Sizing and control optimization of thermal energy storage in a solar district heating system

Abstract

Solar district heating systems have shown significant promise to facilitate the large scale adoption of solar energy technologies and thus substantially reduce greenhouse gas emissions. Given the mismatch between solar energy and district heating demand, energy storage devices play a critical role given their capacity to stockpile solar energy in both the short-term (hours to days) and long-term (months). However, the integration, sizing and control of energy storage technologies is far from simple. This paper investigates sizing and controlling thermal energy storage from the perspective of its performance within a district heating system, highlighting the close link between design and control. A 52-house Canadian solar district heating system, the Drake Landing Solar Community (DLSC), was used as a case study. This system uses solar collectors as main energy supply, borehole thermal energy storage (BTES) for seasonal storage and two 120-m³ water tanks for short-term thermal storage (STTS).
The effect of (a) storage sizing (STTS volume) and (b) storage control (rate at which energy is either injected or extracted from the BTES) was evaluated. A control-oriented model, calibrated and validated with operational data at 10-min intervals, was used along with an optimal rule-based control to gauge system primary energy use. Different scenarios were tested, with STTS volumes ranging from 120 m³ to 480 m³, and BTES loop nominal flow rates between 2.5 and 4.5 L/s. An optimization routine was developed to calculate the optimum parameters of the rule-based control strategy. Results show that, in comparison with the design and control in place, primary energy savings of 13%–30% (with BTES flow rates of 2.5–4.5 L/s) could have been obtained with the proposed rule-based control strategy. By decreasing the STTS volume to 120-m³, energy savings up to 6% could still be achieved; savings could reach 27%–36% by increasing the STTS size to 360-m³.

Full text

Available online at www.sciencedirect.com
ScienceDirect
Energy Reports 7 (2021) 389–400
www.elsevier.com/locate/egyr
17th International Symposium on District Heating and Cooling, DHC2021, 6–9 September
2021, Nottingham, UK
Sizing and control optimization of thermal energy storage in a solar
district heating system
Etienne Saloux∗, José A. Candanedo
CanmetENERGY, Natural Resources Canada, 1615 Boulevard Lionel Boulet, Varennes (QC) J3X 1P7, Canada
Received 27 July 2021; accepted 7 August 2021
Abstract
Solar district heating systems have shown significant promise to facilitate the large scale adoption of solar energy
technologies and thus substantially reduce greenhouse gas emissions. Given the mismatch between solar energy and district
heating demand, energy storage devices play a critical role given their capacity to stockpile solar energy in both the short-term
(hours to days) and long-term (months). However, the integration, sizing and control of energy storage technologies is far
from simple. This paper investigates sizing and controlling thermal energy storage from the perspective of its performance
within a district heating system, highlighting the close link between design and control. A 52-house Canadian solar district
heating system, the Drake Landing Solar Community (DLSC), was used as a case study. This system uses solar collectors as
main energy supply, borehole thermal energy storage (BTES) for seasonal storage and two 120-m3water tanks for short-term
thermal storage (STTS).
The effect of (a) storage sizing (STTS volume) and (b) storage control (rate at which energy is either injected or extracted
from the BTES) was evaluated. A control-oriented model, calibrated and validated with operational data at 10-min intervals,
was used along with an optimal rule-based control to gauge system primary energy use. Different scenarios were tested, with
STTS volumes ranging from 120 m3to 480 m3, and BTES loop nominal flow rates between 2.5 and 4.5 L/s. An optimization
routine was developed to calculate the optimum parameters of the rule-based control strategy. Results show that, in comparison
with the design and control in place, primary energy savings of 13%–30% (with BTES flow rates of 2.5–4.5 L/s) could have
been obtained with the proposed rule-based control strategy. By decreasing the STTS volume to 120-m3, energy savings up to
6% could still be achieved; savings could reach 27%–36% by increasing the STTS size to 360-m3.
Crown Copyright c
⃝2021 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the scientific committee of the 17th International Symposium on District Heating and Cooling, DHC2021, 2021.
Keywords: Control strategy; District heating; Solar collectors; Solar community; Thermal energy storage
1. Introduction
District energy systems (DES) represent an untapped opportunity for integrating renewable energy-based systems
and thermal energy storage (TES) devices at a large scale, paving the way for 4th and 5th generation district heating
∗Corresponding author.
E-mail address: etienne.saloux@canada.ca (E. Saloux).
https://doi.org/10.1016/j.egyr.2021.08.092
2352-4847/Crown Copyright c
⃝2021 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the scientific committee of the 17th International Symposium on District Heating and Cooling, DHC2021,
2021.

