Photovoltaic/Thermal (BIPV/T) System Retrofits in the Canadian Housing Stock

Abstract Techno-economic impact of retrofitting houses in the Canadian housing stock with PV and BIPV/T systems is evaluated using the Canadian Hybrid End-use Energy and Emission Model. Houses with south, south-east and south-west facing roofs are considered eligible for the retrofit since solar irradiation is maximum on south facing surfaces in the northern hemisphere. The PV system is used to produce electricity and supply the electrical demand of the house, with the excess electricity sold to the grid in a net-metering arrangement. The BIPV/T system produces electricity as well as thermal energy to supply the electrical as well as the thermal demands for space and domestic hot water heating. The PV system consists of PV panels installed on the available roof surface while the BIPV/T system adds a heat pump, thermal storage tank, auxiliary heater, domestic hot water heating equipment and hydronic heat delivery system, and replaces the existing heating system in eligible houses. The study predicts the energy savings, GHG emission reductions and tolerable capital costs for regions across Canada. Results indicate that the PV system retrofit yields 3% energy savings and 5% GHG emission reduction, while the BIPV/T system yields 18% energy savings and 17% GHG emission reduction in the Canadian housing stock. While the annual electricity use slightly increases, the fossil fuel use of the eligible houses substantially decreases due to BIPV/T system retrofit.


Methodology 211
This study assesses the performance of PV and BIPV/T systems in existing Canadian houses using 212 the Canadian Hybrid Residential End-Use Energy and GHG Emissions Model (CHREM). 213 CHREM is statistically representative of the Canadian housing stock (CHS) with close to 17,000 214 unique house models [27][28][29]. The simulation engine of CHREM is the thoroughly validated, high-215 resolution building energy simulation software 31]. The appliances and lighting (AL) 216 and DHW energy consumption of the houses in CHREM are estimated using a neural network 217 model of Canadian households [32] and a set of AL and DHW load profiles representing the usage 218 profiles in Canadian households. The reductions in GHG emissions due to reduced on-site fossil 219 fuel consumption and reduced off-site fossil fuel consumption for electricity generation are 220 separately calculated for each household. The validity of the predictions of CHREM was verified 221 earlier [27,28]. 222 The energy savings and associated GHG emissions reductions due to retrofitting PV and BIPV/T 223 systems were estimated using CHREM as follows: 224 (i) To maximize the incident solar energy for both PV and BIPV/T retrofits, houses that have 225 a major roof surface facing south-east, south or south-west were selected as eligible houses. 226 For the BIPV/T retrofit, the presence of a basement or mechanical room (plant room) is an 227 additional requirement for eligible houses to provide a suitable space for the equipment 228 required. Since the presence of a mechanical room is not explicitly identified in the 229 CHREM, houses with heating, ventilation and air conditioning systems that require a 230 mechanical room were selected as eligible houses. Due to the differences in construction 231 characteristics of houses across Canada and the non-uniform population density, the 232 number of eligible houses varies substantially from province to province. Also, due to 233 additional basement or mechanical room requirement for the BIPV/T system, the number 234 of eligible houses for the PV retrofit is larger compared to the number of eligible houses 235 for BIPV/T system retrofit. 236 (ii) The ESP-r input files of the selected eligible houses were modified to add the PV and 237 BIPV/T retrofits, and the post-processing code was revised. 238 (iii) The energy savings and reduction (or increase) of GHG emissions of the CHS with the PV 239 and BIPV/T retrofits were evaluated by comparing the estimated energy consumption and 240 GHG emissions with the "base case" (i.e. current) values. The variation in GHG emissions 241 was estimated using published marginal GHG emission intensity factors [33]. The energy 242 savings and reduction (or increase) of GHG emissions estimated by CHREM were 243 extrapolated to the entire CHS using scaling factors [27,28]. 244

Numerical model 245
In this study, first a PV system model is incorporated into CHREM. A roof-mounted PV module 246 can be modeled in ESP-r as part of the roof construction, or as a separate zone with a small 247 thickness, attached to the roof. In the former strategy, the air gap between the PV and the roof is 248 modeled as one layer of the multi-layer construction. In the latter approach the air gap between top 249 and bottom layers is modeled by an air-flow network. The first approach is simpler and less 250 accurate than the second because of the higher temperatures predicted in the air gap due to 251 neglecting the heat transfer rate increase as a result of the air-flow in the gap. Also, it is less suitable 252 for the modelling of BIPV/T systems because the air flow in the gap between the PV module and 253 the roof is a critical component of the system that delivers thermal energy. Therefore, PV panels 254 are modelled in this work as a separate zone with a narrow air gap above the roof, and the air gap 255 is modelled by an airflow network. [34]. The BIPV/T system is modeled using three interconnected 256 networks (i.e. air flow network, plant network and electric network), as shown in Figure 3 [35]. 257

PV array 258
The current vs. voltage curve (I-V curve) is generally used to characterize a PV system. Power 259 generation of a PV cell can be determined from its operating voltage and current. Mottillo et al. 260 [36] used the special materials approach developed by Evans and Kelly [37] and incorporated the 261 PV model into ESP-r based on an equivalent one-diode circuit model (WATSUN-PV model) 262 recommended by Thevenard [38]. The equivalent one-diode circuit is shown in Figure 1. The 263 circuit output current, I, is the difference between the light generated current, IL, and diode current, 264 ID. The diode current represents the resistance of the cell's junction to current flow [36]. This 265 model is based on the short circuit current, the open circuit voltage and maximum power point the 266 at the reference conditions. The reference curve is adjusted to match the operating conditions. The 267 equations defining the short circuit current, Isc, and open circuit voltage, Voc, in the WATSUN-PV 268 model are given in Equations 1 and 2. 269 where ET,eff, is the effective irradiance incident on the surface (W/m 2 ), Tcell, is the cell temperature 270 ( ○ C), and α, β, γ are empirical coefficients. Beam and diffuse solar radiation (inclusive the 271 reflectance of the front surface of the module) are considered in determining the effective 272 irradiance. Reference irradiance and cell temperature are considered to be 1000W/m 2 and 25 ○ C, 273 respectively [36], and the values of the coefficients are given in Table 1 The PV module surface is represented as a multi-layered construction of several material layers. 278 Individual layers are modeled with one or more nodes. A node within the surface is defined as a 279 special material and represents the location of the PV cells within the structure. Cell temperature 280 is determined by solving the energy balance equation for the special material node. 281 The power conditioning unit (PCU) model [40] where Pin, is the power input (W), Pnom, is the nominal power (W), P0, is the power loss when 284 there is a voltage across inverter (W), Us, is the set-point voltage (V), Uout, is the voltage output 285 (V), Ri, is the internal resistance of inverter (Ω) and The Crystalline-Silicone type PV modules with ethylene vinyl acetate (EVA) encapsulation, low-287 iron glass cover and metal back sheet are modelled in CHREM due to their higher efficiency and 288 commercial availability [34]. 289 The number of PV panels for a given roof area is determined as: 290 where NPV, is the number of PV panels (integer), Href, is the Reference insolation (W/m 2 ), η, is the 291 user defined efficiency, AR, is the roof area (m 2 ) and P indv , is the nominal power of individual module (W).

