Meeting Europe’s new grid code requirements28 March 2017
Work is now underway in European countries to implement grid codes that conform to the EC’s Network Code Requirements for Generators (NC RfG), which apply to all generating units above 800 W, with a deadline for implementation of 19 May 2019. Meeting grid codes is expected to become a major driver in the development of reciprocating engine based generation technology.
Gensets using reciprocating engines have a key role to play helping to accommodate intermittent renewables on the grid and moving our power infrastructure towards lower CO emissions without compromising power supply reliability.
Reciprocating engines provide inertia to the grid, for example, have good short circuit capabilities, are flexible and controllable, offer high efficiency, particularly in CHP applications, and have an important role to play in implementing up and coming technologies such as power-to-heat and power-to-gas.
The top priority in power production is keeping the grid stable. Power production plants feeding electricity into the grid must therefore follow clearly defined guidelines, ie grid codes. For the further development of generation technologies based on reciprocating engines, grid code compliance will be as important a development target as increasing efficiency and meeting emissions requirements.
Historically, big centralised power plants have delivered the system services necessary to ensure the stability of the grid. In the past, when the contribution of renewables and distributed power producers to the total power supply was relatively small, the main rule specified for small generators was that, in the event of a malfunction, they should trip within a very short time. But today, tripping renewables and distributed generators in Europe would produce a complete system blackout, as the UCTE network can only absorb a power drop of 3 GW or less (roughly corresponding to the tripping of two big conventional power plants).
Therefore, it has become necessary for renewable and distributed power producers to deliver system services to the grid in the same way that large conventional power plants are called upon to do. In the case of Germany, for example, this requirement has been incorporated in the German BDEW medium voltage directive (Generating Plants Connected to the Medium-Voltage Network). Other countries have similar medium voltage national directives in place to keep their grids stable, eg, CEI 016 in Italy and G59 in the UK.
The need for action has also been recognised by the European Union. To achieve harmonisation at the European level, the European Commission has established a framework for implementation of national directives, setting out criteria that national grid codes should aim to meet. Called Network Code Requirements for Generators (NC RfG) this new regulation was published in the Official Journal of the European Union on 27 April 2016, with a deadline for final implementation of 17 May 2019.
The new regulation applies not only to renewable generators, but also to diesel and gas gensets using reciprocating engines and combined heat and power plants.
NC RfG only specifies the principles that national regulations should adhere to, requiring the details to be elaborated at national level. This will result in national regulations that differ from country to country. Therefore, manufacturers of gensets using reciprocating engines will have to keep careful track of national regulations as they progress, in order to be able to adapt their products to the changing requirements.
One issue of particular interest will be the approaches adopted to demonstrating compliance with national regulations. In Germany, for example, there is a quite complex — but well established — certification process: first, genset performance characteristics are determined by an independent measurement company, and then, based on these measurements, an independent certification company verifies compliance. If similar procedures end up being established in other European countries, the effort required to provide products compliant with various national regulations will increase.
Frequency and voltage stability
Key factors for normal grid operation are frequency and voltage stability.
As already noted, in the past, gensets were required to disconnect from the grid, in Germany, and in many other countries, where frequency deviated outside very tight limits (49.00 Hz – 51.00 Hz). This poses no problem during normal operation, where primary control measures maintain the UCTE network frequency at 50 Hz ± 200 mHz. The critical issue is maintaining the network during a malfunction. Such a situation occurred in November 2006 when the European power system was effectively divided into three different frequency areas. There was huge production of wind generated electricity in the north of Germany and a power flow of 10 GW from north to south. The result was one area with too much power production and over-frequency, and two areas with less power production and under-frequency, with resulting blackouts.
At today’s levels of renewables and distributed generation, and without the wider frequency ranges we now have for such generators, the situation would have been even worse and not manageable, with all renewable and distributed power producers in one area – a large amount of capacity – disconnecting at the same time due to breaching of the fixed frequency thresholds then prevailing.
To avoid such a sudden drop in generating capacity, grid code rules have been developed for distributed power producers.
The frequency ranges have been widened, typically to a range of 47.5 Hz to 51.5 Hz. Second, special measures are required to deal with over-frequency (LFSM-O) and under-frequency (LFSM-U).
General requirements in NC RfG demand different frequency ranges, between 47.0 Hz and 51.5 Hz, for different areas of Europe (Continental Europe, Nordic, Great Britain, Ireland and Northern Ireland and the Baltics).
For over-frequency scenarios (LFSM-O), where a specific frequency threshold (between 50.2 Hz and 50.5 Hz) has been reached, the power producer has to reduce its actual active power in accordance with a “droop function” to avoid a further increase in frequency over and above the threshold. The droop setting must be adjustable between 2% and 12%.
NC RfG stipulates that LFSM-U mode is only required for bigger power modules (“type C” (above 50 MW and below 110 kV) and “type D” (above 110 kV)). In this case, the power generating unit has to increase power if the frequency is below a threshold, and if it is technically feasible for the power generating unit to do so.
With the rise of renewables and distributed generation, the direction of power flows has changed. Formerly, the power flowed from the transmission network to the distribution network and thence to the consumer. The system’s voltage gradient was specified, and conventional power plants were able to control voltage and reactive power. Nowadays, many conventional power plants are being closed down, and a substantial amount of production takes place at the distribution level and on the low voltage network. In this case, the voltage gradient will no longer be specified, and a distributed power plant can lift the voltage at its point of network coupling. If there are no rules for controlling the voltage or feeding in of reactive power, it could result in unacceptable levels of voltage in some parts of the grid.
In the past, conventional power plants were also responsible for delivering reactive power. Nowadays, renewable and distributed power generation has to deliver the reactive power previously supplied by large power plants, now shut down.
Reactive power is a very powerful tool for adjusting voltage. If a synchronous machine is in under-excited mode, it decreases voltage at the point of network connection. If a synchronous machine is over- excited, it is possible to increase voltage at the point of network connection.
With the development of intelligent rules regarding voltage and reactive power control, it is also possible to avoid further investment in grids. With such rules, it is even possible to connect more distributed power to a connection point in existing networks without reactive power management.
NC RfG does not define the reactive power capability for “type B” (up to 50 MW) synchronous power generation units but says that for such units, the individual system operators have the right to specify the capabilities of synchronous machines.
In the case of Germany, for example, the BDEW directive outlines four different methods for controlling reactive power: a fixed active factor cos f; or an active factor cos f (P); or a fixed reactive power in MVar; or a reactive power/voltage characteristic Q(U).
In the BDEW directive, the demand for reactive power capability is currently specified as being in a range between 0.95 under-excited and 0.95 over-excited. However, system operators want to widen this span to between 0.90 under-excited and 0.90 over-excited. In particular, 0.90 under-excited at undervoltage is very challenging and usually implies oversizing of generators, resulting in higher installation costs and lower efficiency.
In the BDEW directive, the normal voltage operating range for a genset is specified as being between 0.9 Un and 1.10 Un. In NC RfG, there is no specification for the voltage operating range of type B units.
Dynamic grid support and LVRT
The requirement for dynamic grid support is the ability to withstand sudden voltage dips. Gensets must ride through grid failures in a controlled manner and support the network with reactive current to maintain voltage. The dynamic grid support for a voltage dip is called “low-voltage ride-through” (LVRT).
In most cases, the reasons for such voltage dips are fault rectification events on the transmission grid. In the case of a short- circuit on the transmission grid, the affected area of the grid will be isolated within a few hundred milliseconds. As long as the short- circuit is connected to the grid, voltage within the transmission and distribution grid drops. In the past, all renewable and distributed power generators tripped and disconnected from the grid when such an incident occurred. With today’s energy mix, such action is no longer acceptable because of the reduction in the number of conventional power plants. During a malfunction, power plants have to feed in reactive power and deliver short-circuit current. Without this short-circuit current, rapid isolation of the malfunction is not feasible; reactive power is needed to trigger the protection systems.
Another reason why reactive power is necessary during malfunctions is to maintain grid voltage in order to avoid disconnection of further power plants.
In Germany, the goal is to protect the transmission system and, as a result, the European grid. A potential partial breakdown of the distribution system cannot be entirely prevented.
Dynamic grid support is very challenging for gensets with synchronous generators. In the case of a small voltage dip (eg, down to 70% Un), the full mechanical power of the engine could be transferred to the grid. The result of the voltage drop is an equivalent increase in current. However, during a major voltage dip, the electrical power can no longer be transmitted to the grid. If voltage drops to 30% Un, the electrical power will also drop to around 30% of nominal. During the first moment of failure, the engine will deliver the same torque as before the incident. The reaction time of the engine depends on the gas volume between the throttle valve and the cylinder inlet, and on the responsiveness of the engine controller. Depending on engine size, time spans greater than 100 ms are realistic.
