Integration of small and micro combined heat and power (CHP) systems into distributed energy systems
Abstract:
During the last decades electrical systems have undergone a transformation, which is still underway and is supposed to lead from the ‘vertical integration’ model to a fully open electricity market. This process presently overlaps with the development of distributed energy resources (DER) that brings under a single concept electricity generation, storage of energy and demand response, all at small scale. DER development raises a number of issues for the different stakeholders. However, it is obviously an opportunity for new, emerging market players, but it brings also new perspectives for incumbents. Participation of DER in the different energy markets requires the setting up of upgraded structures for aggregating small energy sources. This chapter summarizes why dispersed generation is innovating in the electrical power system context. It shows how distributed generation can represent value for the system and it evaluates the significant economic advantage of being interconnected. It concludes with recommendations to wire companies and to regulatory bodies.
4.1 Distributed energy resources (DER)
4.1.1 Initial developments in power systems
Before addressing the issue of integration of DER in the power system, it is useful to recall briefly the development of the production of mechanical energy during the nineteenth century as well as the transition to electricity as a multi-purpose energy resource.
From the beginning of the industrial revolution, with the introduction of the steam engine, ‘power generation’ meant ‘centralization’. This was the result of the physical and technical principles that lie behind these complex processes and the ‘hierarchy’ of the different forms of energy. These physical principles are still present today and have induced a strong opposition between production, on the one hand, and consumption, on the other. Historically the steam machine was built in front of each workshop where the mechanical power generated was distributed to different machines using a system consisting of transmission shafts, pulleys and belts. The size of the production machine was already significantly larger than the average power consumer process in the workshop. Therefore, ‘centralization’ of power generation existed prior to the birth of the electricity sector.
When electricity development started, things did not change immediately. The structure of the distribution of the mechanical power in workshops remained more or less as it was, partly because electrical machines were extremely costly. Thus the steam engine was replaced by a single electric motor. However, a certain concentration took place upstream, because it was no longer necessary to place generation and utilization of energy side by side. The complexity of the energy transformation process leading to electricity production provoked the development of larger factories. The flexibility of the transmission and distribution of electricity, particularly using alternating current, allowed the increasing separation of power generation and the various industrial processes. Within the factory, the electricity generating units were installed in the same building, called the ‘power plant’, in French ‘la centrale électrique’, an expression where the notion of centralization is obvious. The security of supply for the different workshops was strengthened as they were fed by all these units operating in parallel rather than being dependent on a single local steam engine. This shows that electricity is inherently an activity organized about the network. The next step consisted in the development of the connection between neighbouring factories to ensure the sharing of reserves. This permitted the same level of reliability to be reached while limiting the cost of development of the system. The principles that justify network interconnection were already at work. This is due to the relatively lower cost of high voltage (HV) networks compared to the cost of power plants.
It is important to note that distributed generation (DG), which is a component of DER, is not a return to a solution or situation from the past. It is really a new era for electrical systems. Indeed, for the first time since the beginning of the electricity generating industry, one can consider that for certain applications and certain primary energy sources, the production of electricity by very small units could become profitable in the short term. The DG label is less a question of size than a matter of ratio. Indeed a generation unit of 10 MW installed in a large interconnected system can be considered as a decentralized generation, whereas the same unit installed in a system of 100 MW of peak power, would probably be considered as ‘centralized’.
Soon it will be possible to consider combined production of heat and electricity (cogeneration) facilities designed for being installed in single family houses and leading to profitable operation without any incentives beyond ‘green certificates’. Within 10 to 20 years it may also be the case for photovoltaic conversion. Never in the past has it been possible to use energy conversion processes of such a small size. Today, it is becoming possible to generate electricity with generators having a size which is comparable to the size of load components.
4.1.2 Reliability of supply
The economic value of local generation is highest when it is interconnected with the network. This is a direct consequence of the fundamental properties of any electrical system requesting, among other things, the real-time adjustment of the balance between production and consumption both for active and reactive powers. This requires specific technical solutions that the liberalization of the energy market has made more complex.
