Integrating Economy, Energy, Air Pollution in Building Renovation Plans

1st IFAC Workshop on Integrated Assessment Modelling 1st IFAC WorkshopSystems on Integrated Assessment Modelling for Environmental IFAC WorkshopSystems on Integrated Assessment Modelling 1st for Environmental University of Brescia, Italy, May 10-11,Modelling 2018 Available online at www.sciencedirect.com 1st IFAC on Integrated Assessment for Environmental Systems 1st IFAC Workshop Workshop onBrescia, Integrated Assessment University of Brescia, Italy, May 10-11,Modelling 2018 1st Environmental IFAC WorkshopSystems onBrescia, Integrated Assessment Modelling for University of Brescia, Brescia, Italy, May 10-11, 2018 for Environmental Systems for Environmental Systems University University of of Brescia, Brescia, Brescia, Brescia, Italy, Italy, May May 10-11, 10-11, 2018 2018 University of Brescia, Brescia, Italy, May 10-11, 2018

ScienceDirect

IFAC PapersOnLine 51-5 (2018) 102–107 Integrating Economy, Energy, Air Pollution in Building Renovation Plans Integrating Economy, Integrating Economy, Energy, Energy, Air Air Pollution Pollution in in Building Building Renovation Renovation Plans Plans Economy, Energy, Air Pollution in Building Renovation Plans Integrating Giorgio Guariso*, Matteo Sangiorgio** Integrating Economy, Energy, Air Pollution in Building Renovation Plans Giorgio Guariso*, Matteo Sangiorgio**

