Bio-based CHP as efficient and profitable technology for balancing the energy system

SAFETY AND TRANSPORT

ELECTRIC POWER

SYSTEMS

Bio-based CHP as efficient and profitable technology for

balancing the energy system

Camille Hamon

Amin Nasri

Susanne Paulrud

Bio-based CHP as efficient and profitable technology for

balancing the energy system

Camille Hamon

Amin Nasri

Susanne Paulrud

RISE Research Institutes of Sweden AB RISE Report 2021:24

ISBN: ISBN 978-91-89385-09-2 Stockholm 2021

Content

2.1 Background on electricity markets ... 6

2.1.1 Day-ahead market ... 6

2.1.2 Balancing market ... 7

2.1.3 Principles for profitability in up-regulation ... 8

2.1.4 Principles for profitability in down-regulation ... 9

2.1.5 Overall profits from electricity markets ... 10

2.2 District heating systems ... 10

2.2.1 General description of district heating systems ... 10

2.2.2 Operations of CHP plants ... 11

2.3 Participation of CHP plants in balancing ... 12

2.3.1 Strategy “P down only” ... 14

2.3.2 Strategy “P down and Q down” ... 14

2.3.3 Strategy “P down and Q up” ... 15

2.4 Overall methodology ... 16

2.5 Case studies... 21

2.5.1 Borås Energi & Miljö ... 22

2.5.2 Vattenfall ... 22

3 Results ... 23

3.1 Case study 1: Value of participating in the balancing market (Borås) ... 23

3.2 Case study 2: Value of participating in the balancing market (Nyköping)... 26

3.2.1 Impact of balancing market participation on profits ... 26

3.2.2 Analysis of balancing market participation ... 27

3.3 Case study 3: Value of increasing heat storage usage for participation on the balancing market (Nyköping) ... 33

3.4 Case study 4: Comparison of investment in CHP vs heat-only plants (Borås) 35 4 Conclusions ... 42

Preface

The project Bio-based CHP as efficient and profitable technology for balancing the energy system has been coordinated by RISE and the work is a cooperation between RISE, Vattenfall and Borås Energi och Miljö. The project has been financed by Energy Agency through the program BIOKRAFT.

We would like to thank participants and financiers who all contributed to the implementation of the project.

Summary

The project’s objective is to assess whether flexible bio-based power generation can be an effective and profitable technology option for balancing power system with large amounts of weather dependent production (wind and sun). It includes an analysis which estimates the potential income levels that can be gained by different types of bio-based power plants from selling electricity products and determines if the extra cost of having a combined heat and power (CHP) plant instead of having a heat only (HO) plant is worth investing.

The project's subgoals, how they have been addressed and some follow-up directions are presented below:

• Subgoal 1: Increased knowledge about revenues that can be gained by different technology solutions from selling electricity in different markets, e.g., day-ahead and balancing markets (also called manual frequency restoration reserve market – mFRR market).

A model has been developed to estimate the profits that can be made by district heating owners on day-ahead and balancing (mFRR) markets. The model has been applied in two case studies using historical price data for 2019: one using Borås Energi och Miljö’s (BEM) district heating system and one using Vattenfall’s system in Nyköping. In both cases, the results show an increase in profits when participating in balancing markets. This increase is, however, very small compared to the profits made by only participating in day-ahead markets (less than 0.1% increase in overall profits).

• Subgoal 2: Increased knowledge about investment and operating costs (CAPEX/OPEX) of different bio-based technologies, both HO and CHP plants. A literature survey was done to identify comprehensive sources of costs. Two reports were identified as relevant sources: [1] and [2]. The first one dates from 2014 and the second one from 2019. They provide a wide-ranging collection of investment and operation and maintenance costs for units of different sizes and types.

• Subgoal 3: Increased knowledge about the market conditions needed to justify the extra cost of building a bio-based CHP plant instead of a bio-based HO plant. Two investment case studies have been performed using BEM’s system to compare investment in CHP and HO units. The previous model has been used to evaluate the profits from participating in the day-ahead electricity market for three representative years: 2019, 2030 and 2040. Price conditions for 2030 and 2040 were obtained from Svenska kraftnät’s long-term market analysis [3]. The investment and O&M costs of the new units were obtained from [1] and [2]. The first investment study investigates an investment in new units for redundancy purposes (to ensure sufficient production capacity in case the largest unit is unexpectedly out of operation). The investment results show that profits from selling heat and electricity are not enough to cover the investment and O&M costs. The economic losses are smaller when going for the HO option. The second investment study investigates an investment in base-load units (units meant to run most of the time). In this case, both CHP and HO plants are profitable but the HO plant achieve a higher profitability.

• Subgoal 4: Increased knowledge about the financial risks and uncertainties in such an investment.

This was achieved through discussing the results with the industry partners and identifying aspects that could be further investigated in a follow-up project. The lifetime of new investment is around 25 years. Many sources of uncertainties enter the evaluation of the financial indicators. Among these, the most relevant ones that could be further investigated are: assumptions on the underlying heat demand (colder and warmer years), assumptions on future electricity prices, assumptions on the price for green certificates and purpose of an investment (an investment for redundancy purposes may not recover its costs but is still needed – how to value redundancy?). Many other input parameters to the investment studies - such as O&M costs, fuel prices and discount rate - may also play a role in the results. A thorough sensitivity analysis would shed more light on the magnitude of each one and could be included in a follow-up project.

1 Introduction

The Paris agreement [4], the European Green Deal [5] and the national Swedish climate targets [6] call for an ambitious energy transition in the coming decades. Sweden is to have zero net emissions 2045 at the latest and 100 % renewable electricity production by 2040 [7]. Electrification of industries and transport will play an essential role at achieving the climate targets. For Sweden alone domestic electricity use may increase from today’s 140 TWh to between 180 and 220 TWh in 2040 [3].

Biofuel-based combined heat-and-power plants (CHP plants) in district heating systems can play key roles in meeting the challenges associated with the energy transition, such as alleviating local electric grid capacity constraints, see [8], [9] and [10], and help balance power systems with larger amounts of variable generation [11], [12].

For biofuel-based CHP plants to be able to fulfil these roles, there is a need to investigate the profitability of CHP plant investments as compared to heat only plants (HO) and to evaluate their potential contribution in balancing markets from technical and economical perspectives.

1.1 Project aims

The project’s objective was to assess whether flexible bio-based power generation can be an effective and profitable technology option for balancing power system with large amounts of variable generation (wind and sun). It also includes an analysis which estimates the potential income levels that can be gained by different types of bio-based power plants from selling electricity products and determines if the extra cost of having a combined heat and power (CHP) plant instead of having a heat only (HO) plant is worth investing.

The project's sub-goals were:

• Increased knowledge about revenues that can be gained by different technology solutions from selling electricity in different markets, e.g., day-ahead and balancing (mFRR) markets.

