Energy consumption and economic analyses of a district heating network

Energy 57 (2013) 149e159

Contents lists available at SciVerse ScienceDirect

Energy journal homepage: www.elsevier.com/locate/energy

Energy consumption and economic analyses of a district heating network Marouf Pirouti*, Audrius Bagdanavicius, Janaka Ekanayake, Jianzhong Wu, Nick Jenkins Institute of Energy, Cardiff University, Queen’s Buildings, The Parade, Cardiff CF24 3AA, UK

a r t i c l e i n f o

a b s t r a c t

Article history: Received 28 August 2012 Received in revised form 16 January 2013 Accepted 27 January 2013 Available online 6 March 2013

An approach for minimisation of the capital costs and energy consumption in a district heating network is presented using a case study based on a district heating network in South Wales, UK. A number of different design cases were simulated using the PSS SINCAL, taking into account different supply and return temperatures and target pressure losses. The operation of the district heating network was synthesised under different design cases using four district heating operating strategies. Optimisation was conducted to obtain the optimal flow rate and supply temperature for the variable flow and variable supply temperature operating strategy. The optimisation model was formulated using the FICOÔ Xpress optimisation suite. The objective of optimisation was to minimise the annual total energy consumption and costs. Using each operating strategy, the annual pump energy consumption, heat losses and the equivalent annual cost were found and compared. A variable flow and variable supply temperature operating strategy was found to be beneficial in all cases. Design cases with minimum annual total energy consumption and cost used small pipe diameters and large pressure drops. Further, by increasing temperature difference between supply and return pipes, the annual total energy consumption and the equivalent annual cost were reduced. Ó 2013 Elsevier Ltd. All rights reserved.

Keywords: District heating Design Operation Energy analysis Economic evaluation

1. Introduction In the UK, heat demand for homes, businesses and industrial processes accounts for around 49% of total energy demand and 47% of the carbon emissions [1]. The majority of domestic and nondomestic buildings use individual heating systems such as gas boilers or electricity. Less than 0.5% of heat is from renewable sources [1]. The UK government has a target to reduce carbon emissions to at least 34% below the base year level1 in 2020 [2], and to deliver 15% of the UK’s energy consumption from renewable sources by 2020 [3]. Renewable heat is expected to contribute approximately one-third of this overall renewable energy target. Therefore, to achieve the UK’s overall renewable energy target, around 12% of the total heat demand in 2020 will need to come from renewable sources [4]. Debates are now taking place on how to supply heat for buildings in the future in order to reduce or completely avoid the use of fossil fuels [5e8]. District heating (DH) systems have the potential to contribute to the renewable energy targets. DH systems offer primary energy savings, especially where heat and electricity are generated in a * Corresponding author. Tel.: þ44 (0) 2920875710. E-mail address: [email protected] (M. Pirouti). 1 The base year is 1990 for carbon dioxide, nitrous oxide and methane. 0360-5442/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.energy.2013.01.065

single unit (CHP) or waste heat from existing power plants is recovered. In addition, DH has the flexibility to accommodate heat from a variety of renewable heat sources including: biomass, solar thermal and geothermal. A study conducted by BRE (Building Research Establishment Ltd.) [9], shows that under the right conditions,2 DH could supply up to 14% of the UK’s heat demand. It could be a cost effective and viable alternative to individual heating technologies while reducing bills for consumers. According to the National Heat Map for England [10], 50% of the heat demand in England is concentrated with sufficient density to make a DH network worth investigating.3 However, there are a number of economic barriers which would have to be addressed in order to make DH competitive in comparison with alternative heating technologies. Bringing down the cost of DH pipe infrastructure along with reducing heat losses and pump electrical energy consumption would reduce the capital costs and increase the economic competitiveness of DH compared with other technologies. The high cost of DH is mainly attributed to the capital costs of the hot water pipe network [1]. The investment in the DH pipe network mainly depends on pipe length and diameter [11]. An over-dimensioned 2 District heating is best suited to urban areas with high heat demand and a mix of different building types. 3 District heating becomes economically viable at heat density of 3000 kW/km2 [10].

150

M. Pirouti et al. / Energy 57 (2013) 149e159

_ m Dp P T

Nomenclature

Abbreviations CHP combined heat and power CFeCT constant flow and constant temperature CFeVT constant flow and variable temperature DH district heating EAC equivalent annual cost MPL maximum pressure loss NPV net present value SLP sequential linear programming TPL target pressure loss VFeCT variable flow and constant temperature VFeVT variable flow and variable temperature

x h r

Subscripts and superscripts c consumer e electricity fuel fuel j index for pipe section loss heat losses min minimum M number of pipe section o time step 0 p pump r return s supply src source t time step T total

Parameters A annuity factor C cost, £ d diameter, mm or m E energy, kWh or MWh f friction factor i interest rate, % l pipe length, m

DH network increases total installation and operating costs, while an under-dimensioned DH network may affect the supply of heat. DH systems usually operate at part loading conditions hence a DH pipe network designed to carry full load in an area with low energy density is uneconomic [12]. The heat losses in a DH network are affected by pipe diameters and the insulation material used, as well as the temperature of the heat carrier medium in the supply and return pipes. Pipe diameters also have an impact on pressure loss in DH and consequently on the electrical energy consumption of pumps. The electrical energy consumption of pumps is also influenced by the flow rate of the heat carrier [13]. As a result, special attention needs to be paid to the determination of pipe diameters as well as the way the system is operated. Pressure loss per unit length or target pressure loss (TPL) is a common design parameter used in DH pipe network design. Traditional methods of determining DH pipe sizes involve the use of a size searching algorithm in which the smallest pipe diameter is selected in accordance with the maximum TPL [14]. As a rule of thumb, many DH networks in Denmark and in other European countries have been designed using TPL of 100 Pa/m [15e17]. Wide ranges of TPL were used in various studies to determine pipe diameters in DH networks as shown in Table 1.