Reliability of supply does not play the central role it played under the ‘vertical integration’ paradigm for the planning of generation. Most often in the present situation of the electricity market there is no longer any clear standard for setting the requested over-capacity in terms of generation which should be available to ensure the reliability of supply objectives. At best it is developed at generation company level. The need to build a new plant is only determined by the market.1 Such a position, however, is not absolute. In the United States, PJM, for example, checks on a yearly basis the evolution of future performance in terms of reliability.2
In the description that follows, to simplify the approach and as a first approximation, the system will be considered using the ‘vertical integration’ paradigm. The reliability of supply consists of two complementary aspects: the adequacy on the one hand and security on the other. The first, adequacy, means that the system is able to generate and transmit power from generating plants to the load for a set of standard situations including ‘normal’ and ‘abnormal’ situations which must also be carefully defined. ‘Abnormal’ means that the system is weakened due to the unavailability of some of its elements, whether these be the result of maintenance activities or the occurrence of sudden unexpected ‘events’, hence the concept of ‘secured events’.3 The second, security, means that the system must be robust for guaranteeing a stable operation when faced with these secured events. This robustness is based on the concept of preventive safety margins that must be respected during system operation. In the case of more serious incidents, the preservation of the integrity of the power system involves specific operational procedures and also the deployment of automatic countermeasures known as defence plans.
This decomposition of the reliability concept into adequacy and security aspects is valid independently of the size of the network. Ensuring the reliability of any system involves, therefore, specific means allowing for the adjustment in the short, medium and long terms of the active and the reactive power balance. Indeed, voltage and frequency stability depends highly on the balance between production and consumption. The network physically aggregates loads and production. The safe operation of the electrical power system assumes a ‘sufficient control’ of generation. This is implemented through the ‘aggregation’ of the various generating plants by means of communication channels (this is an extension of the earlier practice, but now considering smaller units) or it is based on ‘statistical’ principles (micro cogeneration status can be stochastically assessed according to weather forecasts for example at lower cost). Furthermore, demand response can be included as an innovative manner of stabilizing system operation.
4.2 The value of distributed generation
Successful integration of distributed generation can result from consistent and progressive actions that consider the technical aspects first, then the market architecture and of course the associated regulatory framework. The methodology that was developed as part of the EU-DEEP research project4 started from a simple assumption: most of the physical properties of electrical systems cannot be circumvented; as a result, the in-depth analysis of system behaviour is the route to the identification of solutions that can be effectively deployed. This then allows the selection of the most effective market mechanisms. In particular it is fundamental to propose solutions that are able to reveal the actual value that DER can represent for the network, and for the system beyond the potential gains that it allows in terms of externalities.
However, the ‘sustainable’ economy of distributed generation, without incentives, has not yet been demonstrated. According to new actors, distributed generation has many advantages which should be remunerated. For traditional players, especially distribution system operators (DSO), the connection of distributed generation raises many questions, such as voltage setting, operation of protection, risks related to island operation (‘anti-islanding’ protection) and the increase in short-circuit power. A lack of clarity results also from the appellation ‘distribution network’. In some cases it means medium and low voltage networks, but for others it means a network from as high a voltage as 132 kV down to the low voltage network. But this can also be a consequence of the attitude of some of the market players who are not willing to play this new game.
4.2.1 Technical aspects
In fact, most of the issues relating to distributed generation can be solved without systematically causing additional costs for the network, especially for production that does not violate the design criteria of the distribution network.
The assessment that the ‘value’ DER represents for the network is based on the comparison of islanded and interconnected situations. Fair comparison of solutions supposes similar performance in terms of actual reliability of supply and comparable technical and economic contexts. For example, an energy balance between generation and consumption set up on an annual basis has no meaning, when considering the operation of the electrical power system.
Determining the advantages or disadvantages of DER requires the application of specific methodologies for the assessment of costs and benefits of various options under consideration. These cost-benefit studies should integrate all positive and negative consequences, in particular those related to the distribution networks.
Making the approach more relevant requires taking into account the increasing penetration rates of DER. This helps distinguish local issues typically related to the distribution network which appear first, then systemic issues that concern the system as a whole. This leads to investigations about power system control in normal as well as in abnormal situations. This includes the impact of the extremely high penetration proportion of DER on the behaviour of the system under emergency conditions that require significant technical upgrades.