Giorgio Guariso*, Matteo Sangiorgio** Giorgio Giorgio Guariso*, Guariso*, Matteo Matteo Sangiorgio** Sangiorgio** Guariso*, Sangiorgio** Department of Electronics Information Giorgio and Bioengineering, Politecnico di Milano, Via Ponzio 34/5, 20133 Milano, Italy  Matteo Department of Electronics Information and Bioengineering, Politecnico di Milano, Via Ponzio 34/5, 20133 Milano, Italy  Politecnico di Milano, Via Ponzio 34/5, 20133 Milano, Italy Department of Electronics Information and Bioengineering, Department of Electronics Information and Bioengineering, Politecnico di Milano, Via Ponzio 34/5, 20133 Milano, Italy Department Bioengineering, di *Tel: 02and 2399-3559; e-mail: Politecnico [email protected]). Department of of Electronics Electronics Information Information Bioengineering, di Milano, Milano, Via Via Ponzio Ponzio 34/5, 34/5, 20133 20133 Milano, Milano, Italy Italy *Tel: 02and 2399-3559; e-mail: Politecnico [email protected]). e-mail: [email protected]) *Tel: 02** 2399-3559; e-mail: [email protected]). ** e-mail: [email protected]) *Tel: 02** 2399-3559; e-mail: [email protected]). *Tel: 2399-3559; e-mail: e-mail: [email protected]) *Tel: 02 02** 2399-3559; e-mail: [email protected]). [email protected]). e-mail: [email protected]) ** e-mail: e-mail: [email protected]) [email protected]) Abstract: Residential buildings ** represent considerable portion of the energy demand of a temperate Abstract: Residential buildings represent a considerable portion of the energy demand of a temperate Abstract: Residential represent a considerable portion the energy temperate country. Old Europeanbuildings regions, where most of the buildings wereofoften built indemand periods of of alow energy country. Old Europeanbuildings regions, where most of the buildings wereof often built indemand periods of of aalow energy Abstract: Residential represent aa considerable portion the energy temperate Abstract: Residential buildings represent considerable portion of the energy demand of temperate country. Old European regions, where most of the buildings were often built in periods of low energy prices, have a large margin for improvement. The study shows how energy saving measures can be Abstract: Residential buildings represent a considerable portion of the energy demand of a temperate prices, have a large margin for improvement. The study shows how energy saving measures can be country. Old European regions, where most of the buildings were often built in periods of low energy country. Old European regions, where most of the buildings were often built in periods of low energy prices, have a large margin for improvement. The study shows how energy saving measures can be optimally planned at regional level, taking into account the specific features of the building stock, and country. Old European regions, where most of the buildings were often built in periods of low energy optimally planned at regional level, taking into account the specific features of the building stock, and prices, aa large for improvement. The study how energy saving measures can be prices, have margin for improvement. The study shows how energy saving measures can be optimally planned at margin regional level, taking account theshows specific features the building stock, what thehave consequences of an optimal choice into are in economic and environmental terms. prices, a large large margin for improvement. The study shows how energyof saving measures canand be what thehave consequences of an optimal choice into are in economic and environmental terms. optimally planned at regional level, taking account the specific features of the building stock, and optimally planned at regional level, taking into account the specific features of the building stock, and what the consequences of an optimal choice are in economic and environmental terms. optimally planned at regional level, taking into account the specific features of the building stock, and Keywords: Building renovation, Air pollution, Air quality planning and control, Machine learning for © 2018, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved. what the consequences consequences of an an optimal optimal choice are in in economic and environmental environmental terms. Keywords: Building renovation, Air choice pollution, Air quality planning and control, Machine learning for what the of are economic and terms. Keywords: Building renovation, Air pollution, Air quality planning and control, Machine learning for what the consequences of anLarge optimal choice are in economic and environmental terms. environmental applications, scale optimization problems environmental applications, Large scale optimization problems Keywords: Building renovation, Air pollution, Air quality planning and control, Machine learning for Keywords: Building renovation, pollution, Air quality planning environmental applications, LargeAir scale optimization problems Keywords: Building renovation, Air pollution, Air quality planning and and control, control, Machine Machine learning learning for for  environmental applications, Large scale optimization problems environmental  problems environmental applications, applications, Large Large scale scale optimization optimization problems al., 2013; Kurnitski et al., 2011) and many software  al., 2013; Kurnitski et al., 2011) and many software 1. INTRODUCTION  al., 2013;allow Kurnitski et al.,determination 2011) and ofmany software packages an optimal the insulation 1. INTRODUCTION  packages allow an optimal determination ofmany the insulation 1. INTRODUCTION al., 2013; Kurnitski et al., 2011) and software al., 2013; Kurnitski et al., 2011) and many software packages allow an optimal determination of the insulation measures to adopt (see, for instance, the list of about 200 Residential buildings represent, as it is well known, a 1. INTRODUCTION al., 2013;allow et for al.,determination 2011) the andlist many software 1. INTRODUCTION toKurnitski adopt (see, instance, of about 200 Residential buildings represent, as it is well known, a measures packages an optimal of the insulation 1. INTRODUCTION packages allow an optimal determination of the insulation measures to adopt (see, for instance, the list of about 200 software tools on www.buildingenergysoftwaretools.com). Residential buildings represent, as it is well known, a considerable portion of the energy demand of a temperate packages allow an optimal determination of the insulation software tools on www.buildingenergysoftwaretools.com). considerable portion of the energy demand of a temperate to adopt (see, for instance, the list of about 200 Residential buildings represent, asdemand it is is also well known, a measures measures to adopt (see, for instance, the list of about 200 Residential buildings represent, as it well known, a software tools on www.buildingenergysoftwaretools.com). Recently, the problem has been dealt with at city level (De considerable portion of the energy of a temperate country. This has obvious consequences in terms of to (see, instance, the list of about Residential represent, asdemand it is also well Recently, theadopt problem hasforbeen dealt with at city level 200 (De country. Thisbuildings has obvious consequences in terms ofa measures software tools on www.buildingenergysoftwaretools.com). considerable portion of the energy of aa known, temperate software tools on www.buildingenergysoftwaretools.com). considerable portion of the energy demand of temperate Recently, the problem has been dealt with at city level (De Miglio et al., 2017; Yamagata and Seya, 2013; Kostevšek et country. This has obvious consequences also in terms of economic and environmental costs. This is particularly true in software tools on Yamagata www.buildingenergysoftwaretools.com). considerable portion of the consequences energy demand of in a temperate Miglio et al., 2017; and Seya, 2013; Kostevšek et economic and environmental costs. This is particularly true in Recently, the problem has been dealt with at city level (De country. This has obvious also terms of Recently, the problem has been dealt with at city level (De country. This has obvious consequences also in terms of Miglio et al., 2017; Yamagata and Seya, 2013; Kostevšek et al., 2013), but much more rarely at a regional (an economic and environmental costs. This is particularly true in old European countries where most of the buildings were Recently, the problem has been dealt with at city level (De country. This has obvious consequences also in terms of al., 2013), but much more rarely at a regional (an old European countries where most of the buildings were Miglio et al., 2017; Yamagata and Seya, 2013; Kostevšek et economic and environmental costs. This is particularly true in Miglio et al., 2017; Yamagata and Seya, 2013; Kostevšek et economic and environmental costs. This is particularly true in al., 2013), but much more rarely at a regional level (an exception is perhaps, Brandoni and Polonara, 2012). old European countries where most of the buildings were built in periods when the attention to energy saving and Miglio et al., 2017; Yamagata and Seya, 2013; Kostevšek et economic and environmental costs. This isenergy particularly true in exception is perhaps, Brandoni and Polonara, 2012). built in periods when the attention to saving and al., 2013), but much more rarely at a regional level (an old European countries where most of the buildings were al., 2013), but much more rarely at a regional level (an old European countries where most of the buildings were Regional plans must indeed evaluate all the consequences