• Increased knowledge about investment and operating costs (CAPEX/OPEX) of different bio-based technologies, both HO and CHP plants.

• Increased knowledge about the market conditions needed to justify the extra cost of building a bio-based CHP plant instead of a bio-based HO plant.

• Increased knowledge about the financial risks and uncertainties in such an investment.

1.2 Scope

The project focused on district heating systems. Other kinds of CHP owners have not been included in the analysis. The project has performed analyses from the perspective of district heating companies today to better understand the profitability of investments

in CHP plants and participation in the balancing market. Analyses on the system level are not in the scope of this project.

1.3 Content

The report is structured as follows. Section 2 presents the methodology and models that have been developed within the scope of this project in order to address the project goals. It also introduces the case studies performed in this project and presents two real-life district heating systems that have been used in the case studies. Section 3 presents the results from the case studies.

2 Methodology

2.1 Background on electricity markets

In this project, participation of CHP plants in the day-ahead electricity market and in the balancing market was studied. “Balancing market” refers to the market for manual frequency restoration reserves (mFRR). Figure 1 illustrates the timings for these two markets. The following sections give more detail about these two markets and how producers can participate in these markets.

Figure 1: Timings for the day-ahead and balancing markets.

To illustrate how markets work and when it is profitable for a producer to participate, we will take the example of a producer with one production unit. We will assume that this unit can produce between Pmin=50 MW and Pmax=125 MW and that the marginal production cost is constant equal to 250 SEK/MWh.

2.1.1 Day-ahead market

Market participants submit production or consumption bids to the day-ahead market. Bids indicate a volume of energy in MWh/h and a price for production or consumption. The transmission system operators (TSOs) submit trading capacities between bidding zones. For each hour of day D, the day-ahead market coupling algorithm matches sell and buy bids to maximize market welfare while ensuring that the cross-border flows between bidding zones remain under the capacities determined by the TSOs. The results of the day-ahead market are published at 12.45 on day D-1. For each hour of day D, the results consist of day-ahead electricity prices for each bidding zone, and of accepted

production and consumption bids. Therefore, at 12.45 on day D-1, market participants have knowledge which bids of theirs were accepted for each hour of day D. These bids are binding, which means that market participants must ensure to produce or consume the volume of energy in their accepted bids. These commitments are referred to as day-ahead production plans.

Note that the market participants can trade on the intraday continuous intraday market up to one hour before hour H to adjust their production plans but, for the sake of simplicity, this is not considered here.

It is economically profitable for producers to participate in the day-ahead market if they can receive an income that is higher than their marginal production cost. Thus, for the producer in the example from the previous section, the following bid can be placed on the day-ahead market:

- Volume: Pmax MWh/h

- Price: 251 MWh (or any price above 250 SEK/MWh, with 250 MWh/SEK MWh being the marginal production cost for the producer)

Note that under the assumption that producers are price takers and not price makers, there is no incentive for the producers to give a bid price much higher than their marginal production cost since that would increase the risk of not being activated, and an accepted bid receives the day-ahead price for bidding area and not the bid price.

To illustrate how balancing markets work in the following subsection, we will assume that the producer got a bid with a volume of 100 MWh/h accepted on the day-ahead market. Let us assume that the day-ahead price was λda_{=300 SEK/MWh.}

The profit from the day-ahead market before any action on the balancing market is then 100 MWh * (300 – 250 SEK/MWh) = 5000 SEK, which is the difference between income (100 MWh * 300 SEK/MWh) and production costs (100 MWh * 250 SEK/MWh). Note that the profit from the day-ahead market does not depend on the day-ahead bid price. The day-ahead bid price does, however, impact whether a bid gets accepted or not since the bids are activated following a merit order to maximize market welfare.

2.1.2 Balancing market

Market participants can submit balancing bids to the mFRR market for up-regulation (increase of electricity production) or down-regulation (decrease of electricity production) up to 45 minutes before hour H. Balancing bids submitted separately for up- and down-regulation. The bids are made up of a volume in MW, a price in SEK/MWh and an activation time (which must be less than 15 minutes). During the hour of delivery H, unexpected events and deviations from the production and consumption plans can lead to system frequency deviations from 50 Hz (50 Hz is the nominal frequency in the Nordics, i.e. the frequency when production and consumption are in balance). TSOs activate mFRR bids from the balancing market in order to restore the frequency to its nominal value 50 Hz when necessary. Bids are activated in price order. TSOs will activate up-regulation bids to increase the system frequency and down-regulation bids to decrease the system frequency. When a balancing bid is activated, the market participant that submitted the bid must change its electricity production by the bid volume within the activation time. The activated bids may be “released” by the TSOs before the end of

hour H, in which case the market participant must return to its electricity production level before the bid was activated. Market participants must be able to sustain the activated bids until the end of the hour if necessary.

After hour H, the system up- and down-regulation prices are computed considering all balancing actions during this hour. The system up-regulation price is set to the most expensive up-regulation bid that was accepted. The system down-regulation is set to the cheapest down-regulation bid that was accepted. The market participants whose balancing bids were activated will receive the system up-regulation price (not their bid’s up-regulation price) or pay the system regulation price (not their bid’s down-regulation price). This scheme ensures that the market participants receive at least their bid’s up-regulation price or pay at most their bid’s down-regulation price.

In the following subsections, more detail about up- and down-regulation is given. Situations in which it is economically profitable for a producer to participate in up- and down-regulation are described.

2.1.3 Principles for profitability in up-regulation

An up-regulation bid means that TSOs buy energy from market participants. In practice, this means market participants increase their production from their day-ahead production plan. The market participants get paid to produce more energy than planned. Market participants will participate in up-regulation only if the income they receive from the balancing market covers at least the production cost associated with increasing their electricity production.

As mentioned in the previous section, the producer will actually receive the system up-regulation price, which corresponds to the most expensive up-up-regulation bid that was activated. The profit made from participating in up-regulation is equal to the activated volume times the difference between the system up-regulation price and the producers’ marginal production cost.

Going back to the example initiated in the previous sections, it is interesting for the producer to participate in up-regulation during the considered hour if the income it gets is larger than the cost of increasing its production from its day-ahead production plan. Therefore, the producer can decide to submit an up-regulation bid to the balancing market for a price slightly higher than its marginal production cost. The capacity available for up-regulation is the difference between the maximal capacity Pmax=125 MW of the unit and its day-ahead production plan. In summary, the producer can decide to submit the following up-regulation bid:

- Volume: 25 MW (difference between Pmax=125 MW and the day-ahead production plant of 100 MWh)

- Price: 251 MWh (or any price above 250 SEK/MWh, with 250 MWh/SEK MWh being the marginal production cost for the producer).

Note that when setting the bid price, the producer can set any price above its marginal production cost to make profit from participating in up-regulation. Under the assumption that producers are price takers and not price makers, there is no incentive for producers to set an up-regulation price that is much higher than their marginal production cost since they receive the system up-regulation price.