4.2.2 The value for the system
Three main issues must be addressed: how to get DER into the wholesale energy market, how to build upgraded use of system charge schemes and rules for transmission and distribution companies, and finally how to determine support and incentive programmes, if they are deemed useful for society or absolutely necessary to initiate a promising technology.
The participation of DER in the wholesale market means taking part in the primary markets, spot market as well as ancillary services market, but also to secondary markets, forward markets as well as to various hedging instruments. Due to the small size of DER units, the accession to these markets requires new solutions, including aggregation, to allow them to reach the required scale.
The additional value of DER is essentially related to its position in the system.5 Time interval metering of load and generation, associated with appropriate treatment (‘profiling’ individual contribution or via ex post detailed analyses) would allow the determination of the footprint of the local generation or local consumption on the network. Similar charging solutions already exist for transmission costs, the compensation of losses and for congestion management, such as the ‘ competitive locational price’. Another example is the ‘TRIAD’ concept as applied by NGET in the UK.
4.2.3 Market regulation
The profitable integration of DER remains questionable for the smallest ones. Indeed economies of scale, here expressed in terms of mass production, do not seem sufficient for the time being to lead to competitive generation costs compared to traditional generating units.
Presently, the price of electricity for low voltage customers includes two main components: the energy price and ‘use of network’ charges. These latter are obviously maximum at low voltage, as all layers of the network are playing a role. If DER is operating at the right moment, that is to say during peak consumption in its neighbourhood, it represents a significant value for the network. New pricing methodologies mentioned above are able to remunerate local generation in a network that is dominated by the load and vice versa.5,6 The sustainable economy of DER is closely linked to the regulatory framework. Innovative regulatory environments may limit in the long term the cost of distribution networks.
Distribution of electrical energy is expensive. A minimum density of customer is required to justify the development of a network. However, as soon as a client is connected to a network it has, as a client, additional benefits as a producer as well as a consumer. The price of the network, paid through ‘use of system charges’ mechanisms is certainly high, but still generally highly competitive when compared to islanded operation. The next section is an attempt to evaluate this additional value.
4.3 Conditions for profitable decentralized generation
Sustainable development of DG, combined heat and power units in the short term and renewable energy units in the longer term means the adequate optimization of the different components of the installation, but also of the interaction with the external system. Indeed the present chapter shows that islanding operation cannot be profitable except in regions of very low density of customers.
Cogeneration of heat and power is most often the best approach for increasing the overall efficiency of a plant and for reducing the consumption of primary energy. But this is seldom true as far as the profitability of the installation is concerred. This is due in large part to the capital intensive character of electricity generation. Furthermore, fairly long amortization periods are usually considered necessary in the electrical supply industry. For example, combined cycle plants based on combustion turbines, heat recovery steam generators and condensing steam turbines are amortized in 15 years, classical coal units are amortized in 20 years and this could be even longer for new nuclear power plants as these are now built for a 60-year operational lifetime. This has to be compared with other branches of the industry where profitability is expected in a fairly shorter period of about 3-5 years.
Reducing generating costs means taking the right action on various parameters that must be optimized as follows:
• the initial investment must be kept as low as possible;
• the overall efficiency of the installation must be high enough;
• the efficiency of electricity generation is less important than the overall efficiency and the reduction of initial investment has to be prioritized;
• a good design based on heat demand is the main objective and the minimum demand for heat is particularly important;
• the installation must be viewed in connection with the electricity network and the electricity power system.
For small or micro-units, the scale effect that led to multi MW power plants up to 1600 MW turbo-generators, becomes the mass production in fully automatic factories. The design of the unit must be such that overall efficiency is kept as high as possible at an acceptable cost. As an example, Stirling conversion units that until recently were characterized by fairly low efficiencies can now be designed with an overall efficiency of about 40%7. However, such high figures can only be obtained at very high costs that are only compatible with the space industry. Lower efficiency machines using equivalent technical solutions, like free-piston and oscillating permanent magnet generator with efficiencies about 20% can be used for making cogeneration boilers.
The additional cost required to turn a boiler into a micro-CHP must be kept as low as possible and the marginal cost of each additional efficiency point for electricity generation must be carefully checked. Present condensing boilers are characterized by a nominal efficiency of about 108% (base: low heating value of the fuel) and by a seasonal efficiency of about 90%. The part of the investment within the cost of heat is extremely low (for large industrial boilers it is of the order of 1%).