of built in periods when the attention to energy saving and pollutant emissions was much lower than today. exception is perhaps, Brandoni and Polonara, 2012). al., 2013), but much more rarely at a regional level (an old European countries where most of the buildings were Regional plans must indeed evaluate all the consequences of pollutant emissions was much lower than today. exception is perhaps, Brandoni and built in periods when the attention to energy saving and exception is perhaps, Brandoni and Polonara, 2012). built in when the attention to energy saving and plans must indeed evaluate all thePolonara, consequences of alternative options considering both 2012). GHG pollutant emissions was much lower than exception isrenovation perhaps, Brandoni and Polonara, built in periods periods thebudget attention to today. energy saving and Regional alternative renovation options considering both 2012). GHG According to thewhen energy ofthan the Lombardy Region Regional plans must indeed evaluate all the consequences of pollutant emissions was much lower today. Regional plans must indeed evaluate all the consequences of pollutant emissions was much lower than today. and NO alternative renovation options considering both GHG emissions and those of local pollutants, i.e. PM According to the energy budget of the Lombardy Region 10 x, Regional must indeed evaluate all the pollutant 2012), emissions was much lower NOof emissionsplans and those of local pollutants, i.e. consequences PMboth According to theforenergy budget the today. Lombardy Region (Sirena, instance, theofthan energy consumption of and 10 and GHG x, alternative renovation options considering both alternative renovation options considering GHG and NO emissions and those of local pollutants, i.e. PM their impacts on the overall air quality. (Sirena, 2012), for instance, the energy consumption of 10 x, According to the energy budget of the Lombardy Region alternative renovation options considering both GHG According to the energy budget of the Lombardy Region and their impacts on the overall air quality. (Sirena, 2012), for instance, the energy consumption of residential buildings constituted 31% of the overall energy and NO emissions and those of local pollutants, i.e. PM 10 and NOx x,, According to the energy budget of the Lombardy Region emissions and those of local pollutants, i.e. PM and their impacts on the overall air quality. 10 residential buildings constituted 31% of the overall energy (Sirena, 2012), for instance, the energy consumption of NOon emissions and those of local pollutants, i.e. PM 10 and x, (Sirena, 2012), for instance, the energy consumption of This study addresses the last problem in three steps: first, residential buildings constituted 31% of the overall energy used in the region in 2012. Accordingly, this sector was and their impacts on the overall air quality. impacts on the air quality. (Sirena, 2012), for inconstituted instance, the energy consumption of and Thistheir study addresses theoverall last problem in three steps: first, on used in the region 2012. Accordingly, this sector was residential buildings 31% of the overall energy and their impacts on the overall air quality. residential constituted of overall energy Thisbasis study last problem in three first, on ofaddresses a large setthe of detailed building data,steps: the trade-offs used in thebuildings region 2012. this sector was2 the 9%31% of NO and 20% of CO responsible for 40% in of PM10, Accordingly, residential buildings constituted of xxthe the overall energy the basis ofaddresses a large setthe of detailed building data,steps: the trade-offs 9%31% of NO and 20% of CO responsible for 40% in of PM10, Accordingly, 2 This study last problem in three first, on used in the region 2012. this sector was This study addresses the last problem in three steps: first, on used in the region in 2012. Accordingly, this sector was the basis of a large set of detailed building data, the trade-offs between energy savings and their implementation costs , 9% of NO and 20% of CO responsible for 40% of PM equivalent emissions. Quite similar results emerge from the 10 x 2 Thisbasis study addresses the last in three steps: first, are on used in theemissions. region in 2012. Accordingly, this20% sector was energy savings and problem theirbuilding implementation costs are equivalent Quite similar results emerge from the2 between the of a large set of detailed building data, the trade-offs , 9% of NO and of CO responsible for 40% of PM 10 x the basis of a large set of detailed data, the trade-offs , 9% of NO and 20% of CO responsible for 40% of PM between energy savings and their implementation costs are determined by repeatedly solving a linear programming equivalent emissions. Quite similar results emerge from the emission inventory ofofEmilia-Romagna region: CORINAIR 10 x 2 the basis of a large set of detailed building data, the trade-offs , 9% of NO and 20% of CO responsible for 40% PM determined by repeatedly solving a linear programming 10 x 2 emission inventory of Emilia-Romagna region: CORINAIR between energy savings and their implementation costs are equivalent emissions. Quite similar results emerge from the the between energy savings and their implementation costs are equivalent emissions. similar results emerge from determined by repeatedly a linear programming Second, a classical cost-benefit analysis has been emission inventory of Quite Emilia-Romagna region: CORINAIR macro sector 2 (non-industrial combustion) is responsible for problem. between energy savings andsolving their implementation costs are equivalent emissions. Quite similar results emerge from the Second, a classical cost-benefit analysis has been macro sector 2 (non-industrial combustion) is responsible for problem. determined by repeatedly solving aa linear programming emission inventory of Emilia-Romagna region: CORINAIR determined by repeatedly solving linear programming emission inventory of Emilia-Romagna region: CORINAIR problem. Second, a classical cost-benefit analysis has been performed to select the best type and spatial diffusion of macro sector 2 (non-industrial combustion) is responsible for emissions, 8% of NO and 31% of CO 52% of PM 10 x region: CORINAIR 2eq determined by repeatedly solving a linear programming emission inventory of Emilia-Romagna performed to select the best cost-benefit type and spatial diffusion of emissions, 8% combustion) of NOx andis 31% of COfor 52% ofsector PM102 2eq problem. Second, aa classical analysis has been macro (non-industrial responsible problem. Second, classical cost-benefit analysis has been macro sector 2 (non-industrial combustion) is responsible for performed to select the best type and spatial diffusion of energy saving measures in the region. Third, the emissions, 8% of NO and 31% of CO 52% of PM emissions (INEMAR, 2014). 10 x 2eq problem. Second, a classical cost-benefit analysis has been macro sector 2 (non-industrial combustion) is 31% responsible for energy saving measures in the region. Third, the emissions (INEMAR, 2014). performed to select the best type and spatial diffusion of emissions, 8% of NO and of CO 52% of PM 10 x performed to select the best type and spatial diffusion of 8% of NO 52% of PM energy saving measures in the region. Third, the and classical pollutants are correspondent reductions of CO emissions (INEMAR, 2014). 10 emissions, x and 31% of CO2eq 2eq 2 performed to select the best type and spatialpollutants diffusion of 8%domestic of NOxheating and 31% of COare 52% of PM classical are correspondent reductions of CO 10is emissions, 2eq The situation critical since emissions 2 and energy saving measures in the region. Third, the emissions (INEMAR, 2014). energy saving measures in the region. Third, the emissions (INEMAR, 2014). computed and their distribution over the regional territory is and classical pollutants are correspondent reductions of CO The situation is critical since domestic heating emissions are 2 saving measures in the region. Third, the emissions (INEMAR, 2014). computed and their distribution over the regional territory is The situation isincritical since and domestic emissions few are energy mainly located urban areas at lowheating levels (normally and classical pollutants are correspondent reductions of CO 2 and classical pollutants are correspondent reductions of CO computed and their distribution the regional territory is evaluated. The study considers in particular the Lombardy 2over mainly located in urban areas and at low levels (normally few The situation is critical since domestic heating emissions are and classical pollutants are correspondent reductions of CO The situation is critical since domestic heating emissions are 2 evaluated. The study considers in particular the Lombardy mainly located in urban areas and at low levels (normally few tens of meters). Thus they remain in the local atmosphere computed and their distribution over the regional territory is The situation isincritical sinceremain domestic heating emissions are evaluated. computed and their distribution over the regional territory is The study considers in particular the Lombardy region, which is characterized by a high number of residential tens of meters). Thus they in the local atmosphere mainly located urban areas and at low levels (normally few computed and distribution thenumber regional is mainly in urban areas and at low levels (normally few region, which istheir characterized byover a high of territory residential tens of located meters). Thus they remain in and the local atmosphere producing a relevant impact on citizens their health. evaluated. The study considers in particular the Lombardy evaluated. The study considers in particular the Lombardy mainly located in urban areas and at low levels (normally few region, which is characterized by a high number of residential buildings (about 1.5 million) and where exceedances of producing a relevant impact on citizens and their health. tens of meters). Thus they remain in the local atmosphere evaluated. Theis study considers in particular the Lombardy tens of meters). Thus they remain in the local atmosphere buildings (about 1.5 million) and where exceedances of producing a relevant impact on citizens and their health. region, which characterized by a high number of residential region, which is characterized aa high number of tens of meters). Thus they remain in adopted the their localwhen atmosphere (about 1.5 million)by where exceedances of pollution limits frequent. One immediate measure, that iscitizens often critical buildings producing aa relevant impact on and health. region, which isare characterized byand high number of residential residential producing impact on and health. pollution limits are frequent. One immediate measure, that iscitizens often adopted when (about 1.5 million) and where exceedances of producing a relevant relevant impact on andatheir their health.critical buildings (about 1.5 million) and where exceedances of pollution limits are frequent. One immediate measure, that adopted when critical pollution level are reached, is iscitizens tooften impose reduction of the buildings buildings (about 1.5 million) and where exceedances of The paper is organized as follows: the next section revises the pollution level are reached, is to impose a reduction of the pollution limits are frequent. One immediate measure, that is often adopted when critical pollution limits are frequent. One immediate measure, that is often adopted when critical The paper is organized as follows: the next section revises the pollution level are reached, is to impose a reduction of the heating temperature within buildings in order to reduce pollution limits are frequent. One immediate measure, that is often adopted when critical The paper is organized as follows: the next section revises the heating temperature within buildings in order to reduce the situation of the residential building stock in the region and pollution level are reached, is to impose aaadopted reduction of the pollution level are is to reduction of situation residential building in the region and heating temperature within inisorder to reduce consequent emissions. Suchbuildings a measure when The paper paperof is the organized as follows: thestock next section revisesinto the pollution level are reached, reached, is to impose impose aadopted reduction of the the The is organized follows: the next section revises the of the residential building stock in the region and explains the type of as energy saving measures taken consequent emissions. Such a measure is when heating temperature within buildings in order to reduce the situation The paper is organized as follows: the next section revises the heating temperature within buildings in order to reduce explains the type of energy saving measures taken into consequent emissions. Such a measure is adopted when the limits for pollution defined by the current European situation of the residential building stock in the region and heating temperature within buildings in order to reduce the situation of the residential building stock in the region and consideration. Section 3 formulates the cascade of three explains the type of energy saving measures taken into limits for pollution defined by the current European consequent emissions. Such a measure is adopted when the situation of the residential building stock in the region and consequent Such aa measure adopted when consideration. Section 3 formulates the cascadetaken of three limits for emissions. pollution defined by theis regulations are reached. This happens frequently in the explains the type of energy saving measures into explains the type of saving measures taken into consequent emissions. Such measure is current adopted European when Section 3 while formulates of inthree problems outlined above, resultsthe arecascade presented the regulations are reached. This happens frequently in the consideration. limits for pollution defined by the current European explains the type of energy energy saving measures taken into limits for pollution defined by the current European problems outlined above, while results are presented in the regulations are reached. This happens frequently in the coldest winter period when the air over the Padana plain is consideration. Section 3 formulates the cascade of three consideration. Section 3 formulates the cascade of three limits for pollution defined by the current European problems outlined above, while results are presented in following section. Limits and possible extensions of the coldest winter period when the air over the Padana plain is regulations areperiod reached. This happens frequently in and the consideration. Section 3 while formulates the cascade ofof three regulations are reached. This happens frequently in the following section. Limits and possible extensions the coldest winter when the air over the Padana plain is particularly stable. Other measures are however possible problems outlined above, results are presented in the regulations areperiod reached. This happens frequently in and the problems outlined above, while results presented in following section. Limits and possible extensions procedure are discussed in the concluding section. particularly stable. Other measures are however possible coldest winter when the air over the Padana plain is problems outlined above, while results are are presented of in the the coldest winter period when the air over the Padana plain is procedure are discussed in the concluding section. particularly stable. Other measures are however possible and consist in renovating the particularly old building stocks following section. Limits and possible extensions of the coldest winter period when the air are overhowever the building Padana plain is procedure following section. Limits and possible extensions of the are discussed in the concluding section. consist in renovating themeasures particularly old stocks particularly stable. Other possible and following section. Limits and possible extensions of the particularly stable. Other measures are however possible and consist in renovating the particularly old building stocks introducing energy efficiency measures. In this case, the procedure are are discussed discussed in in the the concluding concluding section. section. particularly stable. measures are however possible and introducing energyOther efficiency measures. Inbuilding this case, the procedure consist the old consist in renovating the particularly old building stocks introducing efficiency In this case, the procedure are discussed in the concluding section. problemin is renovating toenergy determine theparticularly bestmeasures. type and diffusion ofstocks such consist in renovating the particularly old building stocks 2. DETAILED DATA ON BUILDINGS problem is toenergy determine the bestmeasures. type and In diffusion of such introducing efficiency this case, the introducing efficiency this the 2. DETAILED DATA ON BUILDINGS problem toenergy determine bestmeasures. type diffusion ofmany such reductionismeasures. The the problem has and beenIn addressed introducing energy efficiency measures. In this case, case, the 2. DETAILED DATA ON BUILDINGS reduction measures. The problem has been addressed many problem is to determine the best type and diffusion of such problem is to determine the best type and diffusion of such reduction measures. The problem has been addressed many The Directive 2012/27/EUDATA of the ON European Parliament and of times in the literature with reference to the individual 2. DETAILED BUILDINGS problem is to determine the best type and diffusion of such 2. BUILDINGS The Directive 2012/27/EUDATA of the ON European Parliament and of times in the literature with reference to the individual reduction measures. The problem has been addressed many 2. DETAILED DETAILED DATA ON BUILDINGS reduction measures. The problem has been addressed many The Directive 2012/27/EU of the European Parliament and of times in the literature with reference to the individual the Council on energy efficiency foresees that “Member building (e.g. Machairas et al., 2014; Evins, 2013; Hamdy et reduction measures. The problem has been addressed many the Council on energy efficiency foresees that “Member building (e.g. Machairas et al., 2014; Evins, 2013; Hamdy et The Directive 2012/27/EU of the European Parliament and of times in the literature with reference to the individual The Directive 2012/27/EU of the European Parliament and times in the literature with reference to the individual the Council on energy efficiency foresees that “Member building (e.g. Machairas et al., 2014; Evins, 2013; Hamdy et The Directiveon 2012/27/EU of the European Parliament and of of times in(e.g. the Machairas literature etwith reference to 2013; the individual the Council energy efficiency foresees that “Member building al., 2014; Evins, Hamdy et the Council on energy efficiency foresees that “Member building (e.g. Machairas et al., 2014; Evins, 2013; Hamdy et building (e.g. Machairas et al., 2014; Evins, 2013; Hamdy et the Council on energy efficiency foresees that “Member

Copyright © 2018 102 Hosting by Elsevier Ltd. All rights reserved. 2405-8963 © 2018, IFAC IFAC (International Federation of Automatic Control) Copyright © 2018 IFAC 102 Peer review©under responsibility of International Federation of Automatic 102Control. Copyright 2018 IFAC 10.1016/j.ifacol.2018.06.218 Copyright © 2018 IFAC 102 Copyright © 2018 IFAC 102 Copyright © 2018 IFAC 102

IFAC IAMES 2018 Brescia, Italy, May 10-11, 2018

Giorgio Guariso et al. / IFAC PapersOnLine 51-5 (2018) 102–107

103

States shall establish a long-term strategy for mobilizing investment in the renovation of the national stock of residential and commercial buildings, both public and private. This strategy shall encompass: (a) an overview of the national building stock based, as appropriate, on statistical sampling; (b) the identification of cost-effective approaches to renovations relevant to the building type and climatic zone; (e) an evidence-based estimate of expected energy savings and wider benefits.” To comply with this and preceding directives, Lombardy region has approved a series of regulations (the latest of which is the D.g.r. 17 July 2015 - n. X/3868 “Provisions on the energy efficiency discipline buildings and the relative certificate of energy performance”) to check and improve the energy situation of buildings. Among these regulations is the creation of an open catalogue of energy performance certificates (APE, in the Italian acronym) where all the characteristics of a building from the energy point of view are stored, as assessed by a certified technician. The interest of such a database, named CENED, is that it also includes the suggested energy saving actions with an estimate of their cost and foreseen benefit in terms of reduced energy use.