E. Saloux and J.A. Candanedo Energy Reports 7 (2021) 389–400
and cooling systems [1,2]. Recent research has focused on the modelling and design of these systems. Olsthoorn
et al. investigated the modelling and optimization of district heating systems involving storage and renewable
energy [3]. A new evaluation tool was proposed by Neyer et al. [4] to assess technical and economic aspects
of solar heating and cooling systems. Van der Heijde et al. presented a new technique using representative days to
reduce the computational time when optimizing both design and control of a fictional DES [5].
Although their potential for applications such as peak shaving, peak shifting and flexibility is known [6,7], there
is still a critical need for advanced controls to further optimize the performance of existing DES. This is particularly
the case when it comes to controlling advanced systems using solar thermal collectors, as well as short-term and
seasonal TES devices. Nonetheless, advanced controls face many challenges such as the lack of high-resolution data,
the dynamics of large and complex networks, and the management of intermittent renewable energy sources [8].
1.1. The Drake Landing Solar Community
The Drake Landing Solar Community (DLSC), located in Okotoks (Alberta, Canada), has been in operation since
2007. Shown in Fig. 1a, this district heating system uses mostly solar energy to satisfy the heating demand of a
52-home community [9]. Solar energy is harvested with 2293-m2solar collectors installed on the roof of detached
garages and is stored in a short-term thermal energy storage (STTS), composed of two 120-m3water tanks. From
the STTS, heat can be transferred to the district for direct use in winter or to a seasonal borehole thermal energy
storage (BTES), 144 boreholes 35-m deep, for future use later in the year. A natural gas boiler is available as a
back-up system.
The DLSC has consistently achieved high performance levels, as evidenced by its high solar fraction (i.e. the
fraction of heating demand covered by solar energy). After the first five years of operation, the solar fraction has
been consistently over 90%, even reaching 100% in 2015–16 [12].
In Alberta, 90% of electricity is generated from fossil fuels. Therefore, when compared to using natural gas
on-site, electricity use is more expensive (eight times) and generates more greenhouse gas (GHG) emissions (about
four times) [13]. In this context, the solar fraction provides a good estimate of the contribution of solar energy
to the heating demand; however, it omits the electricity consumed by the circulation pumps, whose economic and
environmental impact may be significant. Assessing total primary energy use could then complement the solar
fraction [13].
1.2. Previous work on control optimization
Previous work was done to optimize the controls in the DLSC district heating system. It aimed to optimize energy
exchanges in the system to minimize primary energy use while maintaining the same solar fraction. Two strategies
were investigated: (a) an optimal rule-based control (ORC) and (b) a model-based predictive control strategy (MPC).
The ORC strategy was first developed to optimize the use of the rules governing the operation of the variable
speed pumps [13]. Simulation results for 2017–18 showed that, in comparison with conventional control, ORC
could reduce electricity use by 43% while maintaining a similar solar fraction; energy costs and GHG emissions
were reduced by 34% and 29%, respectively [13].
A model-based predictive control (MPC) strategy was then developed, in which weather forecasts were used to
estimate the system performance 48 h ahead and adjust the pump operation accordingly [11]. The simulation study
was found that MPC can further improve the performance; 5%–6% savings compared to ORC [11].
Although MPC may lead to better performance, it is far more challenging to implement on-site (more
computational time-consuming, need for bi-directional communication with the control system to retrieve data and
send commands). However, ORC could still provide significant energy savings while being easily incorporated into
the control system.
In this context, the estimation and sensitivity of ORC parameters is worthy of investigation. In order to evaluate
the sensitivity of the calibration dataset on the ORC parameters, the rules developed in Ref. [13] by using 2017–
18 data were applied to the years 2015–16 and 2016–17; electricity savings of 40% and 45% were obtained,
respectively, with similar solar fractions. In conclusion, the ORC parameters are still valid for different weather
datasets.
390