It is of utmost importance to keep the overall efficiency as high as possible. In fact due to the increased complexity of the conversion process, including the increased weight of the boiler, it is generally not possible to maintain the efficiency figures of conventional boilers. It is nevertheless possible today to reach a nominal efficiency slightly above 100% with Stirling boilers. It is here important to note the advantage of the Stirling conversion: its ability to use various primary fuels such as natural gas, fuel-oil and even wood pellets. The degradation of the overall efficiency with micro-turbines can be significant, mainly for smaller machines. This is due to the considerable mechanical difficulty of reducing the clearance between fixed and rotating parts of the machines. This is definitely a disadvantage for the smaller machines. The aforementioned characteristics are to be tuned at the design stage of the generating unit with the best possible balance between reliability, efficiency and cost.
The next step concerns the optimization of the installation. The case of industrial sites, such as the chemical industry, will be considered as an example. Such installations are generally supposed to feed heat at different pressures. These installations are most often designed as follows. The cogeneration unit is designed based on the minimum heat requirement, corresponding to summer conditions. The additional supply of heat is generated by classical boilers. When natural gas is used as primary energy, the initial conversion takes place in a combustion unit coupled to a synchronous machine which generates electricity. The heat content of exhaust gases is converted into steam in a heat recovery steam generator which is often equipped with additional burners that are sometimes able to be switched to autonomous operation in case of a trip of the combustion turbine (this requires a fresh air supply with auxiliary fans). When heat must be supplied to different processes under different pressure conditions, steam turbines are often used, backpressure units or condensing units with extraction of steam located at different stages of the steam expansion. These steam turbines are also coupled to a synchronous, sometimes asynchronous generator, supplying electricity to the industrial steam network. To keep the reliability of the installation at an acceptable level, but at as low investment cost as possible, pressure reducers are installed in parallel and are operated in case of steam turbine trips. Relief valves are also used at different pressure levels in order to keep the required minimum flow in the plant in case of process trip for making the operation of pressure reducers stable. Usually cogeneration plants are designed to operate at full load for about 8760 hours per year, excluding maintenance period.
For micro cogeneration, which can be viewed as the last cogeneration segment of customers that are not yet equipped, things are a little bit more complicated because the equivalent full load duration is most often not high enough (about 3500 hours of full load equivalent period for micro-CHP in the UK).
Figure 4.1 gives the evolution of the mean cost of electricity as a function of the full load duration, from 1000 to 8760 hours per year. The installation is supposed to operate for 15 years with maintenance costs corresponding to 5% per year (this is practically equivalent to three majors overhauls, each of them corresponding to one-third of the initial cost, which is usually the case for internal combustion machines). Natural gas price is €50/MWh, electrical efficiency 35%, total efficiency 85%, competing boiler yearly efficiency 90%, cogeneration investment €2000/kW, and the selected discount rate is 8%.
For domestic customers, an optimized design, based on the minimum demand of heat, would lead to a fairly limited installed power for heating the daily consumption of hot water. For example, for a family of two adults and two children, 1.5 kWh output for heat is sufficient. This is significantly lower than the 6 kWh of heat output of future Stirling boilers. This figure is most often the result of another approach which is based on the electrical side of the installation. Indeed the objective is to feed the home in electricity for compensating approximately the yearly consumption (mean value estimated at about 3500 kWh per year in England), hence the power of the generator of about 1 kWe. This power is lower than the equivalent peak consumption of a domestic customer; consequently, at first sight, it cannot be harmful for the network. This means that the design of the network is not seriously compromised, it would seem.
4.4 Evaluating the ‘full value’ of being network connected
What is the ‘full value’ of being interconnected to the network? A methodological approach has been set up to answer this question in EU- DEEP Deliverable D11.8 This value is in relation to the large flexibility for balancing load and generation when a site is connected to the network. Indeed the short-term power peak can reach the maximum subscribed power of the site (e.g., up to about 15-20 kW for domestic clients) whereas the peak consumption for the network, considering diversity, is only about 1.5 kW (at system level) or 3 kW (at distribution transformer level) for mean European domestic customers (assuming that electricity is not used for house heating).