Fig. 1. Age distribution of the sample of the building stock in Lombardy. Indeed, as shown in figure 2, the median of the energy consumption in residential buildings is around 180 kWh/m2 per year, with 26% using between 180 and 270 kWh/m 2 per year and 24% exceeding the last value. This is a clear indication of how inefficient the energy use is. In fact, more than half (about 51%) of the houses in the sample belongs to the G class (the least efficient, which corresponds to the use of 16 m3 of methane per square meter per year). A renovation of the stock is thus not only possible, but also highly desirable for the strong improvement it may entail in both the use of energy and the air pollution.

More precisely, for the purpose of this study, three possible actions have been considered: -

A thermal insulation of the opaque building envelope

-

A change of windows material which reduces the heat loss from the home

-

A full restructuring of the building including the two preceding actions.

The database presently includes a complete energy analysis of more than 400,000 dwellings and non-residential building thus representing a large sample of the regional building stock. More than 150,000 records represent residential houses. They were thus considered as a sufficient numerical base for the current study. Fig. 2. Distribution of the energy consumption in the sample.

A simple summary of these data is sufficient to provide a general picture of the current regional situation. For instance, the number of buildings with only one dwelling (single family houses) represent the 42% of the total, which perfectly agrees with the general data of the National Statistical Institute (ISTAT, 2011), 16% of the sample refers to buildings with three to eight flats while 34% are houses with more than eight flats.

2.1 A subdivision of the building stock To better manage the overall planning process, the building stock has been subdivided into a number of classes (or archetypes) as already done in various European projects such as EPISCOPE (Stein et al. (eds.), 2016) or InSMART (Gargiulo et al., 2017) and corresponding to some of the characteristics assessed by the National Statistical Institute. These characteristics are:

Figure 1 shows the distribution of building ages of the sample: 11% of the stock dates back to years prior to the ‘30s and only 4% is after 2006, i.e. it was built under the most recent energy saving regulations. The ‘60s were the period of more intense construction activity and coincides with the fastest economic development. The pressure for cheap new houses induced a rush to new constructions not paralleled with an attention to their environmental impact. Buildings from that period still constitute the 31% of the total stock.

103

-

Construction year: the 7 different classes shown in figure 1 were used;

-

Number of dwellings: 4 classes: 1, 2, 3-8, more than 8;

-

Sub-region (or Province): grouped into 6 classes to take into account different construction traditions;

IFAC IAMES 2018 Brescia, Italy, May 10-11, 2018 104

-

Giorgio Guariso et al. / IFAC PapersOnLine 51-5 (2018) 102–107

Elevation: 2 classes, below and above 600 m a.s.l. since the construction characteristics in the mountains have always been quite peculiar.

In the end, a total of 224 classes of buildings is generated, since some of the sub-regions only have one elevation level. This means that each class is represented in the sample by an average of about 700 instances, even if of course there are case for which the number of building reduces to few tens (typically, in the post 2006 class). Each class is also characterized by the cost and possible advantages of implementing an improved insulation of walls and/or windows. More precisely, the investment cost, the reduction in energy use, and the consequent reduction of GHGs emission can be computed from the available data. The total number of buildings of each class actually present in the territory is derived from the national statistics for each municipality. The municipality in turn belongs to a given province and has an average elevations; this allows the extrapolation of the values obtained from the CENED database to the global building stock of the region. Fig. 3. Scattergrams of the training, cross validation and test results for the annual savings (k€) corresponding to full restructuring.

2.2 Determining the joint effect of two actions Unfortunately, the consequences of adopting both the actions on the walls and on the windows are not always reported in the database. The full restructuring activity is obviously more efficient than the two separate actions on walls and windows, but is normally slightly less efficient than the sum of the two. This is confirmed by all the detail studies catalogued for instance in the TABULA project.

We will assume, as normally done in air quality plans (see Guariso and Volta, 2017), that the decision variables are the number nij of buildings in each class i that adopt the action j. These values are sometimes referred to as “application rates” of the improvement action (Guariso et al., 2016). The objectives of the problem to be optimized are the global investment cost and the energy reduction. The first must be minimized and the second maximized. We can thus write:

In order to evaluate the joint effects of both the walls and the windows actions, instead of developing 224 separate functions, we developed three feed-forward artificial neural networks (ANNs). Given the categorical inputs defining the class of the building and the separate effects of the walls and the window actions, they compute the energy savings, the investment cost, and the annual savings respectively (see for instance, Nguyen et al., 2014).

Where cij is the investments cost and rij the energy reduction deriving from the adoption of action j on a building of class i. Such a problem is subject to the following constraints.

Such an approach, that closely follows that proposed by Magnier and Haghighat (2010), gave excellent results as shown by the scatterplots in figure 3. The identification set is composed of 19,800 samples. 70% of the cases were used to train the ANNs, 15% composes the cross validation set, and the remaining 15% represents the test set. The correlation between actual values and reconstructed ones was always above 0.98. Considering energy saving and investment cost, the result obtained are similar to those shown in figure 3 for the money saving.