E. Saloux and J.A. Candanedo Energy Reports 7 (2021) 389–400
Fig. 1. Drake Landing Solar Community district heating system: (a) schematic [10] and (b) control-oriented model [11].
1.3. Objective of the present work: optimize the link between design and control
The existing STTS volume (240 m3) and the current control parameters (maximum allowed flow rate in the
BTES loop of 3.1 L/s) were used in the calculations described above. One step beyond is to perform a parametric
study to investigate the effect of storage sizing (STTS volume) and control (rate at which energy storage devices
are either charged or discharged).
Design and control are often treated in isolation. Such an approach leads to suboptimal results, since design and
control are inextricably linked: a control strategy must consider the design of the system, and vice versa. This is
even more critical when energy storage is used. This realization has motivated the authors to quantify the impact of
controls on the system performance and to focus on the close link between design and control. More specifically,
391

E. Saloux and J.A. Candanedo Energy Reports 7 (2021) 389–400
the present study aims at assessing the impact of STTS sizing and storage control on primary energy use and solar
fraction.
2. Modelling approach
2.1. Control-oriented model of the community
Modelling a solar community district heating system involving solar thermal collectors and TES devices is no
easy task; previous work tackled this aspect. A Machine Learning model was used to estimate the community heating
demand [14] and a novel nodal TES modelling method was used for the STTS [15]. The rest of the community
was mainly modelled using Resistance–Capacitance thermal networks [16]. A schematic is given in Fig. 1b.
2.2. Control strategy for system “as built” and alternative designs
The control strategy is based on the optimal rule-based control developed in Ref. [13] and shown in Fig. 2. It
aims to control (a) the rate at which the seasonal storage (BTES) is either charged or discharged, and (b) both the
rate and temperature at which solar collectors harvest solar energy.
Fig. 2. Optimal rule-based control for (a) the BTES loop flow rate and (b) the setpoint for temperature difference across solar collectors [13].
392

E. Saloux and J.A. Candanedo Energy Reports 7 (2021) 389–400
•BTES loop flow rate: this flow rate enables “fine-tuning” the rate of energy injection and extraction to/from
the seasonal storage; in general, low flow rates are preferred to reduce electricity use. In normal operation, the
flow rate was nearly constant at 2.6 L/s, although manual override allows higher flow rates (up to 3.1 L/s) on
cold days. The strategy presented in Ref. [13] adjusts the BTES loop flow rate as a function of the amount of
“Useful Energy” (UE) in the STTS. High flow rates are used in two cases: (a) when the STTS useful energy
is very low (the STTS must be charged rapidly to avoid natural gas boiler start-up) or (b) when the STTS
useful energy is very high (heat must be evacuated from the STTS to avoid excessively high temperatures in
the tanks). In between, the flow rate can be reduced or energy exchanges can be avoided. The BTES loop flow
rate was limited to 3.1 L/s.
•Temperature difference (DT) across solar collectors: this DT is controlled via pump speed modulation
and affects both the flow rate and the temperature at the solar collectors’ outlet. Although increasing this
temperature difference reduces the total amount of solar energy harvested, it makes it possible to deliver heat
to the STTS at higher temperatures. In normal operation, the DT setpoint is switched seasonally, between
15 ◦C (in winter) and 25 ◦C (in summer). In contrast, the strategy presented in Ref. [13] proposed to adjust
the DT on a daily basis as a function of STTS useful energy (3-hr average value). The DT is low (i.e. high
flow rate) when the STTS useful energy is high in order to avoid excessive temperatures in the solar loop; the
DT is increased (i.e. low flow rate) when the STTS useful energy is low, since the system requires thermal
energy at higher temperatures. It was found that a constant setpoint DT of 30 ◦C was optimum (in both winter
and summer), enabling a significantly reduced pump electricity consumption.
These control strategies were adapted to each of the alternative designs in this study. The control rule parameters
shown in Fig. 2 for BTES loop flow rates and the DT across solar collectors were optimized based on different STTS
volumes and different maximum values for the BTES loop flow rate. The upper limit of 30 ◦C for the DT across
solar collectors remained unchanged as it would be difficult to model higher temperature differences accurately.
2.3. Initial conditions
The control-oriented model needs to be initialized at the beginning of the simulation. The system state is
mainly determined by temperatures at key locations of the system: solar collectors’ absorber temperature, glycol
temperatures at solar collectors’ inlet and outlet, supply and return water temperatures in the district loop, water
temperatures in the STTS and ground temperatures in the BTES. Initial conditions were associated to measured
data [11].
Similar initial conditions were considered for the sake of simplicity. The six water temperature readings inside
the STTS (top, middle and bottom of each tank) were used for all STTS volumes. The effect on the annual system
performance is expected be very minor, given that the STTS mainly affects the short-term performance. Moreover,
long-term effects on BTES charging in the first years of operation are beyond the scope of this paper.
2.4. Optimization problem: selection of control parameters for alternative designs
The optimization problem targets the minimization of primary energy use of the district heating system on an
annual basis and includes the consumption of both the natural gas back-up boiler (Qboiler) and the circulation pumps
(Wpp). This optimization is carried out by selecting the optimal set of parameters shown in Fig. 2 for the rule-based
393