Continuing our example, let λup_{=350 SEK/MWh be the system up-regulation price.} Suppose that the above bid was activated. Then the producer makes a profit of 25*(350-250) = 2500 SEK from participating in up-regulation. Note that the bid’s up-regulation price does not impact the profit. It does however impact whether the bid gets accepted since

2.1.4 Principles for profitability in down-regulation

A down-regulation bid means that TSOs sell energy to the market participants. In practice, this means market participants reduce their production from their day-ahead production plan. The market participants pay TSO to producer less energy than planned (they “buy back” energy from their day-ahead plan). Reducing their energy production means that producers can save production costs. Therefore, producers will participate in down-regulation only if the cost to buy back energy is lower than the avoided costs. from the balancing market covers at least the production cost associated with increasing their electricity production.

As mentioned in the previous section, the producer will actually pay the system down-regulation price, which corresponds to the cheapest down-down-regulation bid that was activated. The profit made from participating in down-regulation is equal to the activated volume times the difference between the avoided production costs and the system down-regulation price.

In our example, it is interesting for the producer to participate in down-regulation during the considered hour if production cost savings are larger than the cost of buying back its production from its day-ahead production plan. Therefore, the producer can decide to submit a down-regulation bid to the balancing market for a price slightly lower than its marginal production cost. That is, it is ready to buy back energy if this costs less than producing this energy. The capacity available for down-regulation is the difference between its day-ahead production plan and the minimum capacity Pmin of the unit and. In summary, the producer can decide to submit the following down-regulation bid:

- Volume: 50 MW (difference between the day-ahead production plan of 100 MWh and the minimum production level Pmin=50 MW)

- Price: 249 MWh (or any price below 250 SEK/MWh, with 250 MWh/SEK MWh being the marginal production cost for the producer).

Note that when setting the bid price, the producer can set any price below its marginal production cost to make profit from participating in down-regulation. Under the assumption that producers are price takers and not price makers, there is no incentive for producers to set a down-regulation price that is much lower than their marginal production cost since they receive the system down-regulation price.

Continuing our example, let λdn_{=150 SEK/MWh be the system down-regulation price.} Suppose that the above bid was activated. Then the producer makes a profit of 50*(250-150) = 5000 SEK from participating in regulation. Note that the bid’s down-regulation price does not impact the profit. It does however impact whether the bid gets accepted since TSOs activate bids in price order.

2.1.5 Overall profits from electricity markets

In this section, the formulation for the overall profits that producers make by participating in electricity markets is given. The formulation considers a time horizon of N days. In the following, we consider one producer and express the profit that this producer makes.

The profit from participating in the day-ahead market: - Profit(DA) = ∑𝑁𝑑=1∑24ℎ=1(𝜆𝑑,ℎ𝑑𝑎 − 𝐶𝑑,ℎ𝑑𝑎)𝑃𝑑,ℎ𝑑𝑎

Where:

- 𝜆𝑑,ℎ𝑑𝑎 is the day-ahead price for day d and hour h.

- 𝑃_𝑑,ℎ𝑑𝑎 _{is the day-ahead electricity production plan (i.e. volume of accepted bids).}

- 𝐶_𝑑,ℎ𝑑𝑎 _{is the production cost corresponding to 𝑃} 𝑑,ℎ𝑑𝑎.

The profit from participating in the balancing market is

- Profit(balancing) = ∑𝑁𝑑=1∑24ℎ=1[(𝜆_𝑑,ℎ𝑢𝑝 − 𝐶_𝑑,ℎ𝑢𝑝)𝑃_𝑑,ℎ𝑢𝑝+ (𝐶𝑑,ℎ𝑑𝑛− 𝜆𝑑𝑛𝑑,ℎ)𝑃𝑑,ℎ𝑑𝑛]

Where:

- 𝜆_𝑑,ℎ𝑢𝑝 is the system up-regulation price for day d and hour h.

- 𝑃_𝑑,ℎ𝑢𝑝 is the increase in production for regulation (i.e. volume of accepted up-regulation bids).

- 𝐶_𝑑,ℎ𝑢𝑝 is the production cost corresponding to 𝑃_𝑑,ℎ𝑢𝑝.

- 𝑃_𝑑,ℎ𝑑𝑛 _{is the decrease in production for down-regulation (i.e. volume of accepted}

down-regulation bids).

- 𝐶_𝑑,ℎ𝑑𝑛 _{is the avoided production cost corresponding to 𝑃} 𝑑,ℎ𝑑𝑛.

Note that the accepted volumes on the day-ahead and balancing markets are always such that

- 0 ≤ 𝑃_𝑑,ℎ𝑑𝑛_{≤ 𝑃}

𝑑,ℎ𝑑𝑎− 𝑃𝑚𝑖𝑛

- 0 ≤ 𝑃_𝑑,ℎ𝑢𝑝

≤ 𝑃𝑚𝑎𝑥− 𝑃𝑑,ℎ𝑑𝑎

These formulations are used in the developed modules presented in the Section 2.4.

2.2 District heating systems

2.2.1 General description of district heating systems

For the purpose of this report, a district heating system is considered to be a portfolio that can include heat production units, heat storage (such as accumulators) and coolers (where excess heat can be dumped), all operated together to deliver heat to a district heating network. Heat production units can be heat-only boilers (including electric boilers), heat pumps or combined heat-and-power (CHP) plants. Figure 2 illustrates what a district heating system can look like.

Figure 2: Illustration of a district heating system, from [13].

District heating systems are planned and operated to deliver the forecast heat demand at all times while minimizing the production costs. If CHP plants are included in the system, the units will be planned in such a way to also maximize revenues from the electricity market. CHP plants participate mainly in the day-ahead electricity market today.

Heat storage provides district heating systems with flexibility. Operational practices include, for example, filling up the storage before very cold periods to have as much stored heat as possible if needed (for example if a unit would fall out of operation unexpectedly).

2.2.2 Operations of CHP plants

Combined heat-and-power (CHP) plants can produce both heat and electricity as illustrated in Figure 3. Usually, steam is generated by heating up water in a fuel-fired boiler. This steam then passes through a turbine that can drive an electric generator. The steam leaving the turbine can then be processed to extract heat. In some plants, the turbine can be bypassed for operation in heat-only (HO) mode. There is a co-dependence between heat and electricity production. This co-dependence is plant specific.

Figure 3: Schematics of a CHP plant, from [14].

Figure 4 illustrates this co-dependence for a fictitious CHP plant. The light blue triangle area is the heat versus electricity production operating region when the plant operates in

CHP mode. The dark blue line at the bottom is the heat production operation region when the plant operates in HO mode. These operating regions will be referred to as the PQ operating region in what follows (P usually denoted electricity generation and Q heat generation). In the figure, point 1 corresponds to an operating point of the plant in CHP mode and point 2 to an operating point in HO mode. Point 3 is not a valid operating point and the plant can never operate there.

