Elsevier

Energy

Volume 160, 1 October 2018, Pages 1101-1114
Energy

Optimal renovation of buildings towards the nearly Zero Energy Building standard

https://doi.org/10.1016/j.energy.2018.07.023Get rights and content

Highlights

  • A model for the optimization of energy renovation of buildings is presented.

  • The model covers both energy supply systems and envelope design.

  • The model is applied to a case study building located in Northern Spain.

  • The designs are limited by the roof availability and economic feasibility.

  • The model can be used for fast design and policy making.

Abstract

In this paper, a Mixed-Integer Linear Programming model is proposed for the design of the energy renovation of existing buildings, considering both Energy Supply Systems and the adoption of Energy Saving Measures to reduce the demand of buildings in retrofitting towards the nearly Zero Energy Building standard. The method is applied to an existing building located in Bilbao (northern Spain), getting the optimal design, i.e. lower annual net cost, for different limits of non-renewable primary energy consumption. The demand reduction produced by the Energy Saving Measures is included as an input from previously validated dynamic simulations and a simple method is presented for its specific distribution in reference days. This simple method, based on degree-days, allows reference days to be generated that, through an Energy Saving Measure based base temperature, consider the weather, the use and the thermal properties' dependency on the distribution of the demand. The optimization method is used to provide the design selection and operation strategy of the renovation of buildings to meet different non-renewable primary energy consumption limits and to provide designs for different constraints: economic, space availability, etc.

Introduction

Nowadays, the world energy scenario is strongly characterized by fossil fuel scarcity and the climate change driven by their use. This reality requires an energy transition consisting of lowering primary energy consumption, increasing energy efficiency and promoting renewable energy sources. In this context, the European Union (EU) is playing a major role towards this transition, enforcing several Directives that seek the reduction of the energy consumption of different sectors. Amongst these sectors, buildings are responsible for 40% of the overall primary energy consumption of the EU [1], and the Directive 2010/31/EU on the energy performance of buildings (EPBD) [2] aims to cut this consumption. In line with this goal, the EPBD requires all new buildings and part of the existing stock to be nearly Zero Energy Buildings (nZEB).

An nZEB is defined as a building of very high energy performance and the nearly zero amount of energy required should be covered to a very significant extent by energy from renewable sources. In this context, a Zero Energy Building (ZEB) would be an nZEB with no net Non-Renewable Primary Energy (NRPE) consumption. The specific definition of nZEB is intended for independent implementation by EU Member States according to a maximum Non-Renewable Primary Energy (NRPE) consumption value defined by the analysis of the cost optimality concept.

Thus, the design of nZEBs will be based on optimization routines that will give rise to best designs constrained to certain legal and technical specifications. These designs will arise from the evaluation of different combinations of technologies and their associated operation strategy. Specifically, building design optimization can be split into two levels: (i) load level, acting on the envelope and ventilation by Energy Saving Methods (ESM); or (ii) consumption level, acting on the Energy Supply Systems (ESS).

Regarding the load level, Huang and Niu presented a comprehensive review on the optimization of building envelope [3]. Different optimization techniques were identified, genetic algorithms, direct search and neural networks being the most used approaches. In relation to consumption level optimization, Attia et al. [4] reviewed the existing literature and concluded that simple genetic algorithms are the most adequate and widespread technique for solving multiple design, operation and control optimization problems with relative ease.

As can be seen from the literature reviews presented above, while many authors have carried out optimization work at one of these two levels, the combined optimization (i.e. considering both envelope and energy supply systems) is less common. Some of the most relevant studies dealing with this integral optimization analysis couple genetic algorithms with state-of-the-art building energy simulation programs. Ferrara et al. used a computing environment that combines TRNSYS, a building simulation environment, with GenOpt®, a generic optimization program, to get cost-optimal designs for a case study situated in France [5]. Analogously, Ascione et al. analyzed the optimal energy retrofit of a hospital, situated in southern Italy, using genetic algorithms through the combination of EnergyPlus and MATLAB [6]. Also worth mentioning is the approach presented by Hamdy et al. [7]. This consists of an optimization scheme that, at the same time, combines a genetic algorithm with detailed simulation programs, divides the whole process into three stages showing the effect of the optimal and speeds up the exploration by avoiding the unfeasible design-variable combinations and using pre-simulated results. More recently, D'Agostino et al. [8] developed a framework for cost-optimal nZEBs based on energy and cost simulations using a sequential search. However, these approaches require a detailed definition of the building and the solutions to be previously defined ad-hoc by the user, i.e. system configuration and operation strategy. In this sense, Hamdy et al. [9] comprehensively compared the performance of seven evolutionary optimization algorithms and they conclude that algorithm selection and settings might involve trial and error. Therefore, simulation-based approaches could get computationally costly when the whole parameter tuning process is considered. To reduce the computation expense, Gilles et al. used a Kriging model surrogate NZEB performance criteria during the optimization process and a genetic algorithm is considered efficient to find the global optimal solutions [10].