E. Saloux and J.A. Candanedo Energy Reports 7 (2021) 389–400
control strategies. The problem can be formulated as follows:
min
xJ,where J=(PEFgas ×Qboiler +PEFelec ×Wpp)
where x=[U Ev,1U Ev,2U Ev,3U Ev,4Vmax,ch Vmax,di s U EDT,1U EDT,2DTmax DTmin]T
s.t.
0<U Ev,1<U Ev,2<U Ev,3<U Ev,4
Vmin ≤Vmax,ch ≤max(VBTES)
Vmin ≤Vmax,dis ≤max(VBTES)
0<U EDT,1<U EDT,2
15 ◦C≤DTmin <DTmax ≤30 ◦C
Tout,s c ≤95 ◦C (98 ◦C if V O L ST T S =120 m3)
Tout,s,h x1≤88 ◦C(93 ◦C if V OL S T T S =120 m3)
(1)
where xrefers to control strategy parameters shown in Fig. 2. PEFgas and PEFelec are primary energy factors
associated to natural gas (1.11) and electricity (2.86). Vmin is the minimum flow rate to operate the pump (0.7
L/s) and max(VBTES) is the maximum allowed value for the BTES loop flow rate (2.5–4.5 L/s). VOLST T S is the
STTS volume. There are also constraints on glycol temperature at the solar collectors’ outlet (Tout,sc) and water
temperature at HX-1 outlet (Tout,s,hx1) to avoid boiling. UE is the STTS useful energy (in MWh), which depends
on outdoor air temperature, STTS volume and temperatures; the definition can be found in [13].
3. Methodology
3.1. Tested scenarios
The STTS volume and the maximum BTES loop flow rate were varied to evaluate the impact of both storage
sizing and control. In Ref. [13], the control parameters were optimized for the system “as built” (STTS volume
of 240 m3, maximum BTES loop flow rate of 3.1 L/s). In this work, the performance was estimated for these
alternative designs:
•STTS volume: 120 m3, 180 m3, 240 m3, 300 m3, 360 m3, 420 m3, 480 m3.
•Maximum BTES loop flow rate max(VBTES): 2.5 L/s, 3.0 L/s, 3.5 L/s, 4.0 L/s, 4.5 L/s.
Examining each of these factors independently allows to better understand their impact. For example, varying
the maximum BTES loop flow rate shows what can be done by adjusting only the control of the charge/discharge
rate. Conversely, varying the STTS volume shows what would be the impact of either decreasing the storage size
to reduce initial capital costs or increasing it to achieve higher performance levels. The specific extreme case of a
small storage volume 120 m3also shows what would happen in case of a failure in the one of the two STTS tanks.
3.2. Simulations and optimization routine
The control-oriented model (Section 2.1) was used to perform calculations at 10-min intervals based on the
data for July 2016 to June 2017. Actual weather conditions were extracted for the whole year. The system state
initialization was done based on measurements for July 1st, 00:00 (Section 2.3).
In each scenario (i.e. a combination of a STTS volume with a maximum BTES loop flow rate), the control strategy
parameters for both the BTES loop flow rate and the setpoint for temperature difference across solar collectors
(Section 2.2) were optimized to minimize the objective function on an annual basis (Section 2.4).
The optimization routine, programmed in MATLAB, consists of a genetic algorithm to find an estimate of the
global minimum, and a local non-linear minimization routine to refine the calculation [13]. Since this is non-linear
minimization problem, there might be many local minima and convergence may not be straightforward. Therefore,
multiple optimizations were performed for each combination to ensure that a consistent optimum was obtained.
394