Table 1 TPL used to determine pipe diameters in district heating networks.a

District heating circulation pumps are another component which should be chosen to ensure sufficient flow circulation in the network. Traditionally, pumps are chosen using the maximum pressure difference for the most remote consumer. Designing a DH pipe network according to the maximum TPL may result in an inefficient and unnecessary costly DH network. The risk of having an inefficient and unnecessary oversized pipe network can be reduced by considering the variable TPL, the annual heating demand and DH operating strategy (or method) during design of the pipe network. However, the variation of TPL, heat demand and DH operating method contribute to the difficulty of the decision-making process. The objective of this study is to investigate various designs of a DH pipe network and develop a method which allows the design of an energy efficient and cost effective DH network. Different design cases, using different TPL and supply and return temperatures were obtained using commercial software PSS SINCAL [25]. These design cases were then operated under four DH operating strategies, assuming that the DH network is fed by an ideal heat source (i.e. a source which has negligible operating cost and can meet the demand without any restrictions). For the variable flow and variable supply temperature (VFeVT) operating strategy the optimum supply temperature and flow rate were found using the FICOÔ Xpress optimisation suite [26]. Finally, economic analysis was conducted using the Microsoft EXCEL. Obtained results were compared. 2. Method

TPL (Pa/m)

Pipe network

References

30e70 50e200 100 150 200 500 1500 2000

Primary Primary Primary Primary Primary Primary Primary Primary

[18] [14,19] [20,21] [22] [23] [24] [24] [11,12]

network network network network network network (main pipes) network (street pipes) network

mass flow rate, kg/s differential pressure, Pa or bar electrical power, kW or MW temperature,  C consumers pressure drop coefficient efficiency, % water density, kg/m3

First, a number of design cases were simulated. Then annual energy performance along with the equivalent annual cost (EAC) of the design cases were investigated when they were operated under varying conditions of outside air temperature using different operating methods. 2.1. District heating topology

a

A district heating network consists of a primary network (long distance heat transport network), and a secondary network (distribution after heat exchange substation or building’s internal heating system).

First, the topographical configuration of the DH network was determined. A real redevelopment project in Ebbw Vale, South

M. Pirouti et al. / Energy 57 (2013) 149e159

151

Fig. 1. Simplified diagram of the Ebbw Vale district heating project.

Wales, UK was used (Fig. 1) [27]. For the sake of simplicity, the consumers within a geographical area are represented by a cluster. The consumers within a cluster may have different building sizes and occupancy patterns. It was assumed that consumers were connected to the network using heating substations.

was calculated at maximum heating load assuming initial pipe diameters in the network. Pressure loss was calculated using the maximum flow rate. A range of TPL values were taken into account and the pipe diameter of each section was calculated at maximum flow rate based on each TPL. The pipe diameters were calculated using the following equation:

2.2. Calculation of energy use in buildings

vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u u 8fj m _ 2j;max 5 ! dj ¼ u u u Dpj t rp2 lj

Maximum heat demand was calculated for each consumer. Energy demand for space heating and domestic hot water were found based on the estimated area and heat load density (W/m2)4 for each building [28,29]. In order to calculate annual heating load, it was assumed that the energy requirement for domestic hot water system was constant over the year. Variation of space heating over the year depends on the outdoor temperature. The variable heating load over the year for space heating was calculated using the concept of heating degree days [30,31]. 2.3. District heating design cases To obtain design cases, maximum supply temperature at the heat source and return temperature at the consumer’s heating substations were assumed to be known5 [32]. Several regimes of supply and return temperatures were considered. For each temperature regime, maximum mass flow rate (or volume flow rate)

4

CIBSE Guide F benchmark for new buildings. Return temperature depends in a non-linear way on the heat load, supply temperature and consumer behaviour. For the sake of simplicity, return temperature was assumed to be known at the consumer’s heating substations.

j ¼ 1.M

(1)

where d and l are pipe diameter and pipe length, f is the friction _ is mass flow rate and Dp is pressure loss. factor, m A pipe diameter calculated based on the TPL value may be different from those available on the market. Therefore, the pipe with a diameter closest to the calculated pipe diameter was selected. By repeating the calculation with actual pipe diameters, the actual maximum pressure loss (MPL) in the network was obtained. Pump size was calculated to overcome loss of pressure along the route with maximum pressure drop in the network. The pump power in kW was calculated using the following equation [33]:

Pp ¼

_ Dpp m 1000rhp

(2)

where hp is the pump’s overall6 efficiency and Dpp is pump differential pressure. It was assumed that pump efficiency was 80% [34,35]. Pump differential pressure was calculated by:

5

6

Mechanical and electrical efficiency.

152

Dpp ¼

M. Pirouti et al. / Energy 57 (2013) 149e159

X



Dpsj þ Dprj þ Dpc

(3)

j

where Dpsj and Dprj are pressure drop in pipe sections of supply and return respectively. Dpc is the pressure drop in the consumer’s heating substation. This was calculated using the following equation:

_2 Dpc ¼ xm

(4)

The consumer’s pressure drop coefficient x was calculated by assuming that the maximum pressure drop at the consumer’s heating substation is 50 kPa [22], at maximum heating load and maximum flow rate. 2.4. District heating operating strategy The DH design cases were operated for one year using several operating strategies. In a DH system intended to supply a consumer’s energy requirement, two parameters can be controlled: supply temperature and flow rate [36]. Therefore, the following four different operating strategies were investigated7:  Constant flow and constant supply temperature (CFeCT): System operated at the maximum heating load, maximum temperature and maximum flow rate.  Constant flow and variable supply temperature (CFeVT): Flow rate was assumed to be constant and system supply temperature was controlled according to variable heat demand.  Variable flow and constant supply temperature (VFeCT): Supply temperature was assumed to be constant, but the flow rate varied according to the heat demand over the year. Therefore, the pressure drop and the required pump head varied accordingly.  Variable flow and variable supply temperature (VFeVT): The VFeVT is the combination of CFeVT and VFeCT operating methods. Control variables, flow and supply temperature were adjusted simultaneously with respect to the variation of heat demand. Using the CFeCT and the CFeVT systems, where mass flow rate is constant, constant speed pumps were used. For the VFeCT and VFeVT operating method, where mass flow rate varied depending on the heating demand, the variable speed pumps with frequency converters were used. In the CFeCT operating method it was assumed that the DH network is operating at the maximum heating load (peak demand) over the whole year. In the other three cases, it was assumed that the load varies over the year according to the variation of the outdoor temperature. In the VFeVT method, the variation of the system flow and system supply temperature increases the complexity of the analysis. Hence, an optimisation model was developed to obtain the optimum flow and supply temperature over the year according to change in heat demand. For all four operating methods, it was assumed that the DH network was fed by an ideal heat source (ideal-DH). The ideal heat source was defined as a heat source which had negligible operating costs, which was capable of delivering required amount of heat. Close examples of an ideal source are: solar thermal or geothermal. For the VFeVT operating strategy, in addition to the ideal-DH, DH network with a boiler (boiler-DH) and CHP (CHPeDH) as a heat

