standardization of input data and accessibility of model libraries Standard performance assessment procedures Better embedding of BPS in practice Operational support and fault diagnosis with BPS Education, training, and user accreditation
== Accuracy == In the context of building simulation models, error refers to the discrepancy between simulation results and the actual measured performance of the building. There are normally occurring uncertainties in building design and building assessment, which generally stem from approximations in model inputs, such as occupancy behavior. Calibration refers to the process of "tuning" or adjusting assumed simulation model inputs to match observed data from the utilities or Building Management System (BMS). The number of publications dealing with accuracy in building modeling and simulation increased significantly over the past decade. Many papers report large gaps between simulation results and measurements, while other studies show that they can match very well. The reliability of results from BPS depends on many different things, e.g. on the quality of input data, the competence of the simulation engineers and on the applied methods in the simulation engine. An overview about possible causes for the widely discussed performance gap from design stage to operation is given by de Wilde (2014) and a progress report by the Zero Carbon Hub (2013). Both conclude the factors mentioned above as the main uncertainties in BPS. ASHRAE Standard 140-2017 "Standard Method of Test for the Evaluation of Building Energy Analysis Computer Programs (ANSI Approved)" provides a method to validate the technical capability and range of applicability of computer programs to calculate thermal performance. ASHRAE Guideline 4-2014 provides performance indices criteria for model calibration. The performance indices used are normalized mean bias error (NMBE), coefficient of variation (CV) of the root mean square error (RMSE), and R2 (coefficient of determination). ASHRAE recommends a R2 greater than 0.75 for calibrated models. The criteria for NMBE and CV RMSE depends on if measured data is available at a monthly or hourly timescale.
== Technological aspects == Given the complexity of building energy and mass flows, it is generally not possible to find an analytical solution, so the simulation software employs other techniques, such as response function methods, or numerical methods in finite differences or finite volume, as an approximation. Most of today's whole building simulation programs formulate models using imperative programming languages. These languages assign values to variables, declare the sequence of execution of these assignments and change the state of the program, as is done for example in C/C++, Fortran or MATLAB/Simulink. In such programs, model equations are tightly connected to the solution methods, often by making the solution procedure part of the actual model equations. The use of imperative programming languages limits the applicability and extensibility of models. More flexibility offer simulation engines using symbolic Differential Algebraic Equations (DAEs) with general purpose solvers that increase model reuse, transparency and accuracy. Since some of these engines have been developed for more than 20 years (e.g. IDA ICE) and due to the key advantages of equation-based modeling, these simulation engines can be considered as state of the art technology.
== Applications == Building simulation models may be developed for both new or existing buildings. Major use categories of building performance simulation include: