Sie sind auf Seite 1von 16
Energy 118 (2017) 156 e 171 Contents lists available at ScienceDirect Energy journal homepage:

Contents lists available at ScienceDirect

Energy

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

Energy journal homepage: www.elsevier.com/locate/energy Hydrogen generator characteristics for storage of

Hydrogen generator characteristics for storage of renewably-generated energy

characteristics for storage of renewably-generated energy Janusz Kotowicz * , Ł ukasz Bartela, Daniel We ˛

Janusz Kotowicz * , Ł ukasz Bartela, Daniel We˛ cel, Klaudia Dubiel

Institute of Power Engineering and Turbomachinery, Silesian University of Technology, Konarskiego 18, 44-100 Gliwice, Poland

article info

Article history:

Received 24 May 2016 Received in revised form 4 November 2016 Accepted 7 November 2016 Available online 18 December 2016

Keywords:

Hydrogen generator Ef ciency Power-to-Gas Economic analysis

abstract

The paper presents a methodology for determining the ef ciency of a hydrogen generator taking the power requirements of its auxiliary systems into account. Authors present results of laboratory experi- ments conducted on a hydrogen generator containing a PEM water electrolyzer for a wide range of device loads. On the basis of measurements, the ef ciency characteristics of electrolyzers were determined, including that of an entire hydrogen generator using a monitored power supply for its auxiliary devices. Based on the results of the experimental tests, the authors have proposed generalized characteristics of hydrogen generator ef ciency. These characteristics were used for analyses of a Power-to-Gas system cooperating with a 40 MW wind farm with a known yearly power distribution. It was assumed that nightly-produced hydrogen is injected into the natural gas transmission system. An algorithm for determining the thermodynamic and economic characteristics of a Power-to-Gas installation is pro- posed. These characteristics were determined as a function of the degree of storage of the energy pro- duced in a Renewable Energy Sources ( RES) installation, de ned as the ratio of the amount of electricity directed to storage to the annual amount of electricity generated in the RES installation. Depending on the degree of storage, several quantities were determined.

© 2017 Elsevier Ltd. All rights reserved.

1. Introduction

It is forecast that the share of renewable energy sources in the world's electricity production structure will increase [1,2] . In Poland, wind turbines and photovoltaic cells are among the most dynamically developing RES technologies [3,4] . The most important advantage of use these sources is lack of emission of harmful sub- stances. In contrast to use of carbon sources the use of renewable energy does not contribute to the emission of dust, sulfur com- pounds, nitrogen oxides, mercury and carbon dioxide. Common- ness of occurrence and renewability of energy potential make that these energy sources are safe and most commonly cheap in exploitation. The popularity of RES is stimulated by support mechanisms regulated by national laws [5,6] . The environmentally- friendly nature of these technologies is a feature that deserves rational support. The basic mechanism of this support, in addition to a system of tradable certi cates, is guaranteed priority access to the electricity grid. This mechanism creates a situation in which the

* Corresponding author. E-mail address: janusz.kotowicz@polsl.pl (J. Kotowicz).

0360-5442/ © 2017 Elsevier Ltd. All rights reserved.

large, un-adapted coal- red units working in the largest centrally organized systems take on a role of regulation system. The Polish energy system has a relatively low level of elasticity. In addition, various renewable energy sources are characterized by consider- ably differing potentials for electricity generation, which rarely correlate with end-user electricity demand, particularly in the case of wind energy [7 e12] . Mentioned features and the necessity to ensure energy security causes that the development of renewable energy sources is limited by the operator of the electricity system. A further increase in installed wind farm capacity can lead to a situation in which coal- red units will reach the limits of their regulatory capabilities, and the power system will become unstable [13,14] . Therefore further increase in installed capacity of wind farms will be possible by improving operational features of coal- red power plants [15 e17] . This will be the result of the startup of new units with supercritical parameters [18,19] . It is also important that the increase of installed capacity in power plants will be based on gas turbines [20 e 22] . However, experts from the Polskie Sieci Elektroenergetyczne S.A. predict that the dynamic of startups of new wind farms will contribute to insecurity of the system by 2025. In addition to the investment in modern power plant units, it is also very important to implement energy storage

J. Kotowicz et al. / Energy 118 (2017) 156e171

157

systems to avoid this scenario [23 e 27] . Energy storage systems would draw power during off-peak demand periods and produce during peak periods [28,29] . Energy storage is important for a wide range of potential ben- e ciaries of technology - from individual users of electric equip- ment, automotive users, and industry to network operators. Depending on the scale, energy storage can be classi ed into ve groups by storage method: mechanical, electrical, chemical, elec- trochemical and thermal. In the case of cooperation with the large power system, the most favorable characteristics have chemical and mechanical solutions, including pumped storage and com- pressed air energy storage (CAES) [30,31] . A signi cant develop- ment potential have also electrochemical systems. Wide discussion on the potential of the various solutions were carried out in Refs. [32,33] . This article presents the results of analyses for Power- to-Gas technology, which is classi ed in the chemical group. Ac- cording to authors' assumptions, the analyzed system is used in a daily cycle and enables the production of hydrogen during the so- called night valleys, i.e., at night. The hydrogen produced is accu- mulated in buffer tanks and directed to the gas transmission grid.

2. Power-to-Gas technology

remembered, however, that hydrogen is dif cult to transport or use for energy generation [48,49] . The literature often emphasizes the problems connected with the use of hydrogen-rich fuels in gas turbines and piston engines [50,51] . Synthetic Natural Gas (SNG) can be a more convenient fuel for these purposes. Synthetic Natural Gas can be produced from carbon dioxide and hydrogen and can be the end-product of a Power-to-Gas installation; however, it re- quires installing a methanation reactor. A Power-to-Gas installation can be of interest in power plants whose emission of carbon dioxide generates higher operations cost. Carbon dioxide can be a product of gasi cation of biomass [52,53] . The advantage of these systems is the possibility of ef cient use of oxygen, which is the by-product of the electrolysis process. Optional ways of the use of oxygen, and their potential has been described in Ref. [54] . Worldwide, pilot installations currently use methanation reactors based both on biological and chemical processes. The Power-to-Gas technology can be also implemented successfully in other sectors of the in- dustry e.g. in pulp mill, which is a large producer of wood-based CO 2 , so the hydrogen can be used for the methanation process and the oxygen in the pulp and paper mill, replacing the existing O 2 production facility [55] . Fig. 1 shows possible connections in a Power-to-Gas system.

The Power-to-Gas technology allows to produce the gas of high potential energy using electricity taken directly from a generation system or the network. The transformation process within a Power- to-Gas installation of electricity to the chemical energy of gas fuel can be justi ed during periods of higher electricity production and lower electricity demand. This overproduction occurs mainly in systems based on renewable sources (mainly solar and wind sources). A positive economic effect for investments in energy storage systems may occur in a situation in which the price of en- ergy in a period of overproduction is very low [34,35] . The over- production usually occurs during low demand periods (the night valley, the weekend valley). Such scenarios commonly occur in countries where the share of renewable energy is signi cant (e.g., Denmark, Spain, and Germany). This is also a justi ed investment from the point of view of the power system operator. In this case, the Power-to-Gas installation can adopt a function of regulator and adapt the available capacity of the system to the demand [36 e 38] . The strong interest in energy storage systems, due to their ability to regulate supply to accommodate demand, is observed particularly in countries where a growing share of RES is accompanied by a signi cant share of coal- red power plants [12,39 e 42] . In Poland for example, increasing installed wind farm capacity with priority access to the grid is in force, resulting in an increase in forced shutdowns of power units. Shutdowns of power units are required by the system operator's instructions; they may result from high outputs from wind-power sources and low electricity demand [43,44] . Increased coal- red power in off-peak periods could be directed to storage; this could contribute to a decrease of the number of the forced shutdowns of units, and thus reduce the related cost [45 e 47] . The criterion for classifying Power-to-Gas technology can be based on the type of gas produced. The basic element of the system, irrespective of the type of gas, is the electrolyzer, whose function is to produce hydrogen. If the installation contains no additional gas conversion equipment, hydrogen is the nal product of the process. In this case, the hydrogen can be injected into the natural gas grid, forming part of the gas transportation infrastructure. Additionally, the product of the electrolysis process can be delivered by tank trucks to industry (e.g., chemical, metallurgical, electronic, re ning and fats industries). It can also be used for energy purposes at the place of production (Power-to-Gas-to-Power). It should be

3. Production of hydrogen

Hydrogen is a product of an electrolysis installation, constituting of battery of currently available compact electrolyzers. The most popular types of electrolyzers for pilot installations worldwide are electrolyzers with a polymer electrolyte membrane (PEM) [45,56,57] and alkaline electrolyzers (AEL) with aqueous alkaline solutions as electrolytes (KOH or NaOH) [58 e 60] . Other types of electrolyzers include devices with a ceramic membrane (SOECs) [61 e66] . PEM and alkaline electrolyzers operate at a process tem- perature between 70 and 100 C. SOEC electrolyzers operate be- tween 700 and 1000 C. These types of electrolyzers require a supply of heat in the form of steam generated in an external process [67] . This limits the potential points of use. Because of the high process temperature, the ef ciency of SOEC electrolyzers (de ned as the ratio of the chemical energy of fuel obtained to input electric energy) may exceed 100%. SOEC electrolyzers are rarely considered in installation planning; this may serve to regulate power systems because of the long heating and cooling periods of system ele- ments. In this respect, the PEM and alkaline electrolyzers are more favorable. In addition, the times required to change their loads are from 10% to 100% of the nominal power per second and from 10% to 25% of the nominal power per second, respectively. The PEM electrolyzers, despite lower technological maturity, are a perfect alternative to alkaline electrolyzers because of their wide load range (5% e 100% of nominal power), very high purity of produced hydrogen ( > 99.999%) and lower operating cost. Unfortunately, the lower technological maturity of PEM electrolyzers is re ected in lower membrane lifetime and higher restoration costs [68,69] . An important aspect of the installation, irrespective of the type of electrolyzer, is that the water used for electrolysis must comply with the strictest standards; therefore, it is necessary to install water treatment in Power-to-Gas systems. Especially, the removal of minerals and ions is required before begin electrolysis process [70,71] . The ef ciency of the hydrogen generators used is important for the ef ciency of the entire Power-to-Gas storage system. There are highly diverse methods to determine this quantity. Knowledge of the ef ciency characteristics as a function of electrolyzer load is the basis for a correct analysis of an energy storage system cooperating with a generation system.

158

J. Kotowicz et al. / Energy 118 (2017) 156e171

158 J. Kotowicz et al. / Energy 118 (2017) 156 e 171 Fig. 1. Diagram of

Fig. 1. Diagram of possible connections of a Power-to-Gas system.

3.1. The ef ciency of hydrogen generators

The ef ciency of a water electrolysis process may be de ned as the ratio of the theoretical amount of energy E t to the actual amount of energy E r required to split 1 mol of water into hydrogen and oxygen, according to the relationship:

h ec ¼ E t

r

E

;

(1)

In real systems operating in the low-temperature technology, the electrolysis process proceeds according to chemical reaction:

H 2 O ð l Þ ¼ H 2 ð g Þ þ

2 1 O 2 ð g Þ ;

(2)

where the water is in a liquid state ( l ), and hydrogen and oxygen are gaseous ( g ). To maintain the electrolysis process it is necessary to supply electricity and heat. The minimum amount of electrical energy required to split of 1 mol of water is equivalent to Gibbs free energy DG 0 , which is related to the free energy voltage E 0 . For splitting liquid water, this energy is:

DG 0 ¼ n $ F $ E 0 ¼ 237 : 22 kJ = mol ;

(3)

where n e the number of electrons transferred during the elec- trochemical reaction of splitting one molecule of water ( n ¼ 2), F e Faraday constant ( F ¼ 96,485 C/mol), E 0 e standard electrolysis potential (at temperature 298 K and pressure 1 bar this voltage is equal E 0 z 1.23 V) [71 e73] . The heat required is the result of the entropy change and the temperature at which water is supplied. The entropy change for the liquid water is DS ¼ 163,15 J/(mol $ K). Hence, the total energy required to split 1 mol of water is the sum of the Gibbs free energy and the heat required, which corresponds to the enthalpy change DH 0 for the hydrogen:

DH 0 ¼ DG 0 þ T $ DS ð T Þ ¼ 285 :84 kJ = mol ;

(4)

With alkaline and PEM electrolyzers, the heat needed for the reaction is generated by current ow through the electrolysis cell. However, this requires supply of a voltage higher than E 0 to the

electrolysis cell. The theoretical amount of energy E t can therefore be de ned by the thermoneutral voltage V 0 , which is related to the enthalpy change for the liquid water:

0

V 0 ¼ DH F z 1 : 48 V ;

n

$

(5)

The theoretical amount of energy can be calculated from the relationship [74] :

E t ¼ V 0 $ I $ t ;

(6)

where: I e current owing through the electrolyzer, t e time. The thermoneutral voltage V 0 speci ed above is related to the higher heating value (HHV) of hydrogen, which is equal to the enthalpy change DH 0 for hydrogen [74] . When calculating the ef- ciency of the electrolysis process, it is usual practice to use the lower heating value (LHV) of hydrogen. Manufacturers of electro- lyzers often use speci c energy consumption in kWh/m 3 n or kWh/ kg, to describe their effectiveness. The actual energy E r needed to split water into hydrogen and oxygen is mainly dependent on the voltage U supplied to the electrolyzer, which is higher than the thermoneutral voltage V 0 because of losses related to the electrolytic resistance, activation of the electrolysis process and mass transport through the membrane [75] :

E r ¼ U $ I $ t ;

(7)

The actual energy is thus the electric energy supplied to the electrolyzer. From equations (1), (6) and (7) , it follows that the ef- ciency of the electrolyzer is equal to the ratio of the thermoneutral voltage V 0 and the voltage supplied the electrolyzer U. The voltage values given above refer to standard conditions (temperature 298 K and pressure 1 bar). However, the voltages depend on temperature and pressure, and in the case of the voltage U, on the current density in the electrolytic cell; thus the theoretical ef ciency of the electrolysis depends on the ratio of voltages, the temperature, the pressure and the electrolyzer load [75] . Due to the prediction of an electrolyzer's operation with varying loads, the changes in values of current owing through the electrolyzer must be taken into account. Generally, the ef ciency of an electrolyzer

J. Kotowicz et al. / Energy 118 (2017) 156e171

159

increases with a decrease in electrolyzer output compared to the nominal output. This is because losses are reduced with a smaller supply current. The reduction in current value causes a nonlinear reduction of losses, mainly related to ohmic resistance in the electrolyzer, which is independent of the load [75,76] . The elec- trolyzer operating temperature has a signi cant impact on its hydrogen production ef ciency; if it increases, it causes an increase of ef ciency. The electrolyzer operating pressure has less impor- tance [74,75,77] . Because the electrolyzer is direct current (DC) powered, and considering equations (1), (6) and (7) , as well as using the higher heating value (HHV) of hydrogen, the electrolyzer ef ciency will be calculated with the formula:

h EL ¼

_

V H 2 $ HHV

N DC

_

;

(8)

where: V H 2 e volume ow rate of produced hydrogen, N DC e DC

electrical power supplied to the electrolyzer (9). Electrical power is determined by the relationship:

N DC ¼ U $ I :

(9)

Hydrogen generators that use electrolysis for hydrogen pro- duction are usually AC powered. This increases the energy re- quirements of the entire system, mainly due to AC/DC converters, which operate with a speci ed ef ciency. In addition, generators are equipped with a series of auxiliary devices, normally powered from the same source. Therefore, the ef ciency of the hydrogen generator is determined by:

h HG ¼

_

V H 2 $ HHV

N AC

;