7 Return temperature was assumed to be known at the consumer’s heating substations.

source was investigated. The optimisation was based on minimum annual total energy consumption or annual total operating costs. The following objective function was used to minimise annual total energy consumption of the system:

min

X

t t Esrc;fuel þ Ept þ Eloss



(5)

t

where Ep and Eloss are pump electrical energy consumption and heat energy losses. Esrc,fuel is the fuel energy consumption of the heat source. Fuel consumption is zero for an ideal source. In the case of optimisation based on cost the following objective function was used:

min

X

t Csrc;fuel þ



  t t þ Cpt þ Closs  Cchp;e

(6)

t

where Cp and Closs are pumping cost and cost associated with heat losses. Csrc, fuel is the fuel cost, also considered as zero for an ideal source. In the case of CHP the electricity revenue, Cchp,e (given within parentheses), was added as a negative cost to the objective function [37,38]. CHP with back pressure steam turbine was assumed in this study. Electricity generation by the CHP plant was described using the model proposed by Savola et al. [21,39]. The constraints used for both optimisation approaches are:

Ts;min  Ts;t  Ts;max

(7)

_tm _ max m

(8)

Dpp;t  Dpp;max

(9)

Using each operating strategy, the DH network parameters including annual pump energy consumption and heat energy losses as well as the EAC of the DH network were calculated and compared. 2.5. Modelling and analysis of district heating PSS SINCAL was used for modelling and simulation of the DH network [25]. Using the Hardy Cross’s method [40], PSS SINCAL allows simulation of large and complex DH pipe networks. PSS SINCAL steady state calculation determines heating parameters such as flow rate, temperature, pressure, heat losses and velocity in the network. Additionally, multiple time series analysis can be investigated. Using PSS SINCAL the main pipe network (primary network) was modelled, including supply and return pipes. It was assumed that return pipes have the same diameters and length as the supply pipes. Pipes were laid in the ground. An average ground temperature of þ7  C over the whole year was assumed [41,42]. Standard size pre-insulated single steel pipes with a pressure rating up to 25 bar and temperature rating up to 140  C were used. Pipe roughness of 0.4 mm was used in calculations [43]. Four different regimes of maximum supply temperature at the heat source and return temperature at the consumer’s heating substations were assumed for the analysis, Ts,max/Tr,max: 120/70  C, 110/70  C, 100/70  C and 90/70  C [44,45]. A range of TPL values were assumed in which the maximum differential pressure of a pump is less than or equal to 16 bar [44]. Supply and return pipe diameters and pump sizes were calculated for a range of TPL values and temperature regimes at the maximum heating load. The design cases were operated using different operating strategies according to annual heating load. The overall block diagram of the study is shown in Fig. 2.

M. Pirouti et al. / Energy 57 (2013) 149e159

153

Fig. 2. Block diagram of the study.

For the VFeVT operating strategy, the concept of graph theory was used to describe the heating pipe network topology [46]. The energy flow and flow rate at each branch and the temperature and pressure at each node of the network were calculated using equations found in thermal engineering text books [25,47]. The nonlinearity of the heat network equations was dealt with through Sequential Linear Programming (SLP) [26]. 3. Case study: Ebbw Vale district heating A real redevelopment project in South Wales, UK, was used as the basis of a case study to analyse the DH network (Fig. 1). For the calculation of annual heating load, minimum outdoor temperature was 3  C and the base temperature was assumed to be þ15.5  C. The degree days were calculated based on the weather data for Cardiff, South Wales, UK. Annual total heat demand was obtained (space heating plus domestic hot water), for the case study shown in Fig. 1. The calculated annual heating load over the year (left) and load duration curve (right) in MW are shown in Fig. 3. It was observed (Fig. 3) that the annual total heat load is divided into two main seasons (summer and winter). It was assumed that the winter season lasts for 182 days, which includes demand for space heating and domestic hot water. For the rest of the year (summer season) only the demand for domestic hot water was taken into account. For the DH network (Fig. 1), a number of design cases with different sizes of pipes and pumps were obtained, using varying

temperature regimes and TPL values at maximum heating load. The MPL obtained using this procedure is discussed in Section 2.3, and different design cases are presented in Table 2. 4. Results The DH design cases obtained in Section 2.3 are given in Table 2. They were operated for one year using different operating methods described in Section 2.4. First, annual pump energy consumption and heat losses were calculated for each operating method and temperature regime, and then economic analysis was conducted. 4.1. Energy consumption and heat losses analysis 4.1.1. Energy consumption and heat losses for the CFeCT, CFeVT and VFeCT methods The pump electrical energy consumption and heat losses for all four temperature regimes were obtained when the DH network was operated under CFeCT, CFeVT, and VFeCT operating methods. The results obtained for the design cases based on temperature regime of Ts,mas/Tr,max: 120/70  C, are shown in Fig. 4. For the CFeCT method, the supply temperature at the source side and the return temperature at the consumer’s substation were fixed at 120  C and 70  C over the year. Since the analysis was carried out under maximum load, the flow rate was maximum. Using this operating strategy the energy demand variation for space heating and domestic hot water was not taken into account.

Fig. 3. Annual heat demand (left), and load duration curve (right).

154

M. Pirouti et al. / Energy 57 (2013) 149e159

Table 2 Design cases with obtained maximum pressure loss. Temperature regimes: Ts,max/Tr,max,  C TPL (Pa/m) 120/70

110/70

100/70

90/70

DH case MPL (Pa/m) Maximum DH case MPL (Pa/m) Maximum DH case MPL (Pa/m) Maximum DH case MPL (Pa/m) Maximum velocity (m/s) velocity (m/s) velocity (m/s) velocity (m/s) 50 100 150 200 250 300 350 400 450 500 550 600 700 800 900 1000 1100 1200

DH1 DH2 DH3 DH4 DH5 DH6 DH7 DH8 DH9 DH10 DH11 DH12 DH13 DH14 DH15 DH16 DH17 DH18

52 99 156 214 248 304 339 407 448 483 552 592 661 790 847 977 1276 1403

1.25 1.73 1.73 1.96 2.48 2.31 2.48 3.48 2.92 2.92 3.48 2.92 3.48 3.48 3.48 3.48 4.20 4.20

DH19 DH20 DH21 DH22 DH23 DH24 DH25 DH26 DH27 DH28 DH29 DH30 DH31 DH32 DH33 DH34 DH35 DH36

45 99 153 196 262 304 331 402 470 524 576 630 694 748 800 854 1023 1312

1.08 1.55 2.14 2.14 2.43 2.48 2.48 2.86 2.86 3.08 4.32 4.32 3.62 3.62 4.31 4.31 4.31 4.31

The temperature and flow rates for the CFeVT and VFeCT methods are shown in Fig. 4a and b, respectively. For the CFeVT, flow rate is constant for the winter and summer seasons. Supplying the consumer’s energy requirement during winter season

DH37 DH38 DH39 DH40 DH41 DH42 DH43 DH44 DH45 DH46 DH47 DH48 DH49 DH50 DH51 DH52 DH53 DH54

51 101 143 191 269 299 345 382 462 536 582 723 828 923 1015 1110 1223 1318

1.20 1.73 2.06 2.13 2.83 2.69 2.83 3.21 3.21 3.21 3.21 4.75 3.78 4.07 5.71 5.71 4.79 4.79

DH55 DH56 DH57 DH58 DH59 DH60 DH61 DH62 DH63 DH64 DH65 DH66 DH67 DH68 DH69 DH70 DH71 DH72

55 97 152 197 244 318 367 389 425 496 577 667 748 851 948 1029 1195 1297

1.41 2.13 2.13 2.37 2.57 3.06 2.66 3.06 3.17 3.17 4.78 4.01 4.78 4.78 4.21 4.78 4.78 4.78

requires the variation of supply and return temperatures between assumed limits, according to annual heat demand (see Fig. 3). For the summer season there is only demand for domestic hot water hence, system temperature is fixed at its minimum level. Using the

Fig. 4. (a) Flow, (b) Temperature, (c) Annual pump energy consumption and heat losses, and (d) Annual total energy consumption, using CFeCT, CFeVT and VFeCT operating methods.

M. Pirouti et al. / Energy 57 (2013) 149e159

155

Table 3 Design cases with minimum annual total energy consumption, using different temperature regimes and CFeCT, CFeVT and VFeCT operating methods. Operating method

CFeCT

CFeVT

VFeCT

Ts,max/Tr,max,  C

DH case

ET,min (MWh/year)

DH case

ET,min (MWh/year)

DH case

ET,min (MWh/year)

120/70 110/70 100/70 90/70

DH2 DH20 DH37 DH55

823 839 884 953

DH2 DH20 DH38 DH55

522 546 580 646

DH12 DH26 DH42 DH62

486 518 527 577

VFeCT operating method, supply and return temperatures are fixed for the winter and summer seasons. Flow rate varies according to annual heat load in order to supply the consumer’s energy demand. The annual pump energy consumption and heat losses for the CFeCT, CFeVT and VFeCT methods are shown in Fig. 4c. The annual total energy consumption (annual pump energy consumption plus heat energy losses), are presented in Fig. 4d. It is seen from Fig. 4c that the annual pump energy consumption and heat losses vary, when different design cases shown in Table 2 (only few are indicted in Fig. 4), are operated under different operating methods. From Fig. 4, it can be seen that the annual total energy consumption is the highest for the CFeCT method. For the CFeVT and VFeCT methods, heat energy losses and pump energy consumption are reduced substantially. For all operating methods, almost a linear increase of pump energy consumption and a reduction of heat losses are observed along with increases in the MPL. For the CFeCT method, the case with minimum annual total energy consumption is the design case DH2 (MPL: 99 Pa/m). For the CFeVT method, the most desirable case is the design case DH2 (MPL: 99 Pa/m). However, the case with minimum annual total energy consumption using the VFeCT method is the DH design case DH12 (MPL: 592 Pa/m), which has relatively higher pressure loss. Table 3 summarises the cases with minimum annual total energy consumption under all four temperature regimes. It is shown that when different temperature regimes are used the minimum annual total energy consumption changes. Reducing temperature difference between supply and return pipes increases system flow rate which in turn increases annual pump energy consumption. The results indicate that a better energy performance of the DH network can be achieved with the VFeCT method compared with the other operating methods. For this operating method, relatively small pipes with large pressure losses can be designed. 4.1.2. Energy consumption and heat losses for the VFeVT method The VFeVT operating method complements the CFeVT and VFe CT methods. Using the VFeVT, both control variables of supply temperature and flow rates were adjusted simultaneously to meet the demand. Three different heat sources were analysed to examine