Figure 4:So-called PQ operating region showing the co-dependence between heat and electricity production. The dots show three operating points for the plant: point 1 (black) and 2 (grey) are valid operating points in CHP and HO mode, respectively; point 3 (red) is not a valid operating point since it is outside the CHP and HO mode operating regions.

CHP plants enable district heating companies to participate in electricity and balancing markets. Today, district heating companies typically participate on the day-ahead (and sometimes intraday) electricity markets but not on the balancing market.

2.3 Participation of CHP plants in balancing

To illustrate the participation of CHP plants in balancing, down-regulation will be considered. The same principles apply for up-regulation for which electricity production is increased instead of decreased.

There are three strategies with which CHP plants can participate in down-regulation on the balancing electricity markets, depending on whether heat production is changed or not. As discussed in Section 2.2.1, the electricity and heat production in CHP plants are interdependent. The three strategies are illustrated in the PQ operating region in Figure 5, Figure 6 and Figure 7:

- “P down only” strategy: Decrease in electricity production only.

- “P down and Q down” strategy: Decrease in both electricity and heat production. - “P down and Q up” strategy: Decrease in electricity production and increase in heat production. Note that this strategy may appear counterintuitive because increasing heat production increases fuel costs. The economic profitability of this strategy is discussed further down.

Figure 5: “P down only” - decrease in electricity production only.

Figure 7: "P down and Q up": decrease in electricity production and increase in heat production.

The three strategies are explained in more detail below. It is important to recall from Section 2.1.4 that the profitability of participating in down-regulation is proportional to the difference between the avoided production costs (avoided fuel costs) and the system down-regulation price.

2.3.1 Strategy “P down only”

Electricity production is decreased during the hour in which the plant participates in down-regulation. The heat production remains unchanged. The avoided fuel costs thanks to down-regulation are proportional to the decrease in electricity production and the efficiency to produce electricity.

Using this strategy, it is possible for the producer to participate in down-regulation when the system down-regulation price is below these avoided fuel costs.

2.3.2 Strategy “P down and Q down”

When decreasing both electricity and heat production, the shortfall in heat production must be compensated in the rest of the district heating system since the heat demand must always be satisfied, for example by:

- Reducing dumping in a cooler: this is only possible when a cooler unit is dumping heat. This is essentially a free action.

- Increasing heat production in another unit: this is possible if there is capacity in some other units. Therefore, this compensation action has an associated cost. - Reducing storage to or increasing discharge from heat storage: this is only

possible if there is capacity in the storage to perform these actions. If the storage is scheduled to follow a given plan during the day, any deviation from this plan will have to be compensated for at some point by additional charging (i.e. increased production or reducing dumping in a cooler) during some future hours. This was illustrated in Figure 10. Hence, there may be some costs associated with using the storage in that way. See Section 2.4 and Figure 10 for further explanation of the use of the heat storage for balancing.

The costs of these compensation actions must be accounted for when taking decisions on whether it is profitable to perform down-regulation. Among these compensation actions, only increasing heat production has a cost.

In summary, this strategy entails the following costs and cost savings:

- Avoided fuel costs corresponding to decrease in heat and electricity production in the CHP plant. These avoided fuel costs are larger than for strategy “P down only” since also heat production is reduced, thus leading to additional avoided fuel costs.

- Additional fuel costs corresponding to an increase in heat production of other units or the same unit in subsequent hours (increase to compensate the decrease in heat production in the CHP plant).

The net fuel cost savings are the difference between the avoided fuel costs and the additional fuel costs.

Using this strategy, it is possible for the producer to participate in down-regulation when the system down-regulation price is below these new fuel cost savings.

2.3.3 Strategy “P down and Q up”

Electricity production in the CHP plant is decreased while the heat production is increased during the hour in which the plant participates in down-regulation. The increase heat production must be compensated for by decreasing heat production in some other units.

If the operating cost of these other units are higher than that of the CHP unit, it may become economically profitable to increase the heat production in the CHP plant and decrease the heat production in other more expensive units. This is a form of heat production redispatch.

One may ask why this cannot be done without electricity down-regulation. The reason for this is that the prices on the day-ahead market may have been interesting enough to incentivize a higher electricity production in the CHP plant, which may be followed by a reduction in heat production in order to stay within the feasible operating region. When participating in balancing market, however, it may become interesting to buy back electricity production to re-dispatch heat production as explained above, if the down-regulation price is low enough.

In summary, this strategy entails the following costs and cost savings:

- Avoided fuel costs corresponding to a decrease in electricity production in the CHP plant, and a decrease in heat production in other units.

- Additional fuel costs corresponding to an increase in heat production in the CHP plant.

The net fuel cost savings are the difference between the avoided fuel costs and the additional fuel costs.

Using this strategy, it is possible for the producer to participate in down-regulation when the system down-regulation price is below these new fuel cost savings.

2.4 Overall methodology

The purpose of the project has been to develop a methodology to simulate district heating owners’ decisions to (1) evaluate the profitability of investing in new CHP units as opposed to HO units and (2) evaluate the impact of participating in the balancing market on their overall profitability.

To address this purpose, a model consisting of several modules has been developed. The specifications of the model and its modules have been designed based on discussions with the industrial partners of the project (Vattenfall and Borås Energi och Miljö (BEM)). The following modules are included in the model:

- Module 1: A deterministic rolling scheduling module. This module takes as input a description of the technical and operational cost characteristics of the following units in a district heating system: heat generating units (including CHP units), heat storage facilities (such as hot water storage tanks) and coolers (units to dump excess heat). It also takes as inputs the hourly profile for the heat demand to meet and hourly electricity prices to compute the income from selling electricity produced by CHP units. It returns the optimal hourly schedule of all units (schedule for heat and electricity production, for heat storage and for coolers) for a coming week. The module is meant to be used in a rolling daily manner so the schedule can be updated every day. The module is implemented as a mixed linear integer optimisation problem. It maximizes the operational profit of a heat and electricity portfolio owner while meeting the required heat demand. The operational profit that is maximized is the sum of the hourly revenues from selling electricity minus the sum of the hourly production costs (fuel costs, start-up and shut-down costs). The model is based on [13]. More detail about how module 1 works is given farther down.

- Module 2: A deterministic balancing market participation module. This module takes as input a week-ahead schedule obtained from the module 1 and hourly up- and down-regulation prices on the balancing market for one hour. The module maximizes profits obtained through the participation of the units in the balancing market for one specific hour, while making sure that it is possible to return to the week-ahead electricity production plans for the subsequent hours. The module is implemented as a mixed linear integer optimisation problem. More detail about how module 2 works is given farther down.

- Module 3: A deterministic investment study module. This module takes as input CAPEX, OPEX and the results from the previous two modules over an investment period of interest. It computes the annual cash flows (yearly income from selling heat and electricity minus yearly O&M and fuel costs) and returns a set of investment profitability indicators: internal rate of returns, payback time and net present value.

Table 1 gives a summary of the inputs and outputs of each module.

Table 1: Summary of inputs and outputs of the three modules.

Module 1 List of heat generating units. Technical characteristics of the generating units (efficiency,

minimum and maximum

capacities, ramp rates, min on and off times, PQ operating regions for CHP plants).

Cost data of the generating units (fuel prices, start-up and shut-down prices).

List of coolers and heat storage facilities and their capacities. Hourly time series for heat demand for one week.

Hourly time series for day-ahead electricity prices for one week. Average billing price for heat.

Week-ahead hourly plan for heat and electricity production in each unit.

Week-ahead hourly plan for coolers and heat storage facilities. Revenues from selling heat over one week.

Hourly revenues from day-ahead electricity markets.

Hourly production costs (fuel costs, start-up and shut-down costs).

Profits over one week (sum of revenues minus sum of costs).

Module 2 Same as for module 1. Time series needed for all hours of the day (instead of all hours of the week as for module 1).

Week-ahead hourly plans for heat and electricity production, coolers and heat storage from module 1. Up- and down-regulation prices for one hour.

Up- and down-regulation volumes activated on the balancing market for one hour. Adjusted plans for head and electricity production, coolers and heat storage until the end of the day.

Module 3 Existing district heating systems. New unit to be invested in. Lifetime of the new unit. Discount rate.

Investment cost. O&M cost.

Representative years to be simulated during the lifetime of the new unit.

All inputs for module 1 for these representative years.

Outputs from module 1 for the representative years.

Yearly cashflows (yearly revenues minus production costs minus O&M costs).

Internal rate of return of the investment in the new unit. Payback time of the investment. Net present value of the investment.

Figure 8 illustrate how module 1 works: it determines an optimal week-ahead hourly schedule for all units (including heat storage), given heat demand and day-ahead price forecasts1_{. To determine the optimal schedules, the following inputs are used:}

- Technical data about the units. In particular: minimum and maximum heat production, ramp rates for heat and electricity production, minimum up- and down-time, fuel-to-heat and fuel-to-electricity efficiencies, co-dependence between heat and electricity production.

- Cost data about the units: fuel prices, start-up and shut-down prices.

- Billing price for selling heat (average price paid by the consumers for 1 MWh of heat) obtained from the publicly available statistics in [15].

- Exogenous data:

o Heat demand for the coming week o Day-ahead spot prices for coming week

Revenues from heat and electricity is considered as follows in module 1:

- Heat demand is a constraint that must always be satisfied. The revenues from selling heat is obtained from the yearly heat demand and the average billing price for heat from [15]. Therefore, the decisions taken by module 1 do not impact the revenues from selling heat since the heat demand is always satisfied.

- Revenues from the day-ahead electricity markets are formulated as in Section 2.1.5.

The objective function of the optimization problem solved by module 1 is the sum of the hourly revenues from the electricity day-ahead market minus the production costs (fuel costs, start-up costs and shut-down costs).

The heat storage is constrained to return to a predefined value at the end of the week.

1 _{Module 1 does not generate a bid strategy. Rather, it takes the day-ahead electricity prices as}

inputs and schedules district heating systems optimally to maximize profits from selling electricity. Therefore, it gives a best-case analysis of what the revenues from selling electricity can

Figure 8: Hourly week-ahead schedules from module 1.

Figure 9 illustrates how module 2 works: it models participation in the balancing market for one specific hour (hour 00:00 to 01:00 in the figure). It takes as input the hourly plan for all hours of that day from the week-ahead hourly plan computed by module 1. It also takes as input balancing price forecast for the current hour and updated heat demand forecast. The outputs from module 2 are balancing actions on the electric balancing market (decrease or increase in electricity production) and a re-schedule of all units to allow for these balancing actions.

Module 2 allows balancing actions only in the current hour, that is, no change in electric production is allowed for the subsequent hours. This is equivalent to locking the electricity production for the subsequent hours to the week-ahead electricity production plan. Module 2 allows changes from the week-ahead plan in heat generating units, coolers and heat storages in all hours of the day. A constraint is put onto heat storages to ensure that their stored heat is back at the same level as in the original week-ahead hourly plan (from module 1) at the end of the day This is illustrated in Figure 10. This ensures that the scheduling performed by module 1, which has a longer look-ahead time horizon, is considered in the balancing actions determined by module 2. The capacity that can be used for balancing is constrained by the technical limits of the plants and the day-ahead production plan (first day in the week-ahead production plan), as was explained in Section 2.1.

Note that, in the Swedish setting, the minimum bid size is 10 MW for SE1, SE2 and SE3 and 5 MW for SE4. However, no minimum down- or up-regulation participation limit is set in module 2.

Figure 9: Balancing actions and production re-scheduling at 00.00 from module 2.

Figure 10: Illustration of usage of the heat storage for enabling regulation. At 10.00 down-regulation was performed in one CHP unit by decreasing both electricity and heat production. The heat production decrease was compensated by discharging the storage. From then, the storage deviates from its week-ahead plan but must come back to it at the end of the day.

In addition to the inputs already available in module 1, module 2 requires up- and down-regulation prices for the current hour and an existing week-ahead schedule obtained from running module 1. Revenues from heat and electricity is considered as follows in module 2:

- As for module 1, heat demand is a constraint that must always be satisfied. - Revenues from the balancing markets are formulated as in Section 2.1.5 and are

Module 1 and 2 are thus used in a sequential manner. Module 1 is run first to get an hourly week-ahead plan for heat and electricity production and storage. After that, module 2 is run for each hour of a day to simulate actions on the balancing market. Taken together, the three modules compute the optimal hourly schedule, optimal hourly balancing actions and overall investment profitability of a district heating system. Several use cases are covered by these modules:

- Optimal hourly heat and electricity production schedule to maximize profit on the day-ahead electricity market while meeting the heat demand (module 1) - Optimal hourly participation in the electricity balancing market (up- and

down-regulation) to maximize profit while ensuring the compliance with the overlying day-ahead schedule (model 2).

- Investment profitability indicators when adding new production units to a district heating system.

Modules 1 and 2 have been validated against data provided by the industry partners.

2.5 Case studies

In the case studies, data were collected from the participating companies (Vattenfall’s system in Nyköping and Borås Energi & Miljö’s system) to identify what obstacles and opportunities different actors see when it comes to investments in relation to financial risks, return requirements and owner directives. The following subsections provide more detail about these two district heating systems.

Based on these two systems, the following case studies were performed:

- Case study 1 (Borås): Value of participating in the balancing market. This case study simulates one year of operation. The profits generated by participating on only the ahead electricity market (using only module 1) and on both the day-ahead and balancing market are compared (using both modules 1 and 2).