The basic assumption is to evaluate the performance of the installation presenting an equivalent reliability of supply. The reference is given by the network performance (valid for the region where the considered site is located). Then a simplified model of the site is set up, with the site operating as an island. The size of the machine is selected for being able to feed the load. The demand is supposed to be given by interval metering. The ‘time window’ could be a quarter of an hour, half an hour, etc. The number of machines operating in parallel is selected to achieve the required performance in terms of reliability of supply. The full operational details are not necessarily taken into account as orders of magnitude only are considered.
The present worth evaluation for the whole expected operational life of the installation is calculating taking into account different assumptions (investment, operational and maintenance costs, discount rate, utilization of the plant, etc.). This permits the determination of the expected mean costs of electricity.9 The comparison of this cost with the ‘real’ price of electricity, including network costs, gives an evaluation of the ‘true value’ of being network connected.
Given the experience of different experiments that have been implemented within the EU-DEEP project, it is not necessarily straightforward to develop this type of ‘equivalent’ installation. Some adjustments must be made to make this comparison possible. This is particularly the case for domestic installations. This is essentially due to the lack of diversity when a single installation is considered, the difference between load and generation functions (load and generation curves) and finally due to the the large instantaneous variations that usually characterize load behaviour. Technical performance during operation is evaluated by determining the dynamic behaviour of the site disconnected from the network, based on measurements made in the field using sufficiently high sample rates (e.g., up to 50 samples per second).
The methodology has been used for determining an order of magnitude of the ‘full value’ of being connected to the distribution system. It corresponds to the difference between the total cost of electricity supplied by the system and the mean generating cost of electricity produced locally, in islanding. The determination has been made assuming that the islanded installation presents the same performance as the distribution network in terms of overall reliability of supply. In the presented examples the mean system reliability has been set at about 99.99 to 99.995, which corresponds to the performance of a network presenting ‘good reliability’.
This methodology has been used to develop five cases based on the tests implemented within the EU-DEEP project. Two are based on ‘market segments’ experiments implemented in Grenoble (internal combustion cogeneration plant with islanded capabilities - two variants, one with three or four generators and another one with a combination of UPS, CHP and one or two backup generators) and in Athens (a trigeneration plant based on a micro-turbine with storage and two variants with respectively 100% and 50% loading factor). The three additional cases have been selected to complete the picture: one ‘big’ site characterized by a high diversity of demand (Kapodistrian University of Athens) and two cases inspired from the domestic customer installations in the Berlin test where 10 micro-CHP systems have been installed and remotely operated for one year. However, the specific issues of islanded operation for such sites require characteristics that are not found in the Berlin site tests. Therefore two fictitious cases have been built. The first one considers a CHP installation, based on an efficient fuel cell able to work marginally in open cycle. The second one considers a combination of photovoltaic (PV) and cogeneration with different PV investments. This gives a large diversity of situations allowing for extrapolating to futuristic plants.
The results are summarized in Figs 4.2 and 4.3. Figure 4.2 considers the first sites corresponding to generation of some hundreds of kW up to MW. Costs figures are given assuming two values corresponding to two different levels of reliability of supply. For each case the upper value (square) corresponds to the installation fulfilling the reliability performance; the lower value (diamond) corresponds to the second best solution in terms of reliability. This permits a range of costs to be set up for which the performances are ‘acceptable’. In general the second site corresponds to the same installation where one generator has been removed. This is the case for the Grenoble SM (synchronous machine), Grenoble UPS with one or two backup synchronous generators and for the NTUA inspired tests. For Kapodistrian University (NKUA) two different installations have been considered; the first one is characterized by 21 units (defined by the strict application of the reliability calculation) and the second one by 12 units (this is more in line with practical approaches). The generation costs are in that case determined for CHP as well as for electricity generation.
Figure 4.3 summarizes the results for the domestic customer installations. The solid oxide fuel cell (SOFC) case includes the cost reductions that are expected for the near future, whereas the renewable energy source and combined heat and power solution (RES-CHP) presents results assuming the extension of an already existing installation. Both these installations are marginally able to supply the electrical energy with the reliability objectives set up initially.