292
The input data used in modelling the PV modules are given in Table 1 [34]. 293

Heating system 294
The preheated outside air exiting the roof integrated PV system is fed into the evaporator of the 295 heat pump as shown in Figure 2. The heat pump is modeled as a grey-box component. Under the 296 grey-box modeling strategy the system behaviour is expressed by performance related equations 297 including the power consumption and COP of the heat pump. This strategy has been used for 298 modeling several plant components in 42]. The empirical expressions used for the COP 299 of the heat pump are given in Equation 6 300 where Tw, is the water temperature and Tsup, is the air supply temperature. The heat pump operation 301 is limited to the supply air temperature above -15 ○ C. At temperatures below -15 ○ C, the heat pump 302 is turned off and the auxiliary boiler supplies heat. The values of the coefficients are determined 303 using manufacture's data given in Table 1 [43]. 304 Hot water leaving the condenser of the heat pump is stored in the hot water tank, which is used for 305 energy storage and heat transfer within the hydronic system. The space and DHW heating loops 306 are served by two heat exchangers immersed in the thermal storage tank, modeled as a stratified 307 tank [44]. When the heat pump energy supply is not enough to satisfy the demand, the auxiliary 308 boiler shown in Figure 2 supplies the shortfall. The auxiliary boiler is modeled as a condensing 309 boiler in locations where natural gas is available, and as a non-condensing boiler where natural gas 310 is not available and oil is used. The boiler performance data given in Table 1 where ηb is the boiler efficiency, η0 is the full load boiler efficiency at the reference temperature, 315 φ is the slope of the efficiency curve, Tref is the reference temperature, and Tret is the return water 316 temperature. Then, the heat supply of the boiler is calculated using the boiler efficiency, fuel 317 heating value (HHV for NG and LHV for oil), and instantaneous fuel use, and added to the energy 318 balance equation as a source term. Boiler heat supply, water flow rate, and heat loss to the 319 environment are used to determine the boiler output temperature. To model the transient operation, 320 the thermal mass of the boiler is included in the energy balance. 321 A hydronic heat delivery system using commercially available radiators [48] is assumed for space 322 heating. The number of radiators in each zone is determined to satisfy the design heating load. The 323 power for the hot water circulation pump is estimated using empirical equations as shown in 324 Equation 8  where Pel,pump.SH and Pel,pump.DHW are the power of the pump operating in the space and DHW 326 heating circuit, respectively. The Pnom,burner is the nominal capacity of the auxiliary boiler. 327

Control strategy 328
The BIPV/T system control sensors and actuators are described in Table 2. The system shown in 329 Figure 2 is controlled by the space and DHW heating demand. The hot water tank stores heat and 330 transfers it to the space and DHW heating loops. The temperature of the hot water tank is 331 maintained between 50 ○ C and 55 ○ C by controlling the hot water pump operation. When the 332 temperature drops below 50 ○ C, the pump turns on and continues to operate until the temperature 333 reaches 55 ○ C. 334 When the outlet temperature from the heat pump evaporator is above the cut-out temperature of 335 the heat pump (-15 ○ C), the heat pump compressor turns on to extract heat from the supply air and 336 heats the water in the hot water tank. If the temperature of the water leaving the heat pump is below 337 50 ○ C, the auxiliary boiler turns on to increase the water temperature to 55 ○ C. With this control 338 scheme, the auxiliary boiler operates only at the BIPV/T system shortfall. 339 If the main zone temperature drops below the thermostat set-point of 20 ○ C, the pump supplies hot 340 water from the hot water tank to the radiators. Since the other zones including the basement are 341 slave to the main zone, hot water is supplied to all radiators in the house, until the main zone 342 temperature exceeds the upper temperature threshold of 22 ○ C. 343 The DHW supply temperature is maintained in the range of 55±1 ○ C. To simplify the model, the 344 operation of the combination of DHW service valves is simulated as a small fully mixed adiabatic 345 tank held at 55±1 ○ C, and the DHW draw and equivalent main water supply are applied into this 346 tank, emulating the operation of the valves in a real system. 347