(10)

where: N AC e AC electric power supplied to the hydrogen generator. Some analyses additionally account for hydrogen losses in a hydrogen puri cation system and in some safety devices in the hydrogen ow path. This is important, especially in small units with capacities of several m 3 n /h in which the ow of hydrogen produced by the generator can be even a few percent lower compared to the theoretical capacity that results from the value of the current [78 e 80] . Because the hydrogen ow is measured at the outlet of the hydrogen generator, all possible losses occurring within the entire generator are included in determining its ef ciency. Commonly available hydrogen generators are composed of several basic systems:

electrolyzer e a basic element of the generator in which water is split into hydrogen and oxygen; it is usually in the form of a stack of electrolytic cells supplied with direct current. A single generator may comprise several stacks. monitoring and control system - regulates the proper supply of the electrolyzers and controls the electrolysis process; controls the operation of power systems and auxiliary equipment to maintain operating parameters of the generator. auxiliary equipment - pumps, fans, cooling systems, heaters, hydrogen puri cation systems [81] .

converter nominal power and the type of elements used in the recti er circuit are of great importance [80] . In some solutions it is possible to connect the electrolyzer directly to a DC power source, which may be, e.g., photovoltaic cells (PV). It is required to match the voltage and current levels of the PV system to the electrolyzer or use adequate DC/DC converters. In every case, power losses of approximately 10% are to be expected when operating in the range of 50% e 100% of nominal power, with losses increasing as electro- lyzer power decreases [75,83] . The auxiliary equipment and the monitoring and control sys- tems of hydrogen generators are usually supplied from the same energy source as the electrolyzer. However, often they require different voltage values. In this case, it is necessary to apply voltage converters other than those in the supply system of the electro- lyzers. It is also possible to supply, e.g., pumps or fans directly from the AC mains. Power required for these devices is constant over the entire capacity range of the most frequently used hydrogen gen- erators. Therefore, the applied AC/DC converter will always work with the same ef ciency. With the above assumptions, the power N AC can be determined according to the relationship:

N AC ¼ N DC þ N aux ;

(11)

where: N aux e sum of the auxiliary power and losses.

N aux ¼ DN þ N AU þ N CS þ N p ;

h

auxAC = DC

(12)

where: DN e power losses in AC/DC converter of electrolyzer, N AU e electric power of auxiliary equipment, N CS e electric power of the monitoring and control systems, N p e electric power of pumps, h auxAC/DC e ef ciency of any AC/DC converter supplying auxiliary systems (not a function of the electrolyzer power). The ef ciency of a hydrogen generator also depends on the size of the device. The construction of larger units with higher elec- trolyzer capacities allows reducing losses. Based on the literature data and technical data of currently available PEM electrolyzers, the impact of nominal capacity on the achievable ef ciency of elec- trolyzers and the entire system for hydrogen production can be

determined ( Fig. 3 ) [47,78,84 e 89] . The tendency of increasing ef- ciency with increasing capacity is clear. However, there is a sig-

ni cant difference between the ef ciency of the cell itself and that

of the entire system for hydrogen production (hydrogen generator); this ef ciency difference may even reach 20%. It should be noted that an increase in nominal capacity has less impact on the ef - ciency of electrolyzers than on that of the hydrogen generator. This is related to a lower auxiliary power index d in large hydrogen- production units, which takes a value of approximately 10%. This index is de ned by:

d ¼ N aux

N AC

;

(13)

The largest units with PEM technology currently offered are

characterized by ef ciencies (HHV) of 70%. All these values are obtained at the nominal power. A partial electrolyzer load results in

a change in its hydrogen production ef ciency.

A block diagram of a hydrogen generator system, along with the direction of power ow, is shown in Fig. 2 . An electrolyzer is normally powered from the AC mains through an AC/DC converter, which is a source of relatively high energy losses. The ef ciency of an AC/DC converter is load-dependent and in the case of small converters uctuates near 90% at its nominal output power; at low load it may drop below 70% [75,82] . The

3.2. Experimental determination of hydrogen generator ef ciency

In order to determine the ef ciency characteristics of an actual hydrogen generator and electrolyzer, a device manufactured by the TsvetChrom company was tested. The subject of experimental research was a hydrogen generator with four PEM electrolyzers whose total rated capacity as speci ed by the manufacturer was

160

J. Kotowicz et al. / Energy 118 (2017) 156e171

160 J. Kotowicz et al. / Energy 118 (2017) 156 e 171 Fig. 2. Diagram of

Fig. 2. Diagram of hydrogen generator.

118 (2017) 156 e 171 Fig. 2. Diagram of hydrogen generator. Fig. 3. The ef fi

Fig. 3. The ef ciencies of commercially available PEM electrolyzers.

0.108 m 3 n /h. Nominal parameters of a single electrolyzer were:

I zn ¼ 13 A, U zn ¼ 14 V, which translates into a total power of all electrolyzers of N DC ¼ 728 W [90] . Electrolyzers in the hydrogen generator were electrically connected in series and were connected to an external DC power source that allowed changing the power supplied; they all contributed to a common hydrogen outlet. A diagram of the tested circuit is shown in Fig. 4 , which presents all electrolyzers schematically as a single electrolytic cell. The points at which measurements were made are indicated also. During testing, the entire system should be in steady state operation to maintain constant values: current and voltages on electrolyzer stacks, operation temperature and pressure of elec- trolyzers. It was assumed that all of the auxiliary equipment operated stably with constant power consumption. In determining the auxiliary power periodic run the pumps for example re lling water tanks, periodic opening the valves in order to clean hydrogen tubes from residual water, switch on the signaling lights was omitted. DC voltage and current was measured directly on the

electrolyzer. The operating temperature was determined based on temperature measurements of water owing out of the electro- lyzer. The hydrogen pressure was controlled at the outlet of hydrogen from the generator. The electricity demand of the hydrogen generator was determined based on measurement of the

AC power, which supplied all circuits of the generator.

Accurate measurement of hydrogen ow rate obtained from the electrolyzer was very important for the proper determination of ef ciency. This measurement directly out of the electrolyzer is very dif cult because some water and oxygen penetrate the electrolyzer membrane and ow together with the hydrogen. It is necessary to use a puri cation system, which dries the hydrogen and removes impurities. Theoretically, these systems produce hydrogen with a purity of more than 99.999%. Only this puri ed hydrogen is fed to

the hydrogen generator outlet and measured using a thermal ow

meter. During the experiments, the DC current supplying the electro- lyzers was varied in the range of approximately 15% e 108% of the rated value (2 e14 A), and the operational temperature kept be- tween 22 and 32 C. Steady state operation of the electrolyzers was

dif cult to maintain because of the periodically opening water purge valve in the hydrogen puri cation system; this resulted in momentary changes in electrolyzer voltage and uctuations in hydrogen ow. All measurements were performed with a hydrogen outlet pressure equal to atmospheric pressure. Auxiliary equipment was powered from the AC power source. The measured power consumed by these devices (the second part of the formula (12) ) was independent of the electrolyzer current and was equal to 57 W, also accounting for the AC/DC conversion losses. Based on this study, the characteristics of the hydrogen generator were determined. Fig. 5 shows the characteristics of the sum of all electrolysis

voltages and the total DC power supplied to the electrolyzers versus

DC current. The relationship between voltage and current is loga-

rithmic [91,92] . The electrolyzer ef ciency h EL was determined based on

J. Kotowicz et al. / Energy 118 (2017) 156e171

161

J. Kotowicz et al. / Energy 118 (2017) 156 e 171 161 _ Fig. 4. Diagram

_

Fig. 4. Diagram of tested hydrogen generator, where points of the measured values are: V e hydrogen volume ow rate, t 1 e electrolyzer temperature, t 2 e cooling water tem-

perature, p 1 e hydrogen pressure, p 2 e water pressure, A e current, V e voltage; DC e DC power source, AC e AC power source.

e voltage; DC e DC power source, AC e AC power source. F i g .

Fig. 5. Basic characteristics of the electrolyzers.

equation (8) , while the calculation of the hydrogen generator ef - ciency h HG takes into account only the auxiliary equipment power. Because the electrolyzers were supplied from an external DC source, losses in the AC/DC converters suppling electrolyzers were not taken into account. Based on the measurements, it was found that the power consumed by the auxiliary equipment of the hydrogen generator was constant, and at nominal capacity was 7.2% of the nominal power of hydrogen generator ( N AC ) nom . As the production of the hydrogen generator decreased, the auxiliary power system index d changed signi cantly, and at a load of 0,2 N nom , the auxiliary power system index was approximately 36%. Fig. 6 shows the ef ciency characteristics of the electrolyzer and hydrogen generator as a function of the AC power, referred to the nominal power. The gure shows also the change of the relative volume ow of hydrogen produced as a function of the relative power change. On the basis of these studies, it was found that for the range of measured capacity of the hydrogen generator, the ow rate of

capacity of the hydrogen generator, the fl ow rate of Fig. 6. Characteristics of the ef

Fig. 6. Characteristics of the ef ciency and capacity of a hydrogen generator.

hydrogen produced was linearly dependent on the supplied AC power. The ef ciency of the tested electrolyzer h EL increased with decreasing capacity; however, this relationship was not linear; it can be described by the equation:

h EL ¼ 0 : 49 $

nom 0 : 134

N

AC

ð N AC Þ

;

(14)

The equation has been t to the minimum of the square of the correlation coef cient, which in this case is 0.925. This relationship has been determined based on several mea- surement series, at which generated repeatability of measurements was about 2%. An important factor that affects on the obtained results is the temperature of the electrolysis process, which was maintained at the level 22 e 32 C. Operation at the higher tem- perature causes reduction of thermoneutral voltage V 0 and AC power N AC , and causes an increase of the electrolyzer ef ciency h EL .

162

J. Kotowicz et al. / Energy 118 (2017) 156e171

However, at the observed temperature uctuations, the voltage varies not more than 0.5%. Reached temperatures are relatively low for the PEM electrolyzers, what can also explain the fact of low

ef ciency. The second factor that affects on the obtained ef ciency

_

is the measurement accuracy of hydrogen ow-rate V H 2 . The used thermal mass ow meter is characterized by a accuracy of 0.5%, so the measurement accuracy of hydrogen ow-rate has no signi cant in uence on the overall shape of the ef ciency characteristic. The equation for the electrolyzer ef ciency can therefore be represented in general form using the coef cients a and m by the

relation:

h EL ¼

a

h

nom i m ;

N

AC

ð N AC Þ

(15)

The ef ciency h EL of the tested electrolyzer at nominal power was 49%, while at the lowest power it achieved better than 60%. This range of ef ciencies is consistent with values described for other small-capacity electrolyzers [93] . The low values obtained for the tested electrolyzer may result from the relatively high electrical losses in the electrolytic cell, which translates into a high supply voltage. Low ef ciency of the electrolyzer is characterized by low power units ( Fig. 3 ). Modern PEM electrolyzers are characterized by ef ciency above 70% at nominal capacity [73,77,79,89,92] . Most of the large power electrolyzers currently built are composed of many small capacity electrolyzer stacks which are connected in parallel [94 e 96] . This solution does not change the operating characteristics of a large unit in relation to a small electrolyzer, it is only scaled depending on the amount of used electrolyzer stacks. Therefore, the ef ciency of the electrolyzer is less dependent on the nominal capacity of this unit. While greater importance on the shape of ef ciency charac- teristics of the hydrogen generator will be auxiliary power index, which clearly decreases with increasing nominal power of the generator hydrogen. The correct determination of this index is possible when is known the own needs power of hydrogen generator. On the shape of the electrolyzer ef ciency characteristics affects the voltage ef ciency and the so-called Faraday ef ciency. The voltage ef ciency, de ned as the ratio of thermoneutral voltage V 0 to the voltage U supplied to the electrolyzer, is closely related to the current-voltage characteristics of the cell electrolyzer, and it de- pends on the occurring losses in the electrolyzer. Whereas the Faraday ef ciency, de ned as the ratio of the volume of gas pro- duced over a given time interval to the theoretical volume that should be produced during that time in accordance with Faraday's laws, is constant in a wide range of electrolyzer capacity and very close to 100%. Accordingly, a current-voltage characteristic has the decisive in uence on the ef ciency characteristics. General the dependence of voltage U supplied to the electrolyzer versus current is described by logarithmic function. Considering the formula on ef ciency h EL (8) and above information, theoretical dependence of ef ciency versus the AC power can be described as a power func- tion in a fairly wide range of electrolyzer capacity. The proposed relation (15) can be used in a range of the hydrogen generator nominal power from 20 to 100%. When the power is less than 20% of nominal power, the relation contain un- certainty, because of the high non-linearity of current-voltage characteristics and less Faraday ef ciency in this area. The ef ciency of the hydrogen generator is at least a few per- centage points lower than the ef ciency of electrolyzer and is almost constant over a fairly large power range. When the power supply is less than 30% of nominal power, ef ciency h HG falls rapidly, so the production of hydrogen becomes ineffective. The ef ciency of the hydrogen generator is proposed to depend on the

ef ciency of the electrolyzer and the auxiliary power index through equation (16) :

h HG ¼ h EL $ ð 1 dÞ ;

(16)

It should be noted that this is similar to describing the net ef- ciency of the power plant, i.e., using the gross ef ciency and auxiliary power index. If we assume N aux ¼ constant, the auxiliary power index of hydrogen generator can be written as:

d ¼ N aux ¼

N

AC

N aux

ð N AC Þ nom

$ ð N AC Þ nom

N

AC

¼

d nom $ ð N AC Þ nom

N AC

;

(17)

where: d nom e auxiliary power system index at nominal AC power. Substituting equations (15) and (17) into equation (16) yields:

h HG ¼

nom i m $ 1 d nom $ ð N AC Þ nom ;

a

h

N

AC

ð N AC Þ

N

AC

(18)

The effects of the parameters a , m and d nom on the hydrogen generator ef ciency h HG , are shown in Figs. 7 e 9 , according to equation (18) . To obtain the maximum value of ef ciency, we differentiate equation (18) :

v h HG

fi ciency, we differentiate equation (18) : v h HG ð N AC Þ nom ¼

ð

N

AC

Þ

nom

¼

a $ m m þ 1 þ a $ d nom $ ð m þ 1 Þ

h

h

nom i

N

AC

ð N AC Þ

nom i m þ 2

N

AC

ð N AC Þ

From the condition:

v h HG

ð N AC Þ nom ð N AC Þ nom

¼ 0 ;

;

(19)

(20)

the generator load relationship was calculated where the ef ciency h HG reaches an extremum:

N

AC

¼

ð N AC Þ nom

d nom $ m þ 1

m

;

(21)

Thus, the maximum value of ef ciency is equal to:

h HG ð max Þ ¼

a 1

m þ

1 ·

d nom m þ 1

m

m :

(22)

Þ ¼ a 1 m þ 1 · d nom m þ 1 m m :

Fig. 7. The ef ciency characteristics of a hydrogen generator (h HG ) for different values of coef cient a for m ¼ 0.2 and d nom ¼ 0.075.

J. Kotowicz et al. / Energy 118 (2017) 156e171

163

J. Kotowicz et al. / Energy 118 (2017) 156 e 171 163 Fig. 8. The ef

Fig. 8. The ef ciency characteristics of a hydrogen generator (h HG ) for different values of parameter m for a ¼ 0.725 and d nom ¼ 0.075.

m for a ¼ 0.725 and d n o m ¼ 0.075. Fig. 9. The ef

Fig. 9. The ef ciency characteristics of a hydrogen generator (h HG ) for different values of auxiliary power system ( d nom ) for a ¼ 0.725 and m ¼ 0.2.

4. The storage of hydrogen

depending on the location of the energy storage system, ranges widely from 5% to 20%. The concentration level of hydrogen in Polish conditions is 5%. The concentration limit may result from quality assurance of gas delivered to consumers or may be imposed by the technical requirements of components of the gas network such as pipes, compressors, etc. In addition to the maximum pressure, other basic design values for determining hydrogen tank capacity include the maximum storage temperature, the injection and out ow velocities, and the maximum quantity of hydrogen delivered to a gas network in a single operation cycle. Finally, a minimum pressure of storage is important in the determination of storage capacity. This pressure level depends on the minimum in- jection pressure of hydrogen into the gas pipeline, which in turn depends on the pressure in the natural gas transmission network.