the impact of the heat source on optimal solution of flow rate and supply temperature during operation, and consequently upon DH pump energy consumption and heat losses. A DH network connected to an ideal heat source (ideal-DH), a DH network with a boiler (boiler-DH) and a DH network with a CHP plant (CHPeDH) were investigated. It was assumed that all types of heat sources provide the full amount of required heat. Using the objective function given in Eq. (5) and the constraints given in Eqs. (7)e(9), the optimum supply temperature and flow rate were calculated at a steady state condition over the year, according to change in annual heat demand. For design cases based on a temperature regime of Ts,max/Tr,max: 120/70  C (see Table 2), the supply temperature was constrained at the heat source between a maximum of 120  C for the winter season and a minimum of 70  C for the summer season. Similarly, for design cases based on temperature regimes of Ts,max/Tr,max: 110/ 70  C, 100/70  C and 90/70  C, a maximum supply temperature of 110  C, 100  C and 90  C and minimum supply temperature of 70  C were considered. For the summer season, when there is demand for domestic hot water but not space heating, the supply temperature was fixed at 70  C on the heat source within all temperature regimes. It was assumed that return temperature was known at the consumer’s heating substations. For the sake of simplicity, a constant return temperature of 40  C at the consumer’s heating substations for the winter season and 30  C for the summer season were assumed, using all temperature regimes [32]. For the case of ideal-DH, the objective function includes pump energy consumption and heat energy losses in the pipes. For the case of boiler-DH, the objective function includes boiler fuel consumption, pump energy consumption and heat losses in the pipes. For the case of CHPeDH, the objective function includes heat losses, pump energy consumption and CHP fuel consumption. The objective of the optimisation is to minimise the annual total energy consumption of each DH design case considering different type of heat sources. Using the obtained optimal supply temperature and flow rate the optimum annual pump energy consumption and heat energy losses were determined. The optimum annual pump energy consumption and heat losses for the design cases based on temperature regimes of Ts,max/Tr,max: 120/70  C are shown in Fig. 5a and

Fig. 5. (a) Optimum annual pump energy consumption and heat losses, and (b) Optimum annual total energy consumption, using VFeVT operating method.

156

M. Pirouti et al. / Energy 57 (2013) 149e159

Table 4 Design cases with minimum optimal annual total energy consumption, using different temperature regimes and VFeVT operating method. System

Ideal-DH

Ts,max/Tr,max,  C

DH case

ET,min (MWh/year)

Boiler-DH DH case

ET,min (MWh/year)

DH case

ET,min (MWh/year)

120/70 110/70 100/70 90/70

DH16 DH35 DH44 DH68

367 378 389 389

DH16 DH35 DH44 DH68

377 390 395 393

DH16 DH35 DH44 DH68

371 383 394 400

the optimum annual total energy consumption (annual pump energy consumption plus heat energy losses) of the DH network are shown in Fig. 5b. It can be observed that for the VFeVT method the impact of the energy source on the optimum annual total energy consumption excluding the source energy consumption is negligible. Design case DH16 (MPL: 977 Pa/m) shows minimum optimum annual total energy consumption. The optimum annual total energy consumption for the VFeVT operating method was calculated under all four temperature regimes. The DH design cases with minimum optimum annual total energy consumption are presented in Table 4. Table 4 shows that optimum annual total energy consumption varies for different temperature regimes. The design cases with larger temperature difference between supply and return pipes have less annual total energy consumption. Reducing temperature difference between supply and return pipes increases system flow rate, and therefore, annual pump energy consumption increases. It is worth noting that changing temperature limit (i.e. return temperature) would affect the obtained optimal supply temperature and flow rate. Therefore, for different temperature settings the optimal annual total energy consumption will be dissimilar.

CHPeDH

Pump investment costs: these consist of costs for two pumps, one for the winter season and one for the summer season. It was assumed that the pumps were equipped with variable speed drives. Prices of the pumps were taken from Ref. [51], and prices of variable speed drives were found in Ref. [52]. Operating costs of a DH network: pumping costs (costs of electricity consumption and CO2) and costs associated with heat losses were taken into consideration. The EAC was calculated using the following equation:

EAC ¼

NPV A

(10)

Net present value (NPV), considering the life time of the systems (n years), were calculated using the equation:

NPV ¼

X

Ct

t

ð1 þ iÞt

!

þ Co

(11)

Annuity factor (A) was obtained using the following equation:

A ¼

ð1 þ iÞt  1 i ð1 þ iÞt

(12)

The data used to calculate EAC is given in Table 5. 4.2. Economic analysis All DH design cases were examined using EAC. The EAC comprises both capital and operating costs of the DH pipe network. The capital costs include pipe investment costs and pump investment costs. Pipe investment costs consist of: a. The price of pre-insulated steel pipes including fittings, site joints and termination seals: this is based on a price list obtained from Ref. [48]; b. The cost of civil works: this depends on the pipe size, ground condition and method of digging. The ground condition and digging type were assumed to be the same for all pipes. Civil work costs between 700 and 1000 (£/m) [49,50] were assumed according to the pipe size.

4.2.1. Economic evaluation for the CFeCT, CFeVT and VFeCT methods The economic analysis was conducted for different design cases using different operating methods. The EAC of different design cases for the CFeCT, CFeVT and VFeCT methods under temperature regime of Ts,max/Tr,max: 120/70  C are shown in Fig. 6 (only a few design cases are shown in the figure; other cases can be obtained by correlating the MPL with that shown in Table 2). The cost comparison clearly shows that the EAC is less when the VFeCT operating method is used. For the CFeCT method the case with minimum EAC, is the DH3 (MPL: 156 Pa/m). For the CFeVT method the EAC is minimum when DH4 (MPL: 214 Pa/m) is used.

Table 5 Physical and economic data. Reference from where data were taken Operating time (h/year) Pump life time (year) Pipe network life time (year) Interest rate (%) Inflation rate (%) Electricity price (£/MWh) Heat price (£/MWh) CO2 emission linked to the grid electricity (kgCO2/kWh) CO2 price (£/tCO2) Fuel cost (£/MWh)

8760 15 30 7 5 95 70 0.422

[53] [54] [53,55] [1] [53]

22 43

[1] [56]

Fig. 6. The EAC of the design cases, using CFeCT, CFeVT and VFeCT operating methods.

M. Pirouti et al. / Energy 57 (2013) 149e159

157

Table 6 Design cases with minimum EAC, using different temperature regimes and CFeCT, CFeVT and VFeCT operating methods. Operating method

CFeCT

Ts,max/Tr,max,  C

DH case

EACmin (£/year)

DH case

CFeVT EACmin (£/year)

DH case

EACmin (£/year)

120/70 110/70 100/70 90/70

DH3 DH22 DH38 DH55

242,969 252,703 270,556 301,220

DH4 DH22 DH40 DH66

197,805 204,802 218,403 248,320

DH12 DH27 DH45 DH63

184,211 193,429 200,336 218,346

For the VFeCT method the EAC is minimum for DH12 (MPL: 592 Pa/ m) where the pressure loss is higher. This relates to the variation of flow rate during operation. Changing flow rate during operation reduces pumping cost. Consequently, design case with smaller pipe diameters was found to be desirable. This in turn reduces the EAC of the heat network installation and operation. It can be observed that the difference between the EAC of DH2e DH4 design cases using CFeCT method, DH3eDH6 for the CFeVT operating method and D11eDH17 for VFeCT is small. This small difference is due to the fact that as the pipe diameters decrease, pressure loss in the system increases. Therefore, pipes investment costs decrease while pump investment costs increase. In addition, when using smaller pipe sizes, the costs of heat losses reduce while pumping costs increase due to higher pressure drop. The EAC was also calculated for other temperature regimes and the most cost effective cases were determined. Results are presented in Table 6. It is seen that the EAC is increased when using DH cases with less temperature difference between supply and return pipes. Reducing temperature difference between supply and return pipes increases system flow rate, and has previously been indicated. Design cases with larger flow rates require pipes and pump with larger sizes, and therefore both investment and pumping costs are increased. 4.2.2. Economic evaluation for the VFeVT operating method The DH cases were operated and analysed using the VFeVT operating method under different heat sources. Using the objective function given in Eq. (6) and the constraints given in Eqs. (7)e (9), the optimum supply temperature and flow rate were determined. The objective of the optimisation is to minimise annual total operating costs of the system. For the ideal heat source (ideal-DH), the total operating costs include pumping costs and the costs associated with the heat losses. For boiler-DH the total operating costs comprise pumping costs, the costs associated with heat losses and boiler fuel consumption. Pumping costs, the costs associated

VFeCT

with heat losses and CHP fuel were taken into account when CHPe DH case was considered. Furthermore, for the CHPeDH, the revenue of electricity was added as a minus cost to the objective function. Using the optimum supply temperature and flow rate, optimum heat losses and pumping costs were calculated for each DH design case. Then, the EAC was determined using Eqs. (10)e(12). The EAC of the design cases based on temperature regime of Ts,max/Tr,max: 120/ 70  C are presented in Fig. 7. These results only include the capital and operational costs of the pipes and pumps (DH network) and do not include capital and operational costs of the sources (the optimisation was based on operational costs of the source and network). It is seen that the type of heat source affects the EAC. For the ideal-DH and boiler-DH cases, the cost is minimum for DH17 (MPL: 1276 Pa/m). For CHPeDH, the minimum cost happens when DH9 (MPL: 448 Pa/m) was used. Using CHPeDH, it is shown that the cost effective design case has less pressure drop in comparison to the ideal-DH and boiler-DH. The reason for this relates to the fact that in CHPeDH, heat and electricity are generated simultaneously and the price of electricity is higher than that of heat. Therefore, to achieve an economical operation of CHPeDH (back pressure steam turbine CHP); electricity production is increased by the optimiser. This results in a reduction of the supply temperature which in turn, reduces heat losses in the DH network. However, it also results in rising flow rate to supply energy demand and, in turn, increased pumping costs. Overall, for the CHPeDH, pumping costs were larger and the costs of heat losses were less compared with the boiler-DH and ideal-DH cases. Therefore, the EAC of the CHPeDH was larger than that of the boiler-DH and ideal-DH cases. The EAC was calculated for all temperature regimes, using the VFeVT operating method. The cases with minimum cost were determined. The results are presented in Table 7. As previously explained, it can be seen that for the design cases with less temperature difference between supply and return pipes, the EAC of the DH network installation and operation are higher than the other cases. Additionally, for the CHPeDH the cost effective design case has larger pipe diameters and smaller pressure drops compared with those obtained for the ideal-DH and boiler-DH. Finally, it is worth mentioning that the obtained results will change when using different temperature settings and energy prices. 4.3. Comparison

Fig. 7. The EAC of the design cases, using VFeVT operating method.

It was shown that for the design cases based on temperature regime of Ts,max/Tr,max: 120/70  C, the minimum annual total energy consumption and the EAC were lowest when using the VFeVT compared with the other operating methods. From the results given in Tables 3 and 4 it can be seen that the design case DH16 (MPL: 977 Pa/m) was the design with minimum annual total energy consumption for all investigated type of heat sources. Similarly, from Tables 6 and 7 the design case DH17 (MPL: 1276 Pa/m) was found to be the most cost effective DH design when a boiler or an ideal heat source was considered with the DH network. DH9 (MPL:

158

M. Pirouti et al. / Energy 57 (2013) 149e159

Table 7 Design cases with minimum EAC, using different temperature regimes and VFeVT operating method. System

Ideal-DH

Ts,max/Tr,max,  C

DH case

EACmin (£/year)

Boiler-DH DH case

EACmin (£/year)

CHPeDH DH case

EACmin (£/year)

120/70 110/70 100/70 90/70

DH17 DH36 DH54 DH68

165,501 170,463 175,852 184,339

DH17 DH36 DH54 DH68

166,289 171,259 176,525 184,814

DH9 DH23 DH45 DH67

180,475 183,780 185,773 188,480

Table 8 Design cases with minimum annual total energy consumption, VFeVT operating method. System

Ideal-DH

Ts,max/Tr,max-operating method

MPL (Pa/m)

Ep (MWh/year)

Eloss (MWh/year)

Boiler-DH MPL (Pa/m)

Ep (MWh/year)

Eloss (MWh/year)

CHPeDH MPL (Pa/m)

Ep (MWh/year)

Eloss (MWh/year)

120/70  C-VFeVT

977

33

333

977

59

317

977

23

347

Table 9 Design cases with minimum EAC, VFeVT operating method. System

Ideal-DH

Ts,max/Tr,max-operating method

MPL (Pa/m)

EAC (£/year)

Boiler-DH MPL (Pa/m)

EAC (£/year)

MPL (Pa/m)

EAC (£/year)

120/70  C -VFeVT

1276

165,501

1276

166,289

448

180,475

448 Pa/m) was the cost effective design case when CHP was chosen as the heat source. These best design cases are summarised in Tables 8 and 9. A comparison of results shows that the major difference between energy efficient and cost effective DH design cases can be shown when CHP is the heat source. For the ideal-DH and boiler-DH, the difference seen in Tables 8 and 9 is due to the pipe investment costs and the cost of heat losses rather than pump investment costs and the cost of pump electricity consumption. When conducting an economic evaluation of the ideal-DH and boiler-DH, it was found that the optimal solution corresponds to smaller pipe diameters. Both analytical approaches suggest that the size of pipe diameters is based on rather large TPL values for the VFeVT operating method. With large TPL values, pipe diameters reduce while pump sizes increases. Therefore, using the VFeVT operating method, the annual total energy consumption will reduce along with the EAC of the heat network installation and operation. 5. Conclusion An approach for the minimisation of the capital costs and energy consumption in a DH network was presented. A number of DH design cases (pipe and pump with different sizes) were modelled. The performance of the DH design cases over the year was conducted using different DH operating strategies such as CFeCT, CFe VT, VFeCT and VFeVT as well as different supply and return temperature regimes. The annual pump energy consumption, heat losses and the EAC of the DH design cases were compared. The results showed that supply and return temperature regime and DH operating strategy had a substantial impact on annual energy performance and the EAC. The design cases with minimum annual total energy consumption and EAC had different pipe diameters and pump sizes under different operating methods and temperature regimes. It was found that when using the VFeVT operating method the annual total energy consumption and the EAC were lower compared with other operating strategies. Furthermore, the most

CHPeDH

economical design also depended on the types of heat source. For the CHPeDH, it was economically more beneficial to have larger pipe diameters and smaller size pump compared with the ideal-DH and boiler-DH. Using the VFeVT operating method the most energy efficient and cost effective design cases of DH pipe network, considering different type of heat sources, corresponded to small pipe diameters and large pressure drops. Comparison of DH design cases with different flow rates showed that in order to achieve energy efficient and cost effective DH design, it was more advantageous to reduce system flow rate by increasing temperature difference between supply and return pipes. By reducing system flow rate, the annual total energy consumption as well as the EAC of the DH network was reduced. In the next study, a tool based on the analyses presented in this study is being developed. Moreover, the low temperature DH network is being investigated.

Acknowledgments The research comprises part of the Supergen-Highly Distributed Energy Future (HiDEF) work stream via EPSRC funding. The authors wish to thank Supergen-HiDEF consortium for their continued financial and technical support. The authors also thank Marc Rees for access to the case study data.

References [1] Davies G, Woods P. The potential and costs of district heating network: a report to the department of energy and climate change; 2009. pp. 1e134. [2] The carbon plan: delivering our low carbon future; 2011. pp. 3e4. [3] UK renewable energy roadmap; 2011. pp. 3e8. [4] The renewable heat incentive: consultation on interim cost control; 2012. pp. 9e14. [5] Lund H, Möller B, Mathiesen BV, Dyrelund A. The role of district heating in future renewable energy systems. Energy 2010;35(3):1381e90. [6] Lund H. Renewable energy strategies for sustainable development. Energy 2007;32(6):912e9.

M. Pirouti et al. / Energy 57 (2013) 149e159 [7] Lund H, Mathiesen BV. Energy system analysis of 100% renewable energy systems d the case of Denmark in years 2030 and 2050. Energy 2009;34(5): 524e31. [8] Lund H, Hvelplund F. The economic crisis and sustainable development: the design of job creation strategies by use of concrete institutional economics. Energy 2012;43(1):192e200. [9] Routledge K, Williams J. District heating e heat metering cost benefit analysis. BRE client report number 278-408. Building Research Establishment Ltd.; 2012. [10] The future of heating: a strategic framework for low carbon heat in the UK; 2012. pp. 59e73. [11] Zinko H, Bøhm B, Kristjansson H, Ottosson U, Rämä M, Sipilä K. District heating distribution in areas with low heat demand density; 2008. pp. 5e11. [12] Dalla Rosa A, Christensen JE. Low-energy district heating in energy-efficient building areas. Energy 2011;36(12):6890e9. _ Svendsen S. Improving the dimensioning of piping networks and [13] Tol HI, network layouts in low-energy district heating systems connected to lowenergy buildings: a case study in Roskilde, Denmark. Energy 2012;38(1): 276e90. [14] Yildirim N, Toksoy M, Gokcen G. Piping network design of geothermal district heating systems: case study for a university campus. Energy 2010;35(8): 3256e62. [15] Hlebnikov A, Siirde A, Paist A. Basics of optimal design of district heating pipelines diameters and design examples of Estonian old non-optimised district heating networks, Doctoral School of Energy- and Geo-technology, January 15-20, 2007, Kuressaare, Estonia. p. 149-153. [16] Hlebnikov A, Dementjeva N, Siirde A. Optimization of Narva district heating network and analysis of competitiveness of Oil Shale CHP building in Narva. Oil Shale 2009;26(3):269e82. [17] Ulloa P. Potential for combined heat and power and district heating and cooling from waste-to- energy facilities in the U.S. - learning from the Danish experience, Master thesis, Department of Earth and Environmental engineering, Columbia University, 2007. [18] Li X, Duanmu L, Shu H. Optimal design of district heating and cooling pipe network of seawater-source heat pump. Energy and Buildings 2010;42(1): 100e4. [19] Gudmundsson O, Nielsen A, Iversen J. The effects of lowering the network temperatures in existing networks. In: DHC13, the 13th international symposium on district heating and cooling, September 3rd to September 4th, 2012, Copenhagen, Denmark. p. 116-121. [20] Phetteplace G. Optimal design of piping systems for district heating; 1995. U.S. Army Corps of Engineers-Cold Regions rResearch & Engineering Laboratory CRREL report. pp. 95e17. [21] Sipilä K, Pursiheimo E, Savola T, Fogelholm CJ, Keppo I, Ahtila P. Small-scale biomass CHP plant and district heating; 2005. pp. 35e52. [22] Skagestad B, Mildenstein P. District heating and cooling connection handbook. IEA District Heating and Cooling; 1999. [23] Vallios I, Tsoutsos T, Papadakis G. Design of biomass district heating systems. Biomass and Bioenergy 2009;33(4):659e78. [24] Li H, Rosa AD, Svendsen S. Design of low temperature district heating network with supply water recirculation. In: The 12th international symposium on district heating and cooling, September 5th to September 7th, 2010, Tallinn, Estonia. p. 73-80. [25] PSSÒSINCAL 6.5 Heating: heating calculations in flow networks. SIEMENS AG; 2010. [26] Fico Xpress optimization suite: user guides. [Online]. Available: http://www. fico.com/en/Products/DMTools/xpress-overview/Pages/XpressDocumentation.aspx. [accessed 30.07.12.]. [27] Rees MT, Wu J, Awad B, Ekanayake J, Jenkins N. A total energy approach to integrated community infrastructure design. In: 2011 IEEE power and energy society general meeting, IEEE, 2011. p. 1-8. [28] Hutt B, Field J, Hobbs D, Johnson T, Palmer J, Watson A. Energy efficiency in buildings: CIBSE guide F. CIBSE; 2004. [29] Beggs C. Energy management supply and conservation. Elsevier Ltd.; 2009. [30] Day T. Degree-days theory: and application. CIBSE; 2006.