- Case study 2 (Nyköping): Value of participating in the balancing market. Same case study as case study 1 but applied on the district heating system owned by Vattenfall.

- Case study 3 (Nyköping): Value of increasing heat storage usage for participation on the balancing market. This case study is similar to case study 2 but uses a slightly different version of module 2 in which participation on the balancing market can make a larger use of the heat storage. This case study evaluates the impact of this increased use on the yearly profits.

- Case study 4 (Borås): Investment studies CHP vs Heat-only. Two investment alternatives are considered, one heat-only unit and one CHP. These investment options are compared in terms of different investment profitability indicators (using modules 1 and 3).

The results are reported in Section 3.

In all case studies, the modules from Section 2.4 are used with historical prices as input data. This means that the case studies must be interpreted as best-case results, in the sense that no uncertainty on the electricity prices on either the day-ahead market or the balancing market was considered.

2.5.1 Borås Energi & Miljö

BEM owns two major plants Ryaverket and Sobacken. Sobacken was taken into operation in 2019.

Figure 11 shows the yearly fuel consumption and heat delivery in GWh for a few past years.

Figure 11: Historical fuel consumption (before Sobacken was put into operation)

Two of the units in Ryaverket are biofuel-fired for a total capacity of 90 MW. The other two use waste fuel for a total capacity of 40 MW. A flue gas condensation unit can be used to generate more heat when the waste units are in operation. In addition to the four CHP units, there are several peak boilers, used mainly during constrained situations with high heat demand. More detail about Ryaverket can be found in [16].

Borås’ new CHP plant at the Energy and Environmental Center at Sobacken was ready in 2019. It has a capacity of 120 MW. The CHP plant at Sobacken will in the long term completely replace the biofuel boilers at Ryaverket and supplement the Ryaverket waste boilers. The CHP unit in Sobacken is complemented with a flue gas condensation unit adding approximately 32 MW heat at full boiler load.

A water tank for heat storage (accumulator) is connected to the district heating system. Its maximal capacity is 1900 MWh.

2.5.2 Vattenfall

Vattenfall operates Idbäcksverket in the municipality of Nyköping. This plant has one biofuel CHP unit complemented with two biofuel heat-only units for baseload operation and some oil-fired peak load boilers. The plant is also equipped with a cooler where excess heat can be dumped. The total installed capacity for heat production is 234 MW and the electric installed capacity is 35 MW. A hot water tank for heat storage with capacity 350 MWh is available. More detail about Nyköping’s district heating system can be found in [17].

3 Results

3.1 Case study 1: Value of participating in the

balancing market (Borås)

This case study considers the period 1 January 2018 to 31 May 2018. We compare two cases:

- Participation on day-ahead electricity market only. The district heating system is simulated for the whole study period by using the rolling scheduling model (module 1 from Section 2.4). The scheduling model considers hourly heat demand and day-ahead prices for the coming week. The model schedules the units to maximize profits. The model is run every day in a rolling manner. - Participation on both day-ahead and balancing electricity markets. The district

heating system is simulated for the whole study period by using the rolling scheduling model and the balancing market participation model together (module 2 from Section 2.4). After each run of the rolling scheduling model (module 1), participation in the balancing hour is simulated for each of the 24 hours separating two rolling scheduling model runs. Each run of balancing market participation module maximizes profits on the balancing market for the hour of consideration. Changes in heat production, heat dumping and heat storage from the rolling schedule are allowed with the constraint that the heat storage must be back to exactly the value determined by the rolling schedule at the end of the next day. See Section 2.4 for a detailed description of how the two modules work together.

In both cases, the profits are maximized. The profits consist of

- Revenues from selling heat (the most recent yearly average billing price from [15] is used), selling electricity on day-ahead market and income from participation in balancing market. In this case study, it is assumed that electricity production also receives green certificates (only Sobacken) and a grid support revenue from the local grid owner.

- Costs due to producing heat and electricity (fuel costs) or starting-up units. The outputs of the model runs are the hourly schedules of all units in the district heating system and associated fuel costs, the revenues from selling heat, the revenues from selling electricity on the day-ahead market (see Section 2.1.5) and, in the second case, the revenues from the balancing market. The revenues and costs are summed up over the whole study period to compute the profits made in each case.

Figure 12 shows the costs, revenues (incomes) and profits (revenues-costs) for the two cases. Participation in the balancing market increases the profits by about 400 000 SEK over the five-month study period. The magnitude of this increase is however marginal (it corresponds to 0.1% increase).

Figure 12: Profits, revenues and costs when participating only on day-ahead market (blue - DA) and on both day-ahead and balancing markets (orange - DA and balancing hourly)

Figure 13 and Figure 14 show the breakdown of the revenues for the two cases. The revenues from selling heat do not change since the heat demand is met equally in both cases. The income increase comes from participation in the balancing market, either in the form of extra revenues from up-regulation or of cost saving from down-regulation. Note that the figures show the costs of buying back power for down-regulation and not the overall profit made by down-regulating. Hence, in the case of down-regulation, the numbers give the extent to which energy was bought back by down-regulating.

Figure 13: Breakdown of the revenues when participating in day-ahead market only.

Figure 14: Breakdown of the revenues when participating in both day-ahead and balancing markets.

To give an example of how participation on the balancing market can impact the operations of the plants, Figure 15 shows one instance of down-regulation on 16 May at 06:00. It can be seen that, during that hour, Sobacken decreased its electricity production by about 10 MWh. Hence, the electricity production during that hour deviated from the week-ahead plan by 10 MWh. At hour 07:00, the electricity production is back to the week-ahead plan. Figure 16 shows the heat production in the week-ahead plan and after down-regulation at 06:00. It can be seen that the decrease in electricity

the model has found optimal to use the strategy “P down and Q down” (described in Section 2.3.2) in that particular instance of down-regulation. It can also be seen that the decrease in heat production at hour 06:00 is planned to be compensated later by an increase in heat production at hours 22:00 and 23:00.

The reason for this compensation can be seen in Figure 17 and Figure 18. Figure 17 shows the heat stored in the heat storage in the week-ahead plan and after down-regulation at 06:00. There is a decrease in stored heat due to down-regulation. The magnitude of this decrease can be read in Figure 17. It is equal to the decrease in heat production at hour 06:00. Because the heat demand must be satisfied, the deviation in heat production from the week-ahead plan was compensated by extracting heat from the heat storage. As explained in Section 2.4, the heat storage must be back to the week-ahead plan at the end of the day after participation in the balancing market. As can be seen in both figures, this is the case in this instance thanks to the increase in heat production at hours 22:00 and 23:00 planned by module 2 to allow down-regulation in hour 06:00.

Figure 15: Electricity production in Sobacken in the week-ahead plan and after down-regulation at 06:00.