These figures must be compared with the cost of electricity delivered by the power system: for example, in the low voltage network, €160-200/MWh (taxes included), where ‘use of system’ charges represent about 45 to 50%. This indicates an advantage when compared to islanded operation of the order of about €200/MWh minimum, except for the case of NTUA characterized by 100% utilization, which is not fully realistic when assuming islanded operation.
4.5 Recommendations to distribution system operators (DSO) and regulators
Autonomous operation of small size sites is generally not competitive when compared to the network connection. Hence, in general, DER will be connected to the distribution network. The rational integration in the system supposes fully open collaboration between DSO and regulatory bodies. Technical analyses have shown that efficient and sustainable solutions exist. They assume first that new designs have been defined, and second that adequate regulatory frameworks have been deployed. Both these aspects are summarized below as recommendations for DSO and regulators.
4.5.1 New designs for distribution networks
Context
Increasing the proportion of DER in the distribution network can have significant impacts on the network infrastructure; the type and the size of the considered DG are of utmost importance. DER has the potential to deliver services to DSOs and transmission system operators (TSOs) via entities aggregating multiple small resources dispersed in the distribution network. It is expected that these opportunities will increase in the future when medium voltage (MV) and low voltage (LV) distribution networks lose their unconditional adequacy (the ‘fit and forget’ principle being no longer fulfilled).
For the sake of clarity it is important to distinguish ‘thick’ from ‘thin’ distribution networks (respectively, operating HV-MV-LV or MV-LV grids only). In ‘thick’ distribution systems the range of services that can be delivered today or in the short term is broader. In ‘thin’ distribution systems, the range of services that can be delivered is limited to balancing services for TSOs. Future services that could be delivered by DER in MV and LV distribution networks are related to limited contribution to voltage control or reactive power compensation, and not to power flow management. For the more critical constraints, such as voltage control in ‘N-1’ situations, the corrective action corresponds rather to generation curtailment than to active management of sources. Active management is an appealing solution but it supposes the existence of control margins, otherwise it means limitation to DER output. Design criteria for distribution networks might be upgraded if limitations to DER output cannot be accepted.
Increasing the ‘DER hosting capacity’ of the network, whether or not active management is used, requires the setting up of new design criteria for developing or exploiting distribution networks. This is necessary because margins are needed if reduction in power injection is to be avoided as much as possible. This means that ‘exogenous’ objectives are necessary for fixing limits: limitation of peak generation for each customer in connection with peak load (diversity included), objective in terms of penetration for DER, limits set to generation control, etc. The impact of an increased number of DER on the cost of the system depends on the types of DER considered and on the network where they are connected. The additional investment costs due to DER integration depend on energy policy choices, including the associated operational rules that are imposed, including the possibility to control power injection in case of contingencies. ‘Exogenous’ objectives, in close connection with the more general objectives of energy policies, are necessary for fixing these targets: limitation of DER generation power per connection, objective in terms of penetration for DER, limits set for generation control in normal and abnormal conditions such as fault handling and DER’s ‘fault-ride-through’ capability, etc. This would lead to extensions to grid codes to comprise equipment with lower power ratings than today.
New design
New design criteria for distribution networks can easily be developed as soon as clear objectives are defined. These objectives should be defined outside of the electrical supply industry, but with its participation. The exact sharing between design upgrades and active management is an integral part of the process. The key aspects here are the existing voltage control margins, homogeneity of the feeders resulting in similar voltage behaviour, DER characteristics, its size, its design and the relationship between load and local generation (i.e. coincidence between the peak generation and peak demand or vice versa). As a result, when determining the ‘hosting capacity’ (the proportion of DER that can be operated in the system without inconvenience), it is important to use the network design criteria as the reference, but also to consider the specificities of the considered DER in connection with the local conditions prevailing in the network.
Distribution networks developed using the traditional ‘fit-and-forget’ principle often exhibit operational margins that allow for accepting a significant proportion of DER, particularly in urban and semi-urban networks. The extension of the ‘fit-and-forget’ principle should be based on the ‘reinterpretation’ of network design criteria. As voltage drops along feeders, the voltage set point is traditionally adjusted near to the upper limit at the feeding HV -MV substation, allowing at the same time system losses to be reduced. In the presence of local generation and in the case of non-homogeneous location of load and generation along feeders, voltage profiles can increase and decrease along the various feeders depending on the coincidence between load and generation on them. The reference voltage must therefore be adjusted downwards which allows the ‘fit and forget’ approach to be preserved. This supposes, however, sufficient regularity in terms of behaviour of the load and generation customers.
