Kostensimulationsmodell zur probabilistischen Lebenszykluskostenermittlung auf Basis digitaler Gebäudemodelle mithilfe einer Monte-Carlo-Simulation

Spinnräker, Eric Wilhelm; van Treeck, Christoph Alban (Thesis advisor); Lützkendorf, Thomas (Thesis advisor)

Aachen (2020)
Dissertation / PhD Thesis

Dissertation, Rheinisch-Westfälische Technische Hochschule Aachen, 2020


The construction industry is dominated by a traditional trade-oriented way of thinking and acting with the focus on „fast and cheap“ construction. This leads to the planning and construction of buildings that are hampered by unnecessarily high operating costs, often function poorly and consume a lot of resources. Furthermore, the insufficient identification and communication of existing uncertainties in the planning phase leads to time delays and an exceeding of costs. The increased usage of digital planning methods such as Building Information Modeling (BIM) as well as the early consideration of life cycle costs and the integration of systematic risk management are central recommendations to solve this problem. However, current calculation models from computer sciences in engineering do not adequately address the above-mentioned aspects, especially not in combination. The calculation and presentation of life-cycle costs can be used in particular to optimise construction design, to compare design variants, for budgeting and to assess economic quality in the context of sustainability assessments. Thus, this thesis proposes a web-based cost simulation model for the probabilistic life cycle cost calculation, BIM5Dplus, based on digital building models. The calculation is based on the net present value method and Monte Carlo simulation. The calculations are implemented on the server-side in Python, and the required computing effort can be minimized by using different methods for drawing samples, such as random sampling, Latin hypercube sampling, or sampling based on Sobol sequences. In addition, the modular design allows cost simulations in the various planning phases at different levels of granularity. For early phases, it is possible to provide statistical planning and cost parameters using a few input parameters. A presented process for the derivation of suitable probability functions based on the least squares method allows an early probabilistic cost estimation. In addition, an IFC interface for transferring data from digital models with the aim of automated quantity determination in accordance with the German standard DIN 276 was implemented to meet the requirements of increasing digital planning using BIM. The IFC interface is validated in this thesis using three test instances. The cost simulation model is used for a total of four examples to illustrate possible applications. The applicationcan help to improve the planning process with regard to the holistic view of structures and the integration of existing uncertainties and to facilitate decision-making.