Meaning that, in each class i, each building may undergo only either action 1 on the walls, or action 2 on the windows, or action 3, i.e. full restructuring.

Where Kj is the allowable number of actions type j that can be adopted. Not all the buildings can in fact adopt the same measure. Kj is estimated on the base of the fraction of buildings in the CENED database for which action j was recommended. Finally, the traditional non negativity constraints of the decision variables hold.

3. OPTIMAL PLANNING PROBLEM Given the information described above, we formulate here the optimization problem that shows the trade-off between investment costs and energy reduction at the regional level.

The Pareto front of this problem is shown in figure 4 and presents an almost linear shape, meaning that the reduction in energy consumption is more or less proportional to the investment in efficiency actions. 104

IFAC IAMES 2018 Brescia, Italy, May 10-11, 2018

Giorgio Guariso et al. / IFAC PapersOnLine 51-5 (2018) 102–107

105

600 m of elevation, despite these houses represent only 11% of the total stock.

Fig. 4. The trade-off between investment cost and energy reduction. Fig. 5. The net benefit curve.

Perhaps the most important result of this analysis is the value of the upper right point of the Pareto front. It shows that the maximum reduction that can be attained is of the order of 38% of the current use of energy, and this corresponds to a total investment of over 46 billion euros. To put this value in perspective, the GDP of the region is about 360 billion euros.

It is interesting to note that action 2 (i.e. windows replacement) is never taken into account. It is recommended only few times (5%) in combination with wall insulation, and this produces a decreasing in the derivative of the Pareto front represented in figure 4. These facts prove that windows replacement has a low benefit-investment ratio.

3.2 Cost-benefit analysis

3.3 Environmental impacts

In order to select a meaningful value within the Pareto front, several different approaches are possible. A traditional one is cost-benefit analysis, which is very close to the point of view of the individual citizens (who should pay for the investment in energy saving actions). Direct benefits are the annual reduction in the energy bills, consequent to the reduced use of fuels. These reductions are always positive since the CENED database reports the time needed to recover the initial investment and this averages around 12 years for action 1, and 20 years for action 2. They occur in time and thus, to be fairly compared with the cost, must by actualized with the traditional formula of the net present value (Kurnitski et al., 2011). To adopt it, one has to fix the temporal horizon and define the discount rate. For the current study, we adopted 15 years and 4%, i.e. something close to real estate loans. An extensive sensitivity analysis around this value has not shown significant variations of the results.

The set of decisions corresponding to the maximum difference between benefits and costs, can also be evaluated under the perspective of environmental impacts. These are of two different types: greenhouse effect, represented by the reduction of CO2eq emissions, and health related, that can be measured by the reduction of emission of traditional pollutants such as PM10 and NOx. As to GHGs emissions, their calculation is straightforward, since they are directly related to the reduction of energy use. The regional emission inventory (INEMAR, 2014) estimates in 10.7 Mt CO2eq per year the emission from residential buildings. The energy reduction corresponding to the economically most convenient choice means a decrease of 2.7 Mt per year or about 4% of the overall GHGs emission of the region (domestic heating being just 15% of regional CO 2 emissions).

We can thus formulate the cost benefit problem as:

Local pollution can be evaluated in the same way. Given that the current emission estimate for PM10 and NOx are respectively 7170 t and 8073 t per year, the adoption of the measures outlined above would entail a reduction of about 1700 and 1750 t per year for PM10 and NOx. This is particularly significant for PM10, since it would represent almost 10% of the total regional emission. However, these values are not sufficient to understand the complete consequences of the plan. It is in fact important also to determine where these reductions are located and thus what is their impact on the overall air quality of the region.

Where the decision variable is the global energy reduction R, and B(R) and C(R) are the correspondent total actualized benefit and total investment costs. This net benefit function is reported in figure 5 and shows that the optimal choice is to select not a full adoption of the energy saving measures, but something which is approximatively around two third of the maximum. This situation corresponds to the adoption of action 1 (i.e. wall insulation) on about 60% of the buildings and of action 3 (combination of action 1 and 2) on about 5%. It would require an investment of 24.6 G€, saving about 36 TWh per year. Action 1 should be implemented on about 80% of the houses built between 1930 and 1976; 58% of those built between 1977 and 1992; 33% of those built after 1992 and before 2006, and only 8% of those built later. About one fourth of action 1 should take place in municipalities above

For this purpose, we have to map the set of decisions computed above on the building stock of each municipality, and compute in this way the reduction with respect to the emission in the regional inventory. Figures 6 and 7 show the result of this operation for PM10 and NOx, respectively. It clearly emerges that emission reduction is not uniform over the territory. For instance, the relatively small province of 105

IFAC IAMES 2018 106 Brescia, Italy, May 10-11, 2018

Giorgio Guariso et al. / IFAC PapersOnLine 51-5 (2018) 102–107

Milan accounts for 17.5% of the reduction of NO x since it is densely built and served by a capillary methane network. On the contrary, the provinces of Bergamo and Brescia, where the tradition of burning wood for domestic heating is more diffused account respectively for 17 and 21% of the PM 10 reduction.

renewable source (specifically, photovoltaic and thermal panels) may play an important role in the overall energy system. However, dealing with this type of changes requires a number of additional assumptions that may take the study outside the current scope. On the contrary, the reductions computed here, since they refer only to the passive behaviour of the building envelope, are somehow granted. Though depending in part from the specific meteorological conditions of each year, they should certainly be undertaken.