E. Saloux and J.A. Candanedo Energy Reports 7 (2021) 389–400
4. Results
4.1. Primary energy use
The system performance was evaluated in terms of primary energy use (Fig. 3). It is worth mentioning that low
flow rates (2.5–3.0 L/s) for low STTS volumes (120–180 m3) did not respect the constraints on temperatures either
in the solar loop or in the STTS. In other words, low flow rates prevent solar energy harvested from being evacuated
rapidly enough from the STTS to the seasonal storage. Small storage volumes require high flow rates, not only to
improve the performance of the system, but also to avoid technical and safety issues (boiling).
The effectiveness of the control strategy can be evaluated by comparing the reference (primary energy use of
99.9 MWh) with the 240-m3scenario. Even with flow rates below the reference case (2.5–3.0 L/s vs. 3.1 L/s), the
proposed control strategy reduced primary energy use by 12.7–23.3%. This number can reach 30.2% when the flow
rate is allowed to go up to 4.5 L/s.
Moreover, in a context of failure analysis where one of the tanks becomes defective, a 120-m3STTS volume
with appropriate control strategy parameters could still achieve savings up to 6.1%, in comparison to the reference
scenario. This can be obtained by increasing the pump flow rate to better transfer solar energy to the BTES in the
summer.
Overall, increasing the BTES loop flow rate seems effective to reduce primary energy use; a plateau is reached
after 3.5 L/s. Similarly, increasing the storage volume is also effective; and the performance levels off after 300
m3. In terms of solar fraction (not shown in Fig. 3), increasing both the BTES flow rate and the STTS volume
enhances the ability to provide heat during peak heating loads, thus reducing gas consumption and increasing solar
fraction. However, this effect is rather modest, and for all scenarios, the change in solar fraction was found to be
within ±2% with respect to the reference scenario.
4.2. Optimal parameters for flow rate control
As in Ref. [13], the optimal setpoint for temperature difference across solar collectors was found to be constant
to 30 ◦C, which means that solar energy should be harvested at high temperature as it allows to significantly reduce
solar loop and HX-1 loop pump electricity consumption (Section 4.3).
For the BTES loop flow rate, Fig. 4 shows the system performance (primary energy use, natural gas consumption,
electricity use) and the BTES loop flow rate profile for 120-m3and 240-m3STTS volumes. First, compared to the
reference scenario, the optimization tends to decrease electricity use in priority, which is due to the higher primary
energy factor associated to electricity (Section 2.4), although natural gas consumption is increased. Second, when
the maximum BTES flow rate is increased, the system operates at full flow only when needed (STTS almost empty
or full) in order to decrease gas consumption during peak periods, despite a slightly higher electricity consumption.
Third, higher STTS volume provides enough storage material to allow smoother transitions from maximum flow
rates to zeros, enhancing the effects of increasing maximum BTES loop flow rate.
Finally, control strategy parameters for 240 m3and 3.0 L/s (in orange in Fig. 4d) were compared to those
obtained in Ref. [13] for a different year (2017–18) with a slightly different pump curve and a maximum BTES
loop flow rate of 3.1 L/s. Values were found to be similar for change points in winter (UEV,1, UEV,2in Fig. 2) and
summer (UEV,3, UEV,4in Fig. 2), and maximum flow rates. It indicates that these profiles are not too sensitive to
the calibration data.
4.3. Energy exchanges
Table 1 shows the main energy exchanges within the system for several STTS volumes. From this table, increasing
the STTS volume allows to: (a) slightly increase the harvested solar energy by avoiding critical temperatures in both
the solar loop and the STTS and (b) reduce energy exchanges between the STTS and the BTES, which ultimately
decrease the BTES loop pump consumption and to some extent that of solar and HX-1 loops.
395