Figure 17: Stored heat in week-ahead plan and after down-regulation at 06:00.

Figure 18: Difference in stored heat. The graph shows stored heat after down-regulation minus stored heat in the week-ahead plan.

The conclusion of this case study is that the participation in balancing market only leads to a marginal increase of the overall profits over the study period. It was also seen the heat storage can be used to allow participation in down-regulation.

3.2 Case study 2: Value of participating in the

balancing market (Nyköping)

This case study is like the previous one but is performed on the district heating system from Vattenfall in Nyköping. The study period is 1 January 2018 to 30 November 2018.

3.2.1 Impact of balancing market participation on profits

The same two cases as for case study 1 are simulated. The profits are maximized. The profits consist of

- Revenues from selling heat (the most recent yearly average billing price from [15] is used), selling electricity on day-ahead market and income from participation in balancing market. In this case study, no additional income is considered. - Costs due to producing heat and electricity (fuel costs) or starting up units. The outputs of the model runs are the hourly schedules of all units in the district heating system and associated fuel costs, the revenues from selling heat, the revenues from selling electricity on the day-ahead market (see Section 2.1.5) and, in the second case, the revenues from the balancing market. The revenues and costs are summed up over the whole study period to compute the profits made in each case.

Figure 19 shows the costs, revenues and profits for the whole study period in the two cases (day-ahead market only and day-ahead + balancing markets). As was the case in case study 1, participation in the balancing market increases profits but only marginally (0.05 %). The conclusion from this case study is thus identical to that of the previous one.

Figure 19: Costs, revenues and profits when participating only on day-ahead market (blue - DA) and on both day-ahead and balancing markets (orange - DA and balancing hourly)

One of the modelling aspects that may restrict participation in balancing market is the constraint on the heat storage to be back to its value in the rolling schedule at the end of next day. In case study 3, this constraint is loosened to study its impact on the overall profit.

3.2.2 Analysis of balancing market participation

In addition to the analysis of profits above, the participation in the balancing market was investigated in more detail to understand how often the units participate on the balancing market and under what conditions that happens. The investigation here focuses on down-regulation. A similar investigation could be made for up-regulation. Figure 20 shows down-regulation volumes (in electricity production) versus the system down-regulation prices for each hour during which the CHP unit had some spare capacity to participate in down-regulation (that is, the day-ahead production plan is above the minimum production level). The price thresholds that give the dot colours are the ones described in the previous section and corresponding to the CHP unit in Nyköping’s system:

- “P down only” is the system down-regulation price threshold under which it becomes economically interesting to down-regulate electricity production only, see Section 2.3.1.

- “P down and Q down” is the system down-regulation price threshold under which it may become economically interesting to down-regulate electricity and heat production in the CHP unit, see Section 2.3.2.

The down-regulation price threshold for the “P down and Q up” strategy (see Section 2.3.3) is not illustrated here since it depends on what other units are online that can be re-dispatched.

It can be seen from the figure that whenever the system down-regulation price is below the “P down only” threshold, it was profitable to participate in down-regulation except

in two instances (only blue dot with 0 MWh in actual down-regulation). The reason for this is that, although the CHP unit had some spare capacity to down-regulate, there is another constraint that must be met and that made it impossible to participate in down-regulation: the constraint that ensures that it is possible to follow the week-ahead schedule after any balancing actions (see description of module 2 in Section 2.4). In these specific instances, the rolling schedule was planning for an increase in heat and electricity production in the next hour. This planned increase was done with full ramping capacity. Hence, any down-regulation action would make it impossible to get back to the rolling schedule since the ramping capacity would not allow it.

When the system down-regulation price is between the “P down only” and “P down and Q down” thresholds, the orange dots without actual down-regulation show that although the CHP unit had capacity for down-regulation of electricity, there was no flexibility in the rest of system to allow for it.

Figure 20: Volume of participation in down-regulation versus system down-regulation price. Each dot represents one hour. Each hour of the study period during which the CHP plant had capacity for regulation are plotted. Blue dots: system regulation price below electricity only down-regulation price threshold. Orange dots: system down-down-regulation price between electricity only and electricity+heat down-regulation price thresholds. Green dots: system down-regulation price above electricity+heat down-regulation price threshold.

Figure 21 summarized Figure 20 by counting the number of hours during which it was profitable to participate in down-regulation for difference system down-regulation price levels. The figure shows that most of the hours during which it is economically profitable to participate in down-regulation are hours with a system down-regulation between the “P down only “ and “P down and Q down” price thresholds.

Figure 21: Number of hours during which it is economically profitable to participate in down-regulation, depending in the system down-regulation price. Example: it was economically profitable to participate in down-regulation when the system down-regulation price was above the “P down and Q down” threshold for 17 hours.

Figure 22 shows for how many numbers of hours the three strategies were used. The most common strategy is by far “P down and Q down”, that is, to down-regulate both electricity and heat production. This is in line with the reasoning in the previous section which showed that, under the assumption that there is some flexibility in the system, this strategy can lead to higher profits.

Figure 22: Number of hours during which different down-regulation strategies were used, depending on the system regulation price thresholds. Example: When the system regulation price was between the “P down only” and “P down and Q down” price thresholds, down-regulation was performed by decreasing electricity production and increasing heat production for 14 hours.

Finally, Figure 23, Figure 24 and Figure 25 show the number of hours in which the CHP unit participated in down-regulation, compared to the total number of hours when it was possible to participate, for different levels of system down-regulation price.

Figure 23 shows that there was a total of about 3800 hours when the system was down-regulated in the study period. Among these, the CHP unit had down-regulation capacity in about 2300 hours. The CHP unit actually participated in about 250 hours, almost always partially.

Figure 23: Number of hours with participation in down-regulation, considering all possible hours in which the system was down-regulated. See Table 2 for an explanation of the legend labels.

Figure 24 shows that there was a total of about 1600 hours when the system down-regulation price was below the price threshold for the “P down and Q down” strategy. Among these, the CHP unit had down-regulation capacity in about 1100 hours. The CHP unit actually participated in slightly more than 200 hours, almost always partially.

Figure 24: Number of hours with participation in down-regulation, considering only hours for which the system down-regulation price was below the “P down and Q down” threshold. See Table 2 for an explanation of the legend labels.

Figure 25 shows that there was a total of about 160 hours when the system down-regulation price was below the price threshold for the “P down only” strategy. Among these, the CHP unit had down-regulation capacity in about 20 hours. The CHP unit actually participated in all of these 20 hours but two, as was discussed above in Figure 20.

Figure 25: Number of hours with participation in down-regulation, considering only hours for which the system down-regulation price was below the “P down only” threshold. See Table 2 for an explanation of the legend labels.

Table 2: Explanation of the legend labels of Figure 23, Figure 24 and Figure 25.