Fig. 6. Municipal emissions of PM10 before (top) and after (bottom) the reduction. Fig. 7. Municipal emissions of NOx before (top) and after (bottom) the reduction.

4. CONCLUSIONS

One of the reasons why they are not being applied so widely is the large initial investment required. The current Italian legislation provides a subsidy to energy saving measures as a 65% reimbursement of the investment in terms of reduced taxes over a ten year period. However, since the investment is of the order of tens of thousands of euros per flat or house, in an economic situation like the current one, the availability of the necessary capital still constitutes a relevant barrier. New forms of bank loans could be possible to overcome this problem and may constitute a win-win solution since they can be effective for both the economic and the environmental points of view.

The problem dealt with in this study could, in principle, be formulated as a complete benefit-cost analysis. The economic value of the avoided GHGs emissions (for instance, by referring to its value on the emission market) and the decrease in human health problems due to reduced pollution concentration can also be considered. However, it should be noted that these indirect economic effects are generally lower than the direct savings due to the reduced use of fuels (see for instance, Chae and Park, 2011). The possible energy saving actions considered in this study are not the only contribution that the building sector can provide to a complete energy and environment plan. A change of the heating system as well as an increased use of 106

IFAC IAMES 2018 Brescia, Italy, May 10-11, 2018

Giorgio Guariso et al. / IFAC PapersOnLine 51-5 (2018) 102–107

REFERENCES

107

Regione Lombardia (2015) Piano Regionale degli Interventi per la qualità dell’Aria (PRIA), (in Italian). Avalable at: http://www.regione.lombardia.it. Last accessed Jan. 2018. Sirena (2012). Sistema Informativo Regionale ENergia Ambiente, (in Italian). Avalable at: http://sirena20.energialombardia.eu. Last accessed Jan. 2018. Stein, B., Loga, T., and Diefenbach, N.(eds.) (2016). Monitor Progress Towards Climate Targets in European Housing Stocks - Main Results of the EPISCOPE Project, Institut Wohnen und Umwelt GmbH, Darmstadt, Germany. Yamagata, Y., Seya, H. (2013). Simulating a future smart city: An integrated land use-energy model. Applied Energy, 112, 1466-1474.

Brandoni, C., Polonara, F. (2012). The role of municipal energy planning in the regional energy-planning process, Energy, 48, 323-338. Chae, Y., Park, J. (2011). Quantifying costs and benefits of integrated environmental strategies of air quality management and greenhouse gas reduction in the Seoul Metropolitan Area, Energy Policy, 39, 5296-5308. De Miglio, R., Chiodi, A., Burioli, S., Giovannini, E., and Gargiulo, M. (2017). Exploring integrated energy action plans for a sustainable transition of the municipality of Cesena, Procedia Engineering, 198, 541-548. Evins, R. (2013). A review of computational optimisation methods applied to sustainable building design, Renewable and Sustainable Energy Reviews, 22, 230245. Gargiulo, M., Chiodi, A, De Miglio, R., Simoes, S., Long, G., Pollard, M., Gouveia, J.P., and Giannakidis, G. (2017). An integrated planning framework for the development of sustainable and resilient cities – the case of the InSMART project, Procedia Engineering, 198, 444-453. Guariso, G., Maione, M., Volta, M. (2016).A decision framework for Integrated Assessment Modelling of air quality at regional and local scale, Environmental Science & Policy, 65, 3-12. Guariso, G. and Volta, M. (2017). Air Quality Integrated Assessment, A European Perspectives. Springer International Publishing. Hamdy, M., Hasan, A., and Siren, K. (2013). A multi-stage optimization method for cost-optimal and nearly-zeroenergy building solutions in line with the EPBD-recast 2010, Energy and Buildings, 56, 189-203. INEMAR (2014). INventario EMissioni Aria, (in Italian). Avalable at: http://www.inemar.eu. Last accessed Jan. 2018. ISTAT (2011). Censimento Popolazione e Abitazioni 2011 (in Italian). Available on http://dati.istat.it. Last accessed Jan. 2018. Kostevšek, A., Petek, J., Čuček, L., and Pivec, A. (2013). Conceptual design of a municipal energy and environmental system as an efficient basis for advanced energy planning, Energy, 60, 148-158. Kurnitski, J., Saari, A., Kalamees, T., Vuolle, M., Niemelä, J., and Tarke, T. (2011). Cost optimal and nearly zero (nZEB) energy performance calculations for residential buildings with REHVA definition for nZEB national implementation, Energy and Buildings, 43, 3279-3288. Machairas, V., Tsangrassoulis, A., and Axarli, K. (2014). Algorithms for optimization of building design: A review, Renewable and Sustainable Energy Reviews, 31, March, 101-112. Magnier, L., and Haghighat, F. (2010). Multiobjective optimization of building design using TRNSYS simulations, genetic algorithm, and Artificial Neural Network, Building and Environment, 45, 739-746. Nguyen, A-T., Reiter, S., and Rigo, P. (2104). A review on simulation-based optimization methods applied to building performance analysis, Applied Energy, 113, 1043-1058.

107