E. Saloux and J.A. Candanedo Energy Reports 7 (2021) 389–400
Fig. 3. (a) Primary energy use and (b) primary energy use reduction for different maximum BTES loop flow rates and STTS volumes. The
reference case primary energy use is 99.9 MWh.
396

E. Saloux and J.A. Candanedo Energy Reports 7 (2021) 389–400
Fig. 4. System performance (left) and BTES loop flow rate profiles (right) for STTS volume of (a)–(b) 120-m3and (c)–(d) 240-m3.
4.4. STTS temperature distribution
Fig. 5 shows the temperature distribution inside the STTS on hot days, a key consideration in the context of
failure analysis. Results show that smaller STTS volumes cause higher temperature fluctuations inside the tank
and eventually higher temperature levels. Such conditions decrease the efficiency of the solar collectors and might
397

E. Saloux and J.A. Candanedo Energy Reports 7 (2021) 389–400
Table 1. Energy exchanges in the district heating system for different STTS volumes and maximum BTES loop flow rates.
STTS volume/Maximum BTES flow rate Reference 120 m3/4.5 L/s 240 m3/4.0 L/s 360 m3/4.0 L/s
Solar energy harvested and transferred through HX-1 1138 MWh 1050 MWh 1083 MWh 1114 MWh
Energy injected to the BTES 706 MWh 683 MWh 650 MWh 618 MWh
Energy extracted from the BTES 317 MWh 353 MWh 322 MWh 319 MWh
Solar & HX-1 loop pump consumption 12.6 MWh 6.3 MWh 5.6 MWh 5.5 MWh
BTES loop pump consumption 8.3 MWh 8.4 MWh 6.3 MWh 4.9 MWh
Fig. 5. STTS temperature distribution for a STTS volume and maximum BTES loop flow rate of: (a) 120 m3, 4.5 L/s, (b) 240 m3, 4.0 L/s.
increase the contribution of the fluid coolers, whose role is to evacuate harvested solar energy to the ambient air to
avoid boiling.
5. Conclusions
This paper has investigated the impact of storage control and storage sizing to improve the performance of a
solar district heating system. A rule-based control strategy was optimized to minimize the primary energy use and
398