Existing margins can be expanded using active voltage management when a certain degree of homogeneity exists between feeders, or even more when considering active management of individual DER sources. When the voltage margins following the ‘fit-and-forget’ principle are insufficient, one needs to dynamically change the voltage control settings in the HV-MV substation. However, this supposes that the load shapes of the different feeders exhibit similar voltage characteristics as significant voltage lack of homogeneity can push the voltage outside of the acceptable range. Due to this lack of homogeneity, or if the DER installations lay outside of the network design rules, i.e. the size of the DER installations is not at all in line with the mean demand of the clients in the neighbourhood, there is a risk of technical issues such as over-voltages or overload of network elements. This implies that either penetration ratios should be reduced in these types of networks, or active management of sources is needed, including occasional generation curtailment. In fact, an optimal design for distribution could be characterized by active voltage management during normal conditions, and active DER management for ‘N-1’ contingency situations only.
Technical objections to DER in distribution are often true in principle, but do not often materialize when real contexts and realistic parameters are considered. Protection schemes as used in the distribution systems are not adequate in the presence of DER. This is one of the critical questions with the protection of micro grids. However, as long as distribution networks are involved, due to the physical properties of the power system, the operation of protection schemes is dominated by the short-circuit power supplied by the HV network. Furthermore, small DER units installed in low voltage generally do not supply short-circuit power, hence they cannot disturb the operation of protection.
For radial networks (MV and LV part of distribution), the EU-DEEP project proposed new design rules that take care of one of the most critical issues: voltage control in rural networks. Indeed, distribution networks experience voltage drops/rises on the different circuits as a function of their respective loading. Control or compensating equipment is therefore provided to offset the resultant variation in voltage. The basic requirements leading to a ‘flexible’ system assume full ‘non-homogeneity’ for local generation and consumption. The main consequences of such distribution network design rules are:
• HV-MV distribution substations become able to operate at nominal power to or from the distribution network.
• Full ‘non-homogeneity’ between feeders can be accepted without issues in terms of voltage control.
• HV-MV substations do operate at nominal medium voltage under all circumstances, which is slightly lower than in the present situation.
• Distribution transformers are set at their nominal transformer ratio, at least for MV feeders, allowing changes of their operating point from consumption to generation, and vice versa.
• In low load density regions, where distribution networks can be near to voltage drop limits, reinforcement of the system may be needed using, for example, larger cross-section lines or cables.
For massive DER deployment in distribution networks or for the connection of quite large lumped DER in the network, several technical issues must be faced:
• Rural networks might be limited by voltage control issues in the presence of DER.
• Urban networks might be limited by thermal capacity as well as fault level issues, whenever rotating machines are connected.
• In general, large amounts of DER connected to the network will lead to more complex power flows in the distribution network.
4.5.2 New regulatory frameworks
Net tariffs can be used provisionally at small scale by default of adequate metering systems. Some regulatory arrangements implement net metering, hence the ability to reduce the ‘use of system’ charges when installing a DER. This can represent a significant part of the revenues needed to cover the costs of the installation. This becomes an issue for DSO when faced with a significant penetration of such DER.
Present ‘use of system’ tariffs are generally built on two terms: fixed charges related to the subscribed peak demand and variable energy-based charges proportional to the consumption. Owing to the lack of metering infrastructure, the above terms are estimated on a specified time window (typically one year), the sites where generation and consumption remain balanced throughout this period do not pay for the ‘use of system’. This can deeply affect the financial status of DSO. Given the fact that hosting of DER units can sometimes lead to increased costs for the DSO, they should not be ‘punished’ for hosting more DERs.