GHG emission estimation 348
The GHG emissions of carbon dioxide (CO2), methane (CH4) and nitrous oxide (N2O) are 349 converted to and reported as "equivalent CO2" (CO2e) based on their global warming potential 350 calculated according to Equation 10 [28,33]. 351 The amount of CO2e emission for onsite fossil fuel and electricity consumption is determined using 352 the applicable GHG emission intensity factors (GHG EIF). The GHG EIF is the level of CO2e 353 emission per unit energy of fuel. The GHG EIF for onsite NG and oil consumption is defined based 354 on the chemical reactions that occur in the combustion of these fuels in residential boilers. Thus, 355 GHG emissions due to onsite fuel consumption is calculated in each simulation time step based on 356 the fuel type and efficiency of the energy conversion device [28]. Wood combustion is considered 357 carbon neutral because the combustion of wood returns to the atmosphere the CO2 that was 358 recently removed by photosynthesis [33]. 359 Electricity generation in Canada is by provincial utilities, and based on the available primary 360 energy sources, each provincial utility uses a different fuel mixture. Furthermore, the efficiency of 361 energy conversion as well as the transmission and distribution losses are also widely different. 362 Thus, CHREM calculates the GHG emissions associated with electricity consumption separately 363 for each province using provincial GHG EIFs. The GHG EIF for electricity generation is defined 364 as the level of CO2e emissions for the generation and delivery of one kWh electricity to the end-365 user. Since different types of technologies are used for base-load and peak electricity generation, 366 published values of provincial average and marginal GHG EIFs given in Table 3 [33] are used. 367

Connection of the PV system to the electrical grid 368
Energy storage is an essential part of renewable energy systems. While the hot water tank is a 369 practical option for thermal energy storage for dwellings, electricity storage is a more complicated 370 issue. Onsite electricity storage can be managed using batteries at households. However, the space 371 requirements, initial investment and additional maintenance may decrease the favourability of this 372 option. Grid connected low energy buildings can be considered as an alternative. As shown in 373 Figure 2, onsite electricity generation is consumed by the heat pump, fan, pumps and AL operation. 374 If the onsite electricity generation is not sufficient to meet the demand, the required electricity is 375 imported from the grid with the meter recording electricity draw from the grid. When the onsite 376 electricity generation exceeds the electricity demand of the household, the surplus electricity can 377 be exported to the grid. In this case the meter spins backwards and subtracts the value of the 378 exported electricity. This billing strategy is known as net metering and allows residential 379 customers to earn credit for onsite electricity generation. With this strategy, the grid acts as an 380 infinite and lossless electricity storage system for individual houses. The specific policies for net 381 metering is defined by local authorities. Since several parameters can affect the energy market and 382 electricity trade in each jurisdiction, balanced net metering approach is assumed in this study. 383 Under the balanced net metering approach, the onsite electricity generation and grid electricity 384 supply have the same price, GHG EIF and source energy intensity. Net metering is approved by 385 utility companies across Canada for micro scale electricity generation [49][50][51][52][53]. Whether the 386 electricity grid could support the large electricity export due to widespread PV system adoption is 387 a question that must be investigated in future research. 388

Economic analysis 389
There are substantial uncertainties in estimating the investment cost for PV and BIPV/T system 390 retrofits in Canada. The cost of purchase, delivery and installation of system components (e.g. 391 piping and pumps, storage tanks, auxiliary heaters, heat pumps) vary substantially from province 392 to province due to widely different economic parameters including market size, population density, 393 geographical area, competitive market conditions, special site requirements and prevailing labor 394 rates. Furthermore, the price of PV panels has been dropping significantly during the past decade, 395 and further reductions are expected in the near future [54]. Additionally, while it is expected that 396 a solar system retrofit would increase the market value of a house, helping to recover part of the 397 investment cost, the magnitude of the increase is uncertain as it is affected by factors such as buyer 398 perception and sophistication, market forces, and energy prices. 399 Due to these uncertainties, it is not realistic or practical to use a conventional economic feasibility 400 analysis to assess the economic feasibility of PV and BIPV/T retrofits. Therefore, as in other 401 similar studies [16,18,19,21,[24][25][26], a reverse payback analysis method, referred to as "tolerable 402 capital cost" (TCC) of the upgrades [55], was used here. TCC is the acceptable initial investment 403 (including the present value of the additional annual maintenance cost over the system's lifetime) 404 for an energy saving upgrade that will be recovered based on the annual cost savings, the number 405 of desired years for payback, and the estimated annual interest of borrowing money and fuel cost 406 escalation rates. 407 The prices of natural gas, heating oil, electricity and wood used for each province are presented in 408   Table 6. The validity of these estimates was verified earlier 429 [28] by comparing them with available statistical data on Canadian energy use. In this study, first 430 the impact of PV retrofit on energy consumption and GHG emissions is investigated. For this 431 purpose, all houses that are eligible for PV retrofit are assumed to receive this retrofit and the 432 energy consumption and GHG emissions are estimated using CHREM. The same approach is used 433 to evaluate the BIPV/T system retrofit for the CHS. 434