An alternative to overcome these disadvantages are exact optimization models, such as Mixed-Integer Linear Programming (MILP). MILP-based optimization problems have largely been applied to the optimal design and operation of energy supply systems in building at different levels. Mehleri et al. presented a MILP model for the optimal design of distributed energy generation systems for neighborhoods, which allowed the optimal selection of the system components among several candidate technologies, including the optimal design of a heating pipeline network, that allows heat exchange among the different nodes [11]. Omu et al. created a MILP model for the design of a distributed energy system that meets the electricity and heating demands of a cluster of commercial and residential buildings while minimising annual investment and operating cost [12]. Analogously, Milan et al. developed a MILP-based cost optimization model for the design of 100% renewable residential energy supply systems [13]. A number of recent papers have considered MILP-based approaches to optimize more specific configurations such as power and heat interchanges [14] or on-site renewable technologies with storage [15]. However, there is a lack of research applying these techniques to envelope optimization and, even less, integral optimization of envelope and energy supply systems. This is mainly explained by the high nonlinearity introduced into the optimization problem by different envelope solutions. Ashouri et al. presented a model for ESSs that also considers a building's thermal mass as an additional storage option [16]. However, it used constant thermal loads that were computed before the optimization was executed, not allowing a proper optimization of the envelope. Milic et al. [17] analyzed the performance of an in-house LCC optimization software, OPERA-MILP. The aim is fulfilled through comparison with building energy simulation software IDA ICE before and after cost-optimal energy renovation. MILP modelling has been also used for the simultaneous design and operation of urban energy systems considering a flexible value web framework for representing integrated networks of resources and technologies [18]. A very interesting solution to this problem has recently been proposed by Schütz et al. [19], who proposed a building model suitable for MILP, based on ISO 13790 and validated according to ASHRAE 140.

The authors recently presented a simple method dealing with the ESS optimization of buildings [20]. The method was based on a general superstructure that made it possible to include any existing or future technologies, covering heating, Domestic Hot Water (DHW), cooling and electricity. The model was linked to a Mixed Integer Linear Programming (MILP) problem that allowed the selection of equipment and its operation for a given set of load profiles. The operation included the unit commitment of technologies with limited load regulation capacity through binary control and the time horizon was discretized in a set of independent reference days, allowing the model to be optimized by state-of-the-art solvers within acceptable calculation times. The model allowed the annual cost minimization based on the net present value for a set of constraints imposed by the designer, such as NRPE consumption limits.

The main objective of this paper is to extend the optimization model to allow an optimization that considers both Energy Supply Systems and Energy Saving Measures alternatives for certain NRPE consumption limits. Thus, the method presented here contributes to the need for simple and reliable tools toward nZEB design and operation. The proposed envelope alternatives will be treated as virtual technologies that produce, at no energy cost, the heating demand reduction caused by their implementation. For this purpose, the different yearly demands are previously defined as inputs to the method, which were obtained from the simulation of a validated TRNSYS model. As previously stated, the MILP method presented by the authors discretized the aggregated thermal demand into a set of reference days. In this paper, the distribution of the input loads into reference days of the different envelope alternatives is obtained from a degree-days based method, using a variable base temperature.

The method is applied to achieve the optimal integral renovation of an existing building located in Bilbao (Northern Spain). A set of 7 envelope retrofitting solutions or ESMs were considered, while the main technologies available in the market have been included in the model using self-tailored cost estimation models. Then, the optimal ESS configuration and ESM solution, which minimize the annual cost, have been obtained for three NRPE consumption limits: cost optimal and two Zero Energy Building scenarios, depending on whether the electric consumption of the users is accounted (ZEB) or not (ZEB′). Only heating, DHW and electricity loads are considered in the case study, which are the typical loads in northern Spain. Cooling load would be included in the same way, but the application of the method to buildings under hot weather conditions will be analyzed by further research.

The rest of the paper is organized in 4 sections, as follows: Section 2 deals with the energy and economic modelling of the building renovation and presents the optimization method. This model is applied to the case study covered by Section 3, giving rise to the optimal configuration and operation in terms of energy and economic performance, as appears in Section 4. Finally, the main contributions of the paper and future research are summarized in Section 5.

Section snippets

Materials and Methods

This section presents the method for the optimal renovation of buildings considering, from an energy and economic perspective, both ESMs and ESSs.

Case study

The proposed method is applied to a multi-family building located in a district of Otxarkoaga, in the city of Bilbao (northern Spain). The district was built in 1959–1961 and can be considered representative of the existing building stock in the region. The building comprises 36 dwelling units, each of which has a net floor area of 55 m2, giving rise to an overall area of 1980 m2 [27]. A picture of the building and its surroundings are presented in Fig. 4.

The external walls of the dwelling are

Results and discussion

The model has been generated and solved with CPLEX v12.6.2 [31] within MatLab R2014a [32]. A computer with Intel Core i5-2430 M CPU @ 2.40 GHz processor and 8 GB of RAM was used for the resolution. The problem consists of 58,858 constraints, 37,440 continuous variables, and 6970 integer variables, from which 6919 are binary ones. Elapsed time is case dependent. The following stop criterion was taken: simulations were run until a gap of 1% or, alternatively, 2 h of operation were reached. In no

Conclusions

The energy renovation of buildings is expected to play a key role in reducing energy dependency and mitigating climate change, especially due to the great stock of existing low efficiency buildings. Even though the nZEB concept has recently been introduced for new buildings, the dominant urgency of changes will require its application to be extended to the renovation of existing ones. In this paper, a method has been proposed for the design of energy renovation, considering ESSs and the

Acknowledgements

This research has been partially supported by the project MTM2015-65317-P from the Spanish Ministry of Economy and Competitiveness and “Fundación Iberdrola” by the Research Grant in Energy and Environment 2017.

References (36)

Cited by (0)

View full text