159

[31] Heating degree days (HDD). [Online]. Available: http://www.degreedays.net/. [accessed 30.07.12.]. [32] Bøhm B, Ha S, Kim W, Kim B, Koljonen T, Larsen HV, et al. Simple models for operational optimisation; 2002. pp. 1e132. [33] Snoek C, Yang L, Onno T, Frederiksen S, Korsman H. Optimization of district heating systems by maximizing building heating system temperature difference; 1999. pp. 88e90. [34] Schofield S. Variable speed driven pumps: best practice guide. British Pump Manufacturers’ Association (BPMA); 2003. [35] Grundfos data booklet NB, NBE, NK, NKE: single-stage end-suction pumps according to EN 733, 50 Hz, GRUNDFOS A/S. DK-8850 Bjerringbro. Denmark; 2010. [36] Gustafsson J, Delsing J, Van Deventer J. Improved district heating substation efficiency with a new control strategy. Applied Energy 2010;87(6):1996e 2004. [37] Benonysson A, Bøhm B, Ravn HF. Operational optimization in a district heating system. Energy Conversion and Management 1995;36(5):297e314. [38] Rolfsman B. Combined heat-and-power plants and district heating in a deregulated electricity market. Applied Energy 2004;78(1):37e52. [39] Savola T, Keppo I. Off-design simulation and mathematical modeling of smallscale CHP plants at part loads. Applied Thermal Engineering 2005;25(8e9): 1219e32. [40] Cross H. Analysis of flow in networks of conduits or conductors. University of Illinois Bulletin 1936;XXXIV(22):7e28. [41] Dalla Rosa A, Li H, Svendsen S. Method for optimal design of pipes for lowenergy district heating, with focus on heat losses. Energy 2011;36(5): 2407e18. [42] Met Office-Wales: climate. [Online]. Available: http://www.metoffice.gov.uk/ climate/uk/wl/print.html. [accessed 30.07.12.]. [43] Dobersek D, Goricanec D. Optimisation of tree path pipe network with nonlinear optimisation method. Applied Thermal Engineering 2009;29(8e9): 1584e91. [44] 2004, District heating substations design and installation. Technical requirements F:101. p. 8-10. [45] Dalla Rosa A, Boulter R, Church K, Svendsen S. District heating (DH) network design and operation toward a system-wide methodology for optimizing renewable energy solutions (SMORES) in Canada: a case study. Energy 2012; 45(1):960e74. [46] Hassine I Ben, Eicker U. Simulation and optimization of the district heating network in Scharnhauser Park. In: 2nd European conference on polygeneration, 30th March-1st April, 2011, Tarragona, Spain. p. 1-12. [47] Moran MJ, Shapiro HN, Munson BR, DeWitt DP. Introduction to thermal systems engineering: thermodynamic, fluid mechanic, and heat transfer. John Wiley & Sons, Inc.; 2003. [48] Pre insulated pipe-price list; 2010. [49] Environment, planning and housing. [Online]. Available: http://www.local. gov.uk/environment-planning-and-housing. [accessed 30.07.12.]. [50] Cost of district heating. [Online]. Available: http://chp.decc.gov.uk/cms/costof-district-heating/. [accessed 30.05.12.]. [51] Grundfos industrial products- price list October 2010. Category 2. Grundfos; 2010. [52] Inverter drive supermarket, variable speed drives (VFD), soft starter, motor and component. [Online]. Available: http://www.inverterdrive.com/group/ AC-Inverter-Drives-400V/Siemens-Micromaster-430-11kW-400V-EMC/ default.aspx. [accessed 30.05.12.]. [53] Weber C, Shah N. Optimisation based design of a district energy system for an eco-town in the United Kingdom. Energy 2011;36(2):1292e308. [54] Office for national statistic. [Online]. Available: http://www.ons.gov.uk/ons/ rel/cpi/consumer-price-indices/september-2011/index.html. [accessed 18.09.11.]. [55] Europe energy’s portal. [Online]. Available: http://www.energy.eu/#IndustrialElec. [accessed 30.05.12.]. [56] Analytical annex to the renewable heat incentive impact assessment: assumptions used in the analysis underpinning the RES. Department of Energy & Climate Change; 2010.