Energy savings 435
The amount of electricity generated by the PV systems retrofitted in all eligible houses (about 35% 436 of the houses in the CHS) and the associated GHG reductions are presented in Table 7. Electricity 437 is used by appliances and lighting (AL), and in some houses additionally for space and DHW 438 heating. Unlike modern, low-energy houses designed to enhance the suitable area for PV panel 439 installation, existing houses were not designed with solar energy utilization in mind, and have 440 limited roof area for PV installation. Thus, the average PV electricity generation per house is 441 considerably low for existing houses. As shown in Table 7, the average electricity generation per 442 house is 10-15 GJ per year in the CHS. To provide a comparison, the average per house electricity 443 consumption by appliances and lighting in eligible houses is also presented in Table 7. Depending 444 on the province, the average PV electricity generation is about half or less than half of the average 445 AL load per house in the CHS. Thus, a standalone PV retrofit will not be sufficient to convert 446 existing houses into low energy buildings. 447 As discussed earlier, the BIPV/T system is an alternative approach that combines the benefits of 448 the PV and heat pump systems in a hybrid system. The total energy savings and associated GHG 449 emission reductions with the BIPV/T retrofit is given in Table 8. Since the BIPV/T system requires 450 the additional eligibility criterion of a suitable mechanical room in a house, the number of eligible 451 houses for BIPV/T retrofit in the CHS is less than that for PV retrofit, also as shown in Table 8. 452 However, the energy savings due to the BIPV/T system retrofit is much higher compared to the 453 electricity generation by PV retrofit although fewer houses are eligible for the BIPV/T retrofit. 454 As shown in Table 8 for each province, the average energy savings per house with the BIPV/T 455 retrofit varies between 90-120 GJ per year, compared to the 10-15 GJ per year of electricity 456 produced per house by the PV retrofit. The significant difference between the PV and BIPV/T 457 retrofit benefits illustrates the importance of space and DHW heating load in the CHS and indicates 458 that efforts to convert existing houses into low energy buildings need to include HVAC system 459 upgrade(s) to be effective. 460 Estimates of the energy consumption in the CHS including the energy consumption in houses not 461 eligible for the BIPV/T retrofit, and energy consumption before and after the retrofit for eligible 462 houses broken down according to the energy sources used are provided in Table 9 indicating the overall effectiveness of the BIPV/T system to reduce energy consumption. The 2.2% 469 increase in NG use in QC is because the auxiliary heating with the BIPV/T system is assumed to 470 be from NG rather than the oil used in some existing houses. Thus, all oil consumption by eligible 471 houses is replaced with NG after the BIPV/T retrofit. 472 Annual energy savings due to BIPV/T system retrofit in the CHS is provided in Table 10. Overall, 473 with the BIPV/T system retrofit, the energy consumption of the eligible houses reduce from 350.8 474 PJ to 123.8 PJ, corresponding to a reduction of 65%. However, due to the low percentage of 475 eligible houses in the CHS for the BIPV/T system retrofit (25% as shown in Table 8), the energy 476 savings across the entire CHS is about 18% as shown in Table 11. This is about six times more 477 than the savings due to the PV retrofit as shown in the same table. 478