Legend label Explanation

“Down-regulation hours” Total number of down-regulation hours on a system level (i.e. during which some down-regulation bids were activated by the TSOs).

“Possible down-regulation hours” Number of hours when the system was down-regulated (previous legend label) and the CHP unit had down-regulation capacity. That is, the day-ahead plan was above the minimum electricity production.

“Participation in down-regulation” Number of hours when the CHP unit

participated in down-regulation.

“Full participation in down-regulation”

Number of hours when the CHP unit participated fully in down-regulation. That is, all the possible down-regulation capacity (day-ahead plan minus minimum electricity production level) was used.

“Partial participation in down-regulation”

Number of hours when the CHP unit participate in down-regulation but only partially. That is, the electricity production was decreased by less than the

maximum possible down-regulation capacity.

In conclusion, they are several strategies for CHP units to participate in down-regulation. These strategies require different system down-regulation prices to be profitable. The simplest strategy is to only decrease the electricity production in the CHP unit (strategy “P down only”). However, this strategy requires a lower system down-regulation price compared to other strategies to be profitable, which, in the case study, didn’t occur very frequently (Figure 25). This strategy only requires available down-regulation capacity in the CHP unit.

The strategy “P down and Q down” in which both electricity and heat production are decreased is profitable for higher down-regulation prices if they are some cheap compensating actions that can compensate for the decrease in heat production in the CHP unit. The availability of these actions depends on the flexibility available in the district heating system (coolers and heat storage).

3.3 Case study 3: Value of increasing heat

storage usage for participation on the

balancing market (Nyköping)

This case study is a follow-up on the previous one. It is performed on the same system (Vattenfall’s district heating system in Nyköping) and for the same study period (1 January 2018 to 30 November 2018). The difference is that for participating in balancing market, the heat storage can deviate from the rolling schedule by 5% at the end of next day. In the previous case study, the heat storage following any balancing actions had to be back at the same exact value as in the rolling schedule. This was also illustrated in the analysis of case study 1 in Section 3.1.

Figure 26 shows the costs, revenues and profits of the new balancing market participation strategy compared to day-ahead only. This can be compared with Figure 19 for the previous balancing market participation strategy. The overall profits have now decreased compared to participating in the day-ahead market only. The decrease is marginal (it corresponds to a 0.1 % decrease).

Figure 26: Costs, revenues and profits when participating only on day-ahead market (blue - DA) and on both day-ahead and balancing markets (orange - DA and balancing hourly)

Figure 27 and Figure 28 show the breakdown of the revenues for the two heat storage strategies (that of case study 2 and that of case study 3, respectively) for participation in the balancing market. Allowing the heat storage to deviate from the rolling schedule by 5% at the end of next day allowed for increased participation in the balancing market. Both the costs for buying back power for down-regulation and the revenues from selling power for up-regulation have increased. However, the revenue from the day-ahead market has decreased due to the more active participation on the balancing market. The net effect of this is the decrease in overall profit as shown in Figure 26.

Figure 27: Breakdown of the revenues (for case

study 2) Figure 28: Breakdown of the revenues (for case study 3)

The reason for this overall decrease in profit is that a more aggressive use of the heat storage in the participation on balancing market may result in less capacity being available for use on subsequent day-ahead market runs. This stems from the different look-ahead horizons in the week-ahead scheduling (module 1) and in the balancing market participation (module 2). Module 1 considers a time horizon of 7 days ahead. Module 2 only considers up to 36 hours ahead (one hour of balancing prices and up to 36 hours of heat demand). Hence, the rolling scheduling model can capture variations in the heat demand over a longer time horizon and optimize the fleet of units (including the

heat storage) for these variations. Therefore, whereas allowing the heat storage to deviate more from this rolling schedule may increase participation in balancing market, it does not necessarily increase the overall profit.

3.4 Case study 4: Comparison of investment in

CHP vs heat-only plants (Borås)

Two investment studies are performed to compare the profitability of CHP and heat-only plants. These investment studies aim at studying the profitability of investing in two new units to replace the two biofuel-fired units at Ryaverket that will likely be decommissioned within a few years. Two alternatives are considered in each investment study: two CHP units or two heat-only (HO) units. The technical characteristics of the two new units are presented in Table 3 and were constructed based on [1] and [2]. The costs for the two alternatives are presented in Table 4 and were taken from [1].

Table 3: Technical characteristics of the new units.

Quantity Unit 1 Unit 2

Installed capacity 50 MW 25 MW

Fuel Wood chips Wood chips

Flue-gas condensation unit capacity

12.5 MW 6.25 MW

Low-load level 10 MW 5 MW

Alpha-value 0.4 at lower electricity

production level 0.5 at highest electricity production level 0.4 at lower electricity production level 0.5 at highest electricity production level Maximum electricity production level 17 MW 8 MW Minimum heat production level 10 MW 5 MW Heat efficiency 89% 89 %

Table 4: Cost assumptions from [1].

Type of plant Quantity Value

CHP Normalized investment cost 1.08 M€ / MW installed capacity HO Normalized investment cost 0.58 M€ / MW installed capacity

CHP Fixed O&M 45 800 € / MW installed capacity / year

HO Fixed O&M 49 100 € / MW installed

capacity / year

CHP Variable O&M 1.4 € / MWh fuel input /

year

HO Variable O&M 3 €/MWh fuel input / year

The two investment studies differ by the assumptions on the units kept in the baseline portfolio2_:

- Investment study 1: 2 units are assumed to be decommissioned in BEM’s portfolio and replaced by investment in two new units. Investment in new CHP and HO units are compared.

- Investment study 2: same as investment study 1 but Sobacken is also excluded from the portfolio.

The reason for performing these two investment studies can be seen in Figure 29 and Figure 30. These figures show the hourly heat demand profile for a whole year (exact values have been hidden for confidentiality purposes), the base load capacity of the units in the portfolio before investing in new units. When Sobacken is kept in the portfolio, Figure 29 indicates that the new units would cover a very small part of the demand (15 hours) and will most likely not be profitable investments. Note that the new units may still be needed to provide redundancy in the system. When Sobacken is excluded, Figure 30 indicates that the new units will run most of the time between October and May. Note that these two figures give a simplified view of the real scheduling of the plants since they only consider heat production capacity and a merit order hour-by-hour without consideration of electricity production, ramp rates and heat storage.

Figure 29: Base load capacity with Sobacken _{Figure 30: Base load capacity without Sobacken.}

Therefore, these two investment studies evaluate the profitability of different kinds of investments:

- Investment study 1 studies an investment for redundancy: new units will probably not run often but may be needed for redundancy reasons in case of failure of a large unit.

- Investment study 2 studies investment in base load units that will run most of the time.

Figure 31 and Figure 32 illustrates these differences by showing the installed capacities considered in the two investment studies. The scale is different in the two figures and, therefore, it is the relative sizes of the blocks and not their absolute sizes that can be compared.