5. A case study of a Power-to-Gas installation

Fig. 10 presents a diagram of a simple Power-to-Gas installation; it is supplied with electricity produced by production systems using renewable energy sources. Hydrogen injected into the gas grid is one nal product of its process. This system concept, the subject of analysis presented in this paper, is also of interest to agencies in Poland concerned with electricity and gas transmission and elec- tricity production. There is great interest in this subject, associated mainly with an increase the share of renewables in electricity production year over year. The structure of power-generation sources in the Polish power system, dominated by centralized systems based on coal with limited exibility, appears to support this technology. The advantage of the system shown in Fig. 10 is its simple structure, which includes devices currently available commercially. The basic elements of the system are the electro- lyzers and the storage tanks for the electrolysis products: hydrogen and oxygen. The hydrogen is directed to a gas grid pipeline, while the oxygen is transported away by tank trucks. The energy-storage system is equipped with a water treatment installation to ensure the quality of the water used in the electrolysis process. It is assumed that the system is integrated with a renewable power source whose generation capacity varies in time.

In Power-to-Gas technology, a hydrogen generation system is often integrated with hydrogen storage. Installations of energy storage can be classi ed as short- and long-term storage, depend- ing on their roles. Storage tanks, depending on the amount of gas stored, can be constructed as above-ground cylindrical or spherical tanks or can be underground. Natural formations, e.g., rock caverns created by extraction of gas or oil, coal or salt mines, can be adopted as underground tanks. The tank can also be caverns created spe- ci cally for this purpose by leaching salt deposits. Salt caverns require large nancial resources (construction of caverns of 750,000 cubic meter capacity is estimated to cost 20 e 30 M V ), but they are suitable for storage tanks because the salt is inert to the hydrogen. Because the hydrogen is stored only in gaseous form, the capacity of energy storage is linked to the storage pressure. For large salt caverns with a capacity of 500,000 m 3 , hydrogen can be stored at pressures in the range of 60 e180 bar. The highest pres- sures of hydrogen storage are achieved in above-ground tanks. Cylinder tanks used for hydrogen storage to supply fuel-cell vehi- cles achieve a high pressure of 700 bar. When integrating hydrogen production into the natural gas grid and the grid receives the produced hydrogen, the gas tanks in a Power-to-Gas installation are optional. In this option, the tank function is to keep the required amount of hydrogen injected into the natural gas pipeline. The ow rate of hydrogen can be deter- mined by the maximum acceptable hydrogen concentration in natural gas. In the literature, this maximum concentration,

5.1. Selection of nominal power of electrolyzers

The basic input data necessary for determining the nominal power of electrolyzers are the power characteristics of the RES installation as a function of time and the assumed operating times of the Power-to-Gas installation. The Power System is characterized by variable load throughout the day, resulting from the changing electricity demand. Within the day there are two main periods:

peak load and minimum load for system. Due to the speci city of the National Power System (NPS), it seems reasonable to power the electrolysis process during periods of lower energy demand (during the so-called valley of the night). This solution is also economically justi able because the electricity market price is lower in this period. According to the Polish Power Exchange yearly average price of electricity in the period from 10pm until 6am in 2015 amounted around 29 V /MWh. For comparison, the average price of electricity between 10am and 6pm was more than 56 V /MWh. Similar considerations, concerning supplying of the Power-to-Gas installation in the valley of the electricity demand, also apply in the systems with predominance of nuclear energy which are characterized by minimal exibility in load change [97,98] . The Polish power generation system based on coal is also in exible in this eld. Continuity of production in large centralized production systems is essential to maintain high national energy security. For the simultaneous development of RES the investments in energy storage systems are necessary [99] .

164

J. Kotowicz et al. / Energy 118 (2017) 156e171

164 J. Kotowicz et al. / Energy 118 (2017) 156 e 171 Fig. 10. Power-to-Gas installation.

Fig. 10. Power-to-Gas installation.

The basic decision in the design process of energy storage sys- tem is the determination of its size. The selected size of the storage system intended for cooperate with a renewable resource should enable to store a suf ciently large quantity of produced energy. Oversizing of the storage system, however, can carry to adverse economic consequences. For determination of size of the energy storage system, it is proposed to use the so-called degree of storage g s , which is de ned as the ratio of the annual amount of electricity directed to storing, E el_s , to the annual total amount of electricity generated in the RES installation, E el_a :

s

g s ¼ E el E el a

:

(23)

The hourly production values of a RES installation in the annual working cycle are the basis of the analysis. The production used in the present analysis was taken from a wind farm. The electricity production varies both daily and annually. The annual power characteristic of the RES installation is shown in Fig. 11. For sample input data, the assumptions were a 15% degree of storage and an 8- h work cycle of the storage installation in the night valley (between 10pm and 6am). With these assumptions, the nominal power of the hydrogen generators, calculated iteratively, was 5.96 MW. The nominal power may be the sum of the nominal powers of several devices, forming a battery. The power supply of the hydrogen

generators is additionally outlined in grey in Fig. 11. Electricity produced by the wind farm outside the night valley was directed to the electrical grid. Between 10pm and 6am, the electrolysis process is powered rst. If RES production exceeds the nominal power of the hydrogen generators, the surplus is directed into the power grid. Therefore, the input power of the installation is limited by the nominal power of hydrogen generator (5.96 MW) or the momen- tary potential of the wind farm. As the power of the wind farm varies, additional units can be turned on or off. The similar characteristics of a randomly-selected week of the RES installation are presented in Fig. 12 . The nigh valleys are marked by vertical dotted lines, while the production of the RES installation directed to Power-to-Gas is shown in grey. The over- capacity during night valleys and production outside these periods are directed to the power grid. The literature appears a number of other algorithms for sup- plying hydrogen generators in collaboration with renewable energy sources. In Ref. [100] the authors analyze the system with small- scale wind turbine supplying generator hydrogen in off-grid con guration and in grid-integrated mode and assume a constant power of electrolysis system. The excess electricity is connected to dump load or sold to the grid. In the case of less production of the turbine than the nominal power of electrolyzer it is switched off. In Ref. [101] a large scale wind-hydrogen plant is analyzed for the oil sands industry. The capacity of wind is 563 MW. The authors

sands industry. The capacity of wind is 563 MW. The authors Fig. 11. A sample annual

Fig. 11. A sample annual cycle characteristic of a power of RES installation with an outline of the input power of hydrogen generators (grey).

J. Kotowicz et al. / Energy 118 (2017) 156e171

165

J. Kotowicz et al. / Energy 118 (2017) 156 e 171 165 Fig. 12. The characteristics

Fig. 12. The characteristics of the power of RES installation in selected week with marked areas of production directed to the supply of electrolysis installation (grey) and to the power grid.

considered a few plant con guration with different models and number of electrolyzer units. The simulation includes 8760 h to yield and presents among others the annual hydrogen ow rate, annual sold electricity, capacity factor and economic results of the minimum hydrogen production cost. The indicator for the assessment of a technical energy storage system may be the rate of utilization of the working time of an installation d HG , de ned by the formula:

d HG ¼ t E

t

N ;

(24)

where: t E - number of hours of working electrolyzers of installation within a year, t N - number of hours provided for the operation of the installation (duration of the night valleys throughout the year, i.e., 365 8 h ¼ 2920 h). One of the criteria that could serve for evaluating the bene ts of hydrogen generators with a speci c nominal power is the utiliza- tion rate d HG of their nominal power, de ned by the equation:

d HG ¼ ð N HG Þ av

ð N HG Þ nom

;

(25)

where: ð N HG Þ av - average power of installed hydrogen generators achieved in the annual cycle, ð N HG Þ nom - nominal power of the installed generators. For the analyzed case study, the rate of utilization of working time of the installation, assuming a degree of storage of 15%, calculated in accordance with the formula (24) , was 92.2%.The utilization rate of nominal power calculated on the basis of the formula (25) was 74.1%. The nominal input power of electrolyzer installations increases with increasing degree of storage, while the utilization rate of nominal power decreases, which is illustrated by Fig. 13 . For the analyzed case, assuming the Power-to-Gas system operates during the night valleys, the maximum value of the degree of storage was 32.8%. It occurred when the power of the hydrogen generator was equal to the maximum night valley power achieved by the wind farm in the analyzed year. The utilization rate of nominal power de ned by equation (25) applies to the annual cycle. This indicator can also be applied to the daily cycle. Sorting the value of this rate for all days of the year allows making an ordered chart. This type of graph for seven selected degrees of storage is presented in Fig. 14 . Comparison of

degrees of storage is presented in Fig. 14 . Comparison of Fig. 13. The characteristics of

Fig. 13. The characteristics of nominal power of electrolyzers: installation and utili- zation rate of nominal power as a function of the degree of storage.

their characteristics shows that increasing the share of energy directed to the Power-to-Gas installation relative to total RES pro- duction decreases the utilization rate of nominal power of hydrogen generators. High degrees of storage are synonymous with an adverse decrease in the number of days per year in which spe- ci c values of utilization rates of nominal power are achieved by an installation. This is due to the high uctuations in the power from renewable energy sources, both in the daily and annual cycles.

5.2. Determination of the amount of hydrogen produced

On the bases of the calculated nominal input power of a Power- to-Gas installation and the assumed rated power of a single hydrogen generator, the number of devices can be determined.

n HG

ð N HG Þ nom

N HGj nom

n 2 N ;

(26)

where: n HG - number of electrolyzers of the installation, ð N HG Þ nom - the calculated nominal input power ð N HGj Þ nom - the nominal input power of a single hydrogen generator. For our example, the number of required devices was deter- mined assuming the nominal power of a single hydrogen generator of 1.5 MW. Similarly, the number of hydrogen tanks was selected to provide storage for the maximum daily production of hydrogen. The assumed capacity of the tank was 100 m 3 . The tanks were selected assuming that the maximum and minimum pressures of

166

J. Kotowicz et al. / Energy 118 (2017) 156e171

166 J. Kotowicz et al. / Energy 118 (2017) 156 e 171 Fig. 14. The ordered

Fig. 14. The ordered chart of daily utilization rate of nominal power of hydrogen generators for different degrees of storage.

hydrogen storage were 3.5 MPa and 1.6 MPa. The numbers of hydrogen generators and tanks as a function of the degree of storage are shown in Fig. 15 . Due to the variability in the power-on time of the RES installa- tion and in the amount of energy directed to storage, the load on one of the electrolyzers may be changed. The power input to the hydrogen generator system at a given moment is:

N HG ¼ ð n HG Þ nom $ N HGj nom þ N HGj D ;

(27)

where: n HG nom - the number of hydrogen generators working with nominal power ( ð n HG Þ nom n HG ), ð N HGj Þ nom - the nominal input power to a single hydrogen generator, ð N HGj Þ D - the input power of a hydrogen generator working under a load lower than nominal value (0 ð N HGj Þ D < ð N HGj Þ nom ). The ef ciency of a hydrogen generator working at partial load changes with input power. On this basis, the chemical energy ux of hydrogen produced at a given moment is determined by the formula:

_

E chH 2 ¼ ð h HG Þ nom $ ð n HG Þ nom $ N HGj nom þ

h HG $ N HGj D :

(28)

n o m þ h H G $ N H G j D : (28) Fig.

Fig. 15. The number of hydrogen generators and tanks of the installation as functions of the degree of storage.

The ef ciency of the generator at loads other than the nominal h HG changes with the load in accordance with (18). For analysis the authors adopted a ¼ 0.725; m ¼ 0.2 and d nom ¼ 0.075. The value a has been selected on the basis of the technical data of large power units, collected from several manufacturers [94 e 96,102,103] . Ef ciency of these hydrogen generators are uc- tuated around 70%. Therefore it is assumed that the newly designed system should be characterized by high ef ciency of 72.5% at nominal capacity of hydrogen generator. The ux of hydrogen chemical energy ð V H 2 Þ in m 3 n /s is then determined using the formula:

_

_

V H 2 ¼

_

E

chH 2

HHV Dr H 2

;

(29)

where: HHV e the higher heating value of hydrogen (141.8 MJ/kg),

and r H2 e hydrogen density in normal conditions (0.0899 kg/m 3 n ).

A graph of the maximum volume ow of hydrogen ð

_

V H 2 Þ max ,

received in system operation with the calculated nominal power, as a function of the degree of storage ( g s ), is shown in Fig. 16 .

The annual production volume of hydrogen ( V H2year ) can be

volume of hydrogen ( V H 2 y e a r ) can be Fig. 16.

Fig. 16. The characteristics of the maximum volume ow of hydrogen as a function of the degree of storage.

J. Kotowicz et al. / Energy 118 (2017) 156e171

167

determined by calculating the integral over a year:

8760

V H 2 year ¼ Z

0

_

V H 2 d t;

(30)

In the case of the rated input power of the installation ð N HG Þ nom of 5.96 MW and the input power of a single gas generator ð N HGj Þ nom of 1.5 MW, the installation has 4 hydrogen generators. The annual hydrogen production was 2 252 875 m 3 n . The relationship between the annual production and the degree of storage is shown in Fig. 17.

5.3. Economic analysis

Power production in Poland is still based on coal- and lignite- red power plants [19] . Simultaneously, sources of power genera- tion based on renewable resources have priority access to the po- wer network. Despite the positive environmental impact of promoting use of energy from renewable sources, this situation contributes to adverse impacts on the power system. The priority of the RES thrusts large conventional coal- red units into the role of marginal regulatory units that must adjust of their generation ca- pacity to the current demand for electricity. The dynamically changing relationship between the production capacity and the size of the electricity demand contributes to the need for frequent changes of load and shutdowns of coal- red plants, which adversely impacts their economic characteristics. This situation presents a serious risk of disrupting the energy security of the country. The power-system operator is obliged to create solutions, including prevention mechanisms. It is important that such mechanisms do not limit the potential of RES. Organization of market mechanisms to enable economically justi ed operation of energy storage systems can solve the problem of low elasticity of coal blocks while enabling further growth in installed capacity from renewable sources. The Power-to-Gas installation can be bene cial for the elec- tricity grid in a few elds. It can meet the regulatory function by receiving excess energy from the system. The energy storage, including Power-to-Gas may contribute to reducing the number of shutdowns of the coal- red plants and thus to increase their average ef ciency, reduce the frequency of repairs and reduce costs associated with restarting of units. In addition, the storage of energy can help to the unloading of power grids, which cooperate with the wind farm characterized by uctuation of electricity production. Thus, the modernization of the network to increase its capacity may become less costly. Natural gas with small amount of hydrogen (about 5% by vol- ume) can be successfully used in combined cycle gas turbines. The

can be successfully used in combined cycle gas turbines. The Fig. 17. The annual production of

Fig. 17. The annual production of hydrogen depending on the degree of storage.