Reduction of GHG emissions 479
The PV and BIPV/T system retrofits not only reduce energy use in the CHS but also replace a 480 portion of the fossil fuel use (including onsite oil and NG as well as offsite fuel use for electricity 481 generation) with more sustainable options. It is assumed here that PV electricity generation only 482 offsets marginal electricity generation. In provinces where marginal electricity generation is 483 mainly from fossil fuels, PV electricity generation translates into a considerable GHG emission 484 reduction as shown in Table 7. However, GHG emission reductions due to PV retrofit is negligible 485 in NF, QC, MB and BC where hydroelectricity is largely responsible for all, including marginal, 486 electricity generation. While PE and SK use fossil fuels for base electricity generation, the 487 marginal GHG EIF is relatively low as shown in Table 3; thus, the GHG emission reductions in 488 those provinces are also negligible. 489 The estimates for total and average per house GHG emission reductions due to BIPV/T system 490 retrofit are presented in Table 8. Although PV electricity generation offsets the fossil fuel use for 491 marginal electricity generation, the heat pump consumes electricity instead of a fossil fuel for space 492 and DHW heating. Thus, in NF, QC, MB and BC where hydroelectricity is widely available, the 493 BIPV/T system is a favorable option. In Atlantic Provinces where oil is widely used for heating 494 purposes by residential customers and fossil fuels are used for electricity generation, the situation 495 is more complicated. While PV electricity generation is favorable for GHG emission reduction, 496 the heat pump electricity use for heating purposes has an adverse effect on GHG emissions. The 497 most negative impact on GHG emission is predicted in AB where fossil fuels, including coal, are 498 used for electricity generation whereas significantly cleaner NG is mainly used for residential 499 heating purposes. 500 The GHG emission reductions due to BIPV/T system retrofit by fuel source is provided in Table  501 10. The GHG emissions of fossil fuels are reduced in all provinces. It should be noted that the 502 GHG emissions associated with oil is replaced with GHG emissions due to NG with BIPV/T 503 retrofits in QC. As a result, the GHG emissions due to NG increases in QC while the overall GHG 504 emissions from fossil fuels decrease. Percent GHG emission reductions due to PV and BIPV/T 505 retrofits are presented in Table 11. Since the largest GHG EIF is in NB and AB, the largest GHG 506 emission reductions by PV retrofit occur in those provinces. However, due to the large marginal 507 GHG EIF in those provinces, the GHG emission reductions due to BIPV/T retrofit is not significant 508 compared to other provinces. Using heat pumps in place of conventional fossil fuel fired heating 509 systems provides a major benefit to reduce GHG emissions in the provinces where hydro-510 electricity is the main source of marginal electricity generation. 511

Economic feasibility 512
The results of the economic analysis conducted for three fuel escalation rates, three interest rates 513 and two payback periods (as discussed in Section 3) are provided in Tables 12 and 13 for the PV  514 and BIPV/T system retrofits. The TCC is highly influenced by the reduction in fossil fuel use and 515 net electricity purchase from the grid, and it varies substantially in the range of 1,250C$ to 7,600C$ 516 for the PV and 550C$ to 43,000 C$ for BIPV/T systems. 517 The energy savings due to electricity generation is used in the calculation of the TCC for PV 518 systems. As shown in Table 12, the TCC is the highest in NB largely because the average 519 electricity generation per house is maximum and the price of electricity is relatively high compared 520 to other provinces. On the other end of the spectrum, the TCC for QC is the lowest in Canada. This 521 is because QC has the lowest price of electricity and the per house electricity generation is close 522 to average amongst all provinces as shown in Table 7. While the price of electricity is third highest 523 in AB, the TCC is one of the lowest because AB has the lowest average electricity generation per 524 house. 525 For BIPV/T systems, TCC is affected by PV electricity generation as well as the change in end-526 use energy consumption. The significant reduction in oil consumption in AT provinces (i.e. NF, 527 NS, PE and NB) and QC substantially increases the TCC for BIPV/T systems compared to the 528 TCC of PV systems. While the fossil fuel consumption decreases due to BIPV/T retrofit, electricity 529 demand increases as a result of the heat pump operation. Thus, if BIPV/T replaces an inexpensive 530 fossil fuel, i.e. NG, with a relatively higher priced electricity, this results in a low TCC as seen in 531 AB. 532 The low TCC values for PV retrofit indicate that the PV systems will not be considered attractive 533 by Canadian households in the absence of substantial subsidies. The BIPV/T systems are 534 economically more attractive with higher TCC values, but considering the higher capital costs 535 required by these systems, external economic forces such as energy rebates, government subsidies, 536 incentive measures, and legislation (such as Carbon tax that panelizes fossil fuel use) will likely 537 be necessary to promote their wide scale adoption. 538