Distribution network charges schemes must be urgently revisited: since net pricing is not cost reflective, DSO’s business could be at risk under large deployment of DER. Therefore, a sustainable framework has been proposed for allocating distribution network investment costs to customers and distributed generation. These charges must be transparent and non- discriminatory, meaning that ideally each market participant (load/generation) must be charged on the basis of a good estimation of the real costs that they impose on the distribution network. A new efficient ‘use of system’ charges method, based on a ‘marginal’ approach has been developed allowing for unveiling the footprint of load or generation on the distribution network infrastructure. The impact of load and generation must be determined separately as they play symmetrical but complementary roles. Ideally they should be determined for all upstream elements in the network. But such an implementation requires large-scale deployment of smart metering with automatic meter reading and heavy ex post data treatment.
New EU targets are pushing towards more renewable energy and energy efficient distributed CHP units. With demand flexibility, these trends will change the way electricity is generated, transported and used. The integration of DER poses a valid challenge to both industry and regulators. Estimation of their technical and cost impacts remains an issue. So far, most regulators have remained fairly passive and short-sighted, whereas it has been shown that engineering models can adequately estimate both the technical and some of the economic aspects of this integration. A crucial and still open question is, then, to what extent and for what purpose can normative engineering models provide relevant information for economic network regulation?
System operators have the responsibility to develop and to maintain the transmission and distributions networks following exogenously defined targets. New conditions mean new targets in relation to updated energy policy objectives. This is a process that must be catalyzed by regulatory bodies. The most suitable regulatory environment must be implemented with longterm objectives for a smooth, sustainable development of the distribution networks.
Long-term planning of the distribution system is an essential part of the activities of a distribution network operator. Normative models are an attempt to model the planning problem without duplicating the industry planning process. Attractive from a regulatory point of view, they help the regulator overcome the difficulties resulting from the lack of information. These norm models are just special cases of engineering cost functions leading to a high-level representation of the considered network. However, they can include all distribution network cost drivers in order to enable finding the right balance between these cost drivers and the investment required in the network. By doing so, the connection and reinforcement costs, as well as the benefits obtained with DER, could be quantified from a system point of view.
Efficient ‘use of system’ tariffs for distribution should consider:
• Bills of customer based on kW and kWh components of the supply.
• A clear separation of ‘use of system’ tariffs from incentives is mandatory.
• For sites equipped with DER, load and generation must be treated independently.
• This allows for the deployment of ‘use of system’ tariffs that are able to reveal the value of DER (or loads in a part of the network dominated by generation) as ‘ network replacement’.
• But for being applicable down to low voltage networks, the method requires large-scale metering systems as well as ex post data management.
• This asks for simplification. Different stages of simplification are possible. However, caution must be exercised to keep the ‘ essence’ of the tariff during this simplification process. Furthermore, equality of treatment principles must also be integrated to avoid penalizing customers due to their location in the network, like the remote end of feeder; and ‘ use of system’ tariffs must be made stable from one year to the next.
4.6 Acknowledgement
This work has been partly funded by the European Commission as part of EU-DEEP, a European Project supported by the Sixth Framework Programme for Research and Technological Development.
4.7 References
[1] Strbac, G., Jenkins, N. Network Security of the Future UK Electricity System. Report to Performance and Innovation Unit (PIU). 2002.
[2] Lambert, J.D. Creating Competitive Power Markets: The PJM Model. Tulsa, OK: PennWell Corporation; 2001.
[3] Ofgem/DTI. Planning and Operating standards under BETTA 2004; Vols 1 and 2 [An Ofgem/DTI consultation document, July].
[4] EU-DEEP Project. www.eu-deep.com.
[5] Deuse, J., Grenard, S., Benintendi, D., Agrell, P.J., Bogetoft, P. ‘Use of system charges methodology and norm models for distribution system including DER’, CIRED 19th Int. Vienna: Conf. on Electricity Distribution; 2007.
[6] Deuse, J., Purchala, K. DER profitability, distribution network development and regulation’, CIRED 20th Int. Conf. on Electricity Distribution, Prague. 2009.
[7] Available at. Sunpower (2010) High performance free-piston Stirling engines. http://www.sunpower.com/lib/sitefiles/pdf/productlit/Engine%20Brochure.pdf, 2010.
[8] EU-DEEP Deliverable D11. Determination of the value of being network connected application to 5 cases. downloadable from www.eu-deep.com/.
[9] Willis, H.L., Scott, W.G. Distributed Power Generation: planning and evaluation. New York: Marcel Dekker; 2000.