natural gas even with hydrogen is very exible in terms of location, the time of demand or energy use [104] . The enriched gas also can be use in the autonomic gas turbines, working as a peak sources, which perform a regulatory function in the system. There is no doubt that Power-to-Gas technologies can be an important mechanism in preventing these adverse effects of the growth of RES. At the current stage of development of the tech- nology, Power-to-Gas is characterized by relatively low ef ciencies. Its main advantages are few requirements for built-up space or for site conditions (the suitable height differences or the geological conditions that are required, e.g., in storage systems such as CAES or pumped-storage power plants). At this writing, energy storage technologies are of interest to the potential bene ciaries, such as representatives of manufacturing industries, of the natural gas transmission system, and entities specializing in electric energy trading under the principles of safe and secure operation of the National Power System. Technologies with a high utilitarian po- tential during the market test phase have options of receiving nancial support. The technologies with high importance for na- tional energy security can be covered by future nancial support mechanisms, such as adequate tariffs or certi cates. Another mechanism put forward for discussion is a link between the amount of co- nancing of construction and the declared pro le of the power of unstable energy sources such as wind farms. It should be noted that in spite of great importance for the energy security the nancial support for storage systems is not guaranteed by law. Obtaining additional income required to economically justi ed operation of the energy storage systems is only possible by periodic provision of services for the operator of system. Investment risk in this case appears to be high. The results of the analysis, which was carried out in this section indicate that support mechanisms are essential to enable the development of energy storage systems. The economic analyses of the energy storage system were conducted, similarly to the technical analysis, for the full range of possible values of the degree of storage. Within the scope of the analyses, the break-even price of hydrogen was calculated from the condition:

NPV ¼ 0 :

(31)

This value was determined from the relationship:

b

e

C

H 2 ¼

P

t

t

¼ N

¼ 0

½ð J D Þþ K S O 2 þ T A L

ð 1 þ r Þ t

P

t

t

N

¼

¼ 0

V H 2 year $ HHV

ð 1 þ r Þ t

;

(32)

where: J - the investment cost, D - the size of the grants, K- the operating costs, S O2 - revenue from sales of oxygen, T - taxes (on income and property), A e depreciation, L - the liquidation value, r e the discount rate, t - another year of the analyses, from t ¼ 0 (the start of construction) to t ¼ N (the last year of operation). The most important assumptions used for the economic analysis were shown in Table 1. The results of the economic analysis for the three levels of project co- nancing of 0%, 50% and 100% are shown in Figs. 18 e 20 . To assess the economic ef ciency of this energy storage system, the values obtained for the break-even prices of hydrogen can be referred to the current market price of natural gas. In both cases, the price must be understood as a cost for the purchase of units of energy contained in the gas. The investment may be economically effective, assuming a current market price of natural gas of 8 V /GJ, only for the small storage systems ( g s < 0 : 17) with 100% nancing and free electricity supplied to the system during the off-peak periods. Note, however, that the hydrogen at its conversion into electricity is not burdened with carbon dioxide emission. Due to the

168

Table 1 Assumptions data for economic analysis.

J. Kotowicz et al. / Energy 118 (2017) 156e171

Parameter

Value

Grant variants Unit investment cost (for all system referred to the nominal power of hydrogen generators) Construction time Construction cost distributed Time of the system operation Discount rate The purchase price of electricity in off-peak demand period Price of water Operating cost Yearly maintenance costs

Price of oxygen Value of the depreciation rate Property tax

The liquidation value of the installation

0/50/100%

1350 V /kW

2 years

30/70%

20 years

5%

0/10/20/30 V /MWh 1.1 V /m 3 3.5 V /GJ H2 0.3% of the investment costs for the rst 4 years of operation, 0.6% for the next four years, 0.9% during next four years, 1.2% during the next four years and 1.5% for the last four years of operation 0.06 V /m 3 n

6.7%

2% of the capital expenditures incurred on the property while it is subject to depreciation, and 2% of the market value of the property on 1 January of the last year of depreciation 20% of the investment

of the last year of depreciation 20% of the investment Fig. 18. Break-even price of hydrogen,

Fig. 18. Break-even price of hydrogen, for four different electricity purchase prices in the off-peak demand period as a function of the degree of storage with 100% nancing.

a function of the degree of storage with 100% fi nancing. Fig. 19. Break-even price of

Fig. 19. Break-even price of hydrogen for four different electricity purchase prices in the off-peak demand period as a function of the degree of storage with 50% co- nancing.

existence Trading Emission System production of electricity from hydrogen fuel can be cheaper than from natural gas. In an extreme case of the analysis the price of electricity which occurs during the night valley was assumed equal to 0. It should however be consider the possibility of occurrence of negative prices, which can signi - cantly improve the economic ef ciency of investment. The

improve the economic ef fi ciency of investment. The Fig. 20. Break-even price of hydrogen for

Fig. 20. Break-even price of hydrogen for four different electricity purchase prices in the off-peak demand period as a function of the degree of storage in the absence of nancing.

appearance of negative prices within the markets of some Euro- pean countries contributed to the stimulation of interest in storage technologies. Simpli ed analysis was conducted for determined economic conditions - constant electricity prices during the night valleys and peak periods. In real market conditions, where elec- tricity prices are shaped by a number of macroeconomic factors, the use of large storage systems, which can allow for storage of energy over long periods, allows to conduct an optimal energy trade strategy and for this purpose the use of the mechanism of price arbitration. It is also to be noted that the further development of the Power-to-Gas technology and renewable energy sources can lead to a reduction in the investment costs. In literature [105] the lower cost levels in an 85% renewable energy scenario are also analyzed:

1000, 750 and 500 V /kW.

6. Summary

Currently, there has been a signi cant increase in the popularity of hydrogen production by electrolysis. Hydrogen generators with PEM and alkaline electrolyzers have experienced the greatest in- terest, as indicated by pilot installations. The greatest interest in storage systems is in countries that have experienced the most dynamic increase in installed capacity of wind or solar systems in recent years. The core function of storage systems should be sta- bilization of power systems while maintaining the full potential of

J. Kotowicz et al. / Energy 118 (2017) 156e171

169

renewable energy sources and providing reassurance of the effec- tive functioning of conventional generation. The important issue for the analyses of energy storage systems using electrolysis is the ability to determine the ef ciencies of hydrogen generators under changing loads. The literature often presents analyses focusing on electrolyzers but ignores auxiliary equipment such as pumps, fans, converters, transformers or control systems. Energy analyses taking these elements into account allow the determination of the potential ef ciency of the work of hydrogen generators belonging to the storage systems provided for cooperation with unstable energy sources. The experimental studies conducted by the present authors have shown that the total auxiliary power for a test hydrogen generator under nominal load constituted 7.2% of the power supply of the system. The reduction of load leads to increasing electrolyzer ef ciency. A constant level of auxiliary power ensures that the reduction of load corresponds to decreased ef ciency of the hydrogen generator. The auxiliary po- wer index at the load of 0.2( N HG ) nom is approximately 36%. Will- ingness to conduct analyses of the system, including the multivariate analyses to determine the energy ef ciency of storage systems, creates a need for suitable models of hydrogen generators. Through the generalized characteristics of the generator ef ciency proposed in this paper, the ef ciency can be determined for any load on the device. The proposed equation may constitute a basis for the analysis of energy storage systems cooperating with un- stable energy sources such as wind farms, where determination of the ef ciency of a hydrogen generator is needed in a whole range of work with changing input power. Having obtained parameteriza- tion, it is possible to match the form of equation to the character- istics of devices available on the market. A very important issue is the selection of the nominal power of hydrogen generators cooperating with a speci c RES installation. In the paper, this quantity was determined as a function of the degree of storage. The degree of storage was de ned as the ratio of the amount of electricity directed to storage to the total annual amount of electricity generated in RES installation. The authors proposed, as the criterion for evaluation of the legitimacy of installing hydrogen generators with a speci c nominal power, the utilization rate d HG of their nominal power. The paper shows that the nominal power of hydrogen generators, the utilization rate of their nominal power, and the economic characteristics of the Power-to-Gas installation are dependent on the degree of storage g s . The use of an energy storage system in Power-to-Gas technology certainly delivers positive effects in terms of the impact on the electricity system. This is mainly because it mitigates a possible adverse impact on the National Electricity System caused by changes of RES production. Although hydrogen generators are already available in the market, they are still characterized by relatively low ef ciencies. Despite its imperfections however, this technology could become more popular. Stimulating factors could be:

- the expected development of the technology of electrolyzers,

- the expected decrease in electrolyzer prices,

- the emergence of considerable decreases in energy prices in off- peak periods, which will continue to experience an increasing share of renewable energy in the structure of primary energy consumption,

- introduction of support mechanisms, that increase the exibility in terms of available power for the network (subsidies, tariffs, certi cates).