Conclusion 539
The performance of PV and BIPV/T system retrofits in the CHS was investigated considering 540 energy savings, GHG emission reductions and economic feasibility. It was assumed that the 541 retrofits were applied to all houses that are suitable for the installation without the need for major 542 renovations. The findings are as follows: 543 • About 35% and 25% of existing houses in the CHS are eligible for PV and BIPV/T retrofits, 544 respectively. 545 • If all eligible houses adopt PV electricity generation, the energy consumption in the CHS 546 will be reduced by 37.5 PJ per year, which is equivalent to 3% annual energy savings. This 547 will results in 3.27 Mt of CO2e equivalent GHG emission reductions, which is 5% of the 548 annual GHG emissions from the CHS. In NF, QC, MB and BC where utility electricity 549 generation is from renewable resources, the impact of PV retrofit on GHG emission 550 reduction is negligible. While the average per house electricity generation by PV systems 551 is similar in all provinces, the reduction in GHG emissions is not. The highest GHG 552 emission reductions occur in regions where the fuel mixture for marginal electricity 553 generation consists mainly of fossil fuels. 554 • If all eligible houses in the CHS implement BIPV/T system retrofits, the energy 555 consumption in the CHS will be reduced by 227 PJ per year, which is equivalent to 18% 556 annual energy savings. This will remove 10.85 Mt of CO2e equivalent GHG emissions, 557 which is 17% of the annual GHG emissions from the CHS. The change in total electricity 558 use of the CHS is almost negligible while the 99.9% of the annual energy savings is 559 associated with oil and NG consumption. Since replacing existing fossil fuel fired heating 560 systems with heat pumps may increase the electricity demand of some houses, the 561 associated GHG emissions due to electricity use increases in OT and AB. The overall 562 impact of BIPV/T system retrofit is favorable from both energy conservation and GHG 563 emission perspectives. 564 • The majority of energy savings and GHG emission reductions from the BIPV/T system are 565 found to occur from the heat pump and not the PV electricity generation. 566 • The economic analysis indicates that the BIPV/T system retrofit is more feasible in the AT 567 region and QC where oil consumption for space and DHW heating is significantly reduced. 568 The lowest TCC is predicted in AB where the relatively inexpensive NG use is substituted 569 with electricity. 570 • Although the maximum suitable roof area for PV panel installation was considered, the 571 standalone PV electricity generation is not sufficient to convert existing houses into 572 NZEBs. On the other hand, the BIPV/T system retrofit can substantially reduce energy 573 consumption and will be a suitable option to be included in the set of potential strategies 574 to be evaluated to achieve near NZE or NZE status for Canadian houses. 575

Acknowledgement 576
The authors gratefully acknowledge the financial support provided to this project through the 577 He, W., Zhang, Y., and Ji, J. Comparative experiment study on photovoltaic and thermal  • Techno-economic performance of PV and BIPV/T systems in Canadian houses is evaluated • Annual energy savings and GHG emission reductions by PV and BIPV/T retrofit are estimated • Net-metering billing strategy is used for accounting impact of PV electricity generation • Majority of energy savings from the BIPV/T system occur from the heat pump • BIPV/T retrofit reduce 18% of annual energy use of the Canadian housing stock The equivalent one-diode circuit.

Figure 2
A typical house with BIPV/T system retrofit.

Figure 3
Modelling of PV-roof system in ESP-r. 1 b a To avoid unnecessary complexity in the components and control algorithms of plant simulation using ESP-r, the combination of mixing valve and three way tempering valve are modeled using a fully mixed adiabatic tank and a DHW pump. b The heating system will not turn on due to the low temperature setpoint during the cooling only season.