E. Saloux and J.A. Candanedo Energy Reports 7 (2021) 389–400
targeted the energy exchanges between the seasonal and the short-term thermal energy storage devices. Both the
short-term storage (STTS) volume and the rate of energy injection/extraction from the seasonal storage (BTES)
were investigated.
Results showed that an appropriate control strategy could achieve significant primary energy savings (13%–
30%) for a similar solar fraction with BTES loop flow rates of 2.5–4.5 L/s compared to the reference scenario. By
decreasing the short-term storage volume from 240 m3to 120 m3, which could be representative of a system failure,
up to 6% energy savings could still be obtained by adjusting control strategy parameters. However, temperatures
in the STTS would be higher and are more likely to cause technical issues. In turn, increasing the STTS volume
would help store more thermal energy and be more efficient during peak periods, resulting in further improved
performance (up to 36%). Nonetheless, the performance tends to level off with STTS volumes higher than 300 m3
and maximum BTES loop flow rates higher than 3.5 L/s. The final selection of a control strategy and a control
volume size would also require a cost–benefit analysis, and other considerations such as robustness and reliability.
These results are representative of the particular context in Alberta, which has high solar radiation levels and
extreme seasonal temperature swings, while electricity is mainly generated from fossil fuels. As the electric grid
becomes gradually decarbonized – or in other contexts (e.g. less sunny conditions, or an already decarbonized grid)
– similar control strategies are expected to be effective, since they allow to properly manage thermal energy storage
devices according to the heating needs; only the control strategy parameters will need to be adjusted accordingly.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could
have appeared to influence the work reported in this paper.
Acknowledgements
The authors gratefully acknowledge the financial support of Natural Resources Canada through the Office of En-
ergy Research and Development, Canada. The authors would like to thank their colleagues from CanmetENERGY,
Lucio Mesquita, Doug McClenahan and Bill Wong, for sharing their experience of Drake Landing and providing
data.
References
[1] Lund H, Østergaard PA, Chang M, Werner S, Svendsen S, Sorknæs P, et al. The status of 4th generation district heating: Research
and results. Energy 2018;164:147–59.
[2] Buffa S, Cozzini M, D’Antoni M, Baratieri M, Fedrizzi R. 5th generation district heating and cooling systems: A review of existing
cases in Europe. Renew Sustain Energy Rev 2019;104:504–22.
[3] Olsthoorn D, Haghighat F, Mirzaei PA. Integration of storage and renewable energy into district heating systems: A review of modelling
and optimization. Sol Energy 2016;136:49–64.
[4] Neyer D, Ostheimer M, Dipasquale C, Köll R. Technical and economic assessment of solar heating and cooling – Methodology and
examples of IEA SHC Task 53. Sol Energy 2018;172:90–101.
[5] van der Heijde B, Vandermeulen A, Salenbien R, Helsen L. Representative days selection for district energy system optimisation: a
solar district heating system with seasonal storage. Appl Energy 2019;248:79–94.
[6] Gravelsins A, Pakere I, Tukulis A, Blumberga D. Solar power in district heating. P2H flexibility concept. Energy 2019;181:1023–35.
[7] Van Oevelen T, Vanhoudt D, Johansson C, Smulders E. Testing and performance evaluation of the STORM controller in two
demonstration sites. Energy 2020;197:117177.
[8] Vandermeulen A, van der Heijde B, Helsen L. Controlling district heating and cooling networks to unlock flexibility: A review. Energy
2018;151:103–15.
[9] Sibbitt B, McClenahan D, Djebbar R, Thornton J, Wong B, Carriere J, et al. The Performance of a High Solar Fraction Seasonal
Storage District Heating System – Five Years of Operation. Energy Procedia 2012;30:856–65.
[10] Quintana HJ, Kummert M. Optimized control strategies for solar district heating systems. J Build Perform Simul 2015;8:79–96.
[11] Saloux E, Candanedo JA. Model-based predictive control to minimize primary energy use in a solar district heating system with
seasonal thermal energy storage. Appl Energy 2021;291:116840.
[12] Mesquita L, McClenahan D, Thornton J, Carriere J, Wong B. Drake Landing Solar Community: 10 Years of Operation. In: ISES solar
world congress, IEA SHC international conference on solar heating and cooling for buildings and industry. 2017, Abu Dhabi, UAE,
Oct 29–Nov 2.
[13] Saloux E, Candanedo JA. Optimal rule-based control for the management of thermal energy storage in a Canadian solar district heating
system. Sol Energy 2020;207:1191–201.
[14] Saloux E, Candanedo JA. Forecasting District Heating Demand using Machine Learning Algorithms. Energy Procedia 2018;149:59–68.
399

E. Saloux and J.A. Candanedo Energy Reports 7 (2021) 389–400
[15] Saloux E, Candanedo JA. Modelling stratified thermal energy storage tanks using an advanced flowrate distribution of the received
flow. Appl Energy 2019;241:34–45.
[16] Saloux E, Candanedo JA. Control-oriented model of a solar community with seasonal thermal energy storage: development, calibration
and validation. J Build Perform Simul 2018;12:523–45.
400