The results of this economic analysis con rm the general view that there is a need for the introduction of market mechanisms supporting energy storage systems.

Acknowledgement

This research was nanced within the project commissioned by the National Center for Research and Development under the GEKON program: " The energy storage in the form of hydrogen in salt caverns " , implemented in 2015 e 2016.

References

[16] Hong L, Lund H, M oller B. The importance of exible power plant operation for Jiangsu's wind integration. Energy 2012;41(1):499 e 507 . [17] Spicker S, Weber C. The future of the European electricity system and the impact of uctuating renewable energy e a scenario analysis. Energy Policy 2014;65:185 e 97 . [18] Atsonios K, Violidakis I, Sfetsioris K, Rakopoulos DC, Grammelis P, Kakaras E. Pre-dried lignite technology implementation in partial load/low demand cases for exibility enhancement. Energy 2016;96:427 e 36 . [19] Kami nski J, Kude ł ko M. The prospects for hard coal as a fuel for the Polish power sector. Energy Policy 2010;38:7939 e 50 . [20] Badyda K. Trends, conditions and prospects for the construction of new gas turbine combined cycle plants in Poland. Rynek Energii 2013;5:26 e 33 . [21] Wo ł owicz M, Badyda K. Gas turbine selection for feedwater repowering. J Power Technol 2015;95(4):302 e 8 . [22] Chmielniak T, Lepszy S, Czaja D. Economic assessment of gas-steam systems taking account of variable loads. J Power Technol 2015;95:54 e 62 . [23] Despr es J, Mima S, Kitous A, Criqui P, Hadjsaid N, Noirot I. Storage as a exibility option in power systems with high shares of variable renewable energy sources: a POLES-based analysis. Energy Econ 2016. http://dx.doi.org/ 10.1016/j.eneco.2016.03.006 . [24] Kondziella H, Bruckner T. Flexibility requirements of renewable energy based electricity systems e a review of research results and methodologies. Renew Sustain Energy Rev 2016;53:10 e 22 .

[15]

170

J. Kotowicz et al. / Energy 118 (2017) 156e171

and impact on storage demand in a European renewable power system of 2050 d sensitivity study. J Energy Storage 2016;6:1 e 10 . [30] Bouman EA, Øberg MM, Hartwich EG. Environmental impacts of balancing offshore wind power with compressed air energy storage (CAES). Energy 2016;95:91 e 8 . [31] Kotowicz J, Jurczyk M. Ef ciency of diabatic CAES installation. Rynek Energii 2015;119(4):49 e 56 . [32] Chen H, Cong TN, Yang W, Tan C, Li Y, Ding Y. Progress in electricity energy storage system: a critical review. Prog Nat Sci 2009;19:291 e 312 . [33] Luo X, Wang J, Dooner M, Clarke J. Overview of current development in electrical energy storage technologies and the application potential in power system operations. Appl Energy 2015;137:511 e 36 . [34] Lund H, Salgi G, Elmegaard B, Andersen AN. Optimal operation strategies of compressed air energy storage (CAES) on electricity spot markets with uctuating prices. Appl Therm Eng 2009;29:799 e 806 . [35] Auer H, Haas R. On integrating large shares of variable renewables into the electricity system. Energy 2016. http://dx.doi.org/10.1016/ j.energy.2016.05.067 . [36] He H, Chang M, Chen A, Ma L, Lu J, Li C. Enhanced hydrogen evolution process of water electrolysis assisted by photocatalysis. Materials for renewable energy & environment (ICMREE). In: International conference, Shanghai, 20 e 22 May 2011; 2011. http://dx.doi.org/10.1109/ ICMREE.2011.5930608 . [37] Jiang W, Wu YK, Yang T, Yu FY, Wang W. Identi cation and Power Electronic Module Design of a solar powered hydrogen electrolyzer. In: Asia-paci c power and energy engineering conference; Shanghai, 27 e 29 March 2012;

2012. http://dx.doi.org/10.1109/APPEEC.2012.6307202 . [38] Paska J, Pawlak K, Surma T. Support systems as an essential element of the impact optimization of new, green sources of electricity on the power sys- tem. Rynek Energii 2013;2(105):48 e 53 . [39] Leading the energy transition. Fact book: electricity storage. SBC Energy Institute; September 2013 . [40] Mirek P, Nowak W. OZE w systemie i co dalej? Potencja ł i ewolucja uk ł ad ow magazynowania energii cz. 1. Energ Ciepl i Zawodowa 2015;4:8 e 12 . [41] Mirek P, Nowak W. W ogonie za Stanami. Potencja ł i ewolucja uk ł ad ow magazynowania energii cz. 2. Energ Ciepl i Zawodowa 2015;5:64 e 6 . [42] Zoulias E, Varkaraki E, Lymberopoulos N, Christodoulou CN, Karagiorgis GN. A review on water electrolysis. Project report Cluster pilot project for the integration of renewable energy sources into European energy sectors using hydrogen . Pikermi, Greece: Centre for Renewable Energy Sources (CRES); 2006 . [43] Bhattacharyya S., Cobben S., Consequences of poor power quality e an overview, Power quality, Andreas Eberhard (Ed.), InTech, ISBN: 978-953- 307-180-0, Available from: http://www.intechopen.com/books/power- quality/consequences-of-poor-power-quality-an-overview . [44] We˛ cel D, Ogulewicz W, Kotowicz J, Jurczyk M. Dynamics of electrolysers operation during hydrogen production. Rynek Energii 2016;122(1):59 e 65 . [45] Keatley P. Cost modelling of coal power plant start-up in cyclical operation. Coal Power Plant Materials and Life Assessment. Elsevier; 2014. p. 358 e 88 . [46] Hentschel J, Babi c U, Spliethoff H. A parametric approach for the valuation of power plant exibility options. Energy Rep 2016;2:40 e 7 . [47] van den Bergh K, Delarue E. Cycling of conventional power plants: technical limits and actual costs. Energy Convers Manag 2015;97:70 e 7 . [48] Singh S, Jain S, Venkateswaran PS, Tiwari AK, Nouni MR, Pandey JK, et al. Hydrogen: a sustainable fuel for future of the transport sector. Renew Sus- tain Energy Rev 2015;51:623 e 33 . [49] Taamallah S, Vogiatzaki K, Alzahrani FM, Mokheimer EMA, Habib MA, Ghoniem AF. Fuel exibility, stability and emissions in premixed hydrogen- richgas turbine combustion: technology, fundamentals, and numerical sim- ulations. Appl Energy 2015;154:1020 e 47 . [50] Chiesa P, Lozza G, Mazzocchi L. Using hydrogen as gas turbine fuel. J Eng Gas Turbines Power 2005;127(1):73 e 80 . [51] Ulleberg Ø, Nakken T, Ete A. The wind/hydrogen demonstration system at Utsira in Norway: evaluation of system performance using operational data and updated hydrogen energy system modeling tools. Int J Hydrogen Energy 2010;35:1841 e 52 . [52] Gassner M, Mar echal F. Thermo-economic optimisation of the integration of electrolysis in synthetic natural gas production from Wood. Energy 2008;33:

189 e 98 .

[61]

82 e 7 .

[65]

~

[71]

[73] Tsiplakides D., PEM water electrolysis fundamentals. Available: http://

J. Kotowicz et al. / Energy 118 (2017) 156e171

171

[97]

[100]

48 e 56 .

[102] Power-to-Gas: how to convert green power into green gas via water elec- trolysis, Denis Thomas, Hydrogenics. Green Gas Research Outlook Sweden; 2015. http://www.hydrogenics.com/home . [103] SILYZER 200 High-pressure ef ciency in the megawatt, Siemens, ran- ge http://www.industry.siemens.com/topics/global/en/pem-electrolyzer/ silyzer/Pages/silyzer.aspx . [104] Walker SB, Mukherjee U, Fowler M, Elkamel A. Benchmarking and selection of Power-to-Gas utilizing electrolytic hydrogen as an energy storage alter- native. Int J Hydrogen Energy 2016;41:7717 e 31 .

254 e 61 .