If the torque is constant and the active power decreases, engine speed will rise. The moment the voltage starts recovering, the generator is under-excited, and the load angle increases and can reach the transient stability limit for the synchronous machine. In this case, the outcome would be a pole slip, which could result in mechanical damage to the synchronous machine.
The weaker the electrical connection, the faster the speed increases, the deeper the duration and the longer the malfunction. The most critical situation is a voltage dip down to 0% Un. In such a scenario, no coupling to the network is left and the speed increases very quickly.
A common way of describing an LVRT requirement is to present a visualisation of the lower required limit of a voltage-against-time profile of the voltage at the network connection point.
NC RfG leaves it up to each member state to determine the shape of this limit profile in order to complete the requirements.
It is worth noting that the current German regulation for gensets with synchronous machines is less strict than the NC RfG’s basic framework.
For power electronic converter technologies (eg as used in PV plants), it is easier to ride through sudden voltage dips because they are not directly coupled to the grid and there is no risk of a pole slip. For that reason, in many countries the LVRT curves for converters are deeper and wider than those for synchronous machines.
The big advantage of synchronous machines over converters is the deleveraging of a huge amount of reactive current during malfunction. Converters are not able to deliver that much short-circuit current. NC RfG takes this into account and provides different limit curves for each technology, which is also the case with the BDEW directive, but typically not for older existing national directives, which were usually focused on wind turbines.
With the adoption of NC RfG, a distinction in fault ride through requirements between synchronous power generating modules and power park modules will also become standard in national European regulations.
Grid code compliance and how to achieve it
Where a grid code is in force, proof of compliance is required by the grid operator, but the procedures for providing this differ from country to country, and may include: manufacturer declarations, plant owner declarations, manufacturer unit test, plant test, models of static and/or dynamic behaviour, and studies of static and/or dynamic behaviour. It is also possible that compliance (full or partial) must be proven via certificates issued by an authorised certifier.
The proof of compliance to NC RfG standards requires an operational notification by the power generating facility owner to the relevant grid operator. For Type B (up to 50 MW) and C (above 50 MW and below 110 kV) generating modules, what is called a power generating module document (PGMD) must be issued that includes the proof of compliance. The exact format of the PGMD and the information in it is to be specified by the relevant system operator. The system operator can request the following information to be included in the PGMD: evidence of an agreement regarding the protection and control settings; itemised proof of compliance; detailed technical data; equipment certificates issued by an authorised certifier; compliance test reports and studies demonstrating steady-state and dynamic performance; simulation models (type C units only).
In Germany, under the BDEW directive, a complex multi-step approach has been established during recent years, as already noted. In the first step, the manufacturer of power generating modules has to undergo a certification process. For this purpose, power generating unit characteristics are measured by an independent measurement institute. Based on these measurements, and on a simulation model provided by the manufacturer, an independent certification company verifies that the power generating module complies with the requirements of the BDEW directive. If this is the case, a certificate for the unit is issued.
For every power plant, a second certificate — the plant certificate – is required during the planning stage. On the basis of the unit certificates for the chosen power generation modules and taking into account data for other power plant systems, such as transformers, medium voltage cables and additional protection devices, a certification company assesses whether the plant meets the requirements as a whole. After construction and commissioning, a declaration of conformity has to be provided to the relevant system operator, confirming that the power plant will be operated as defined by the plant certificate.
It will be interesting to see how the issue of compliance under NC RfG is handled across Europe, especially in view of the wide range of type B power generating modules there are, in the 1 to 50 MW power range, and also bearing in mind that what is economically feasible for a 50 MW power generating module might not be feasible for a 1 MW module.
With gas CHP units from 0.1 to 2.5 MW and diesel gensets from 0.6 to 3.2 MW certified as per the BDEW Directive, MTU Onsite Energy, which played an active part in the process of developing the German grid code, has demonstrated how compliance can be achieved, and with this knowledge believes it is well placed to help generators in other European countries in to fulfill the requirements of emerging national grid codes based on NC RfG.
In Article 14 Table 3.1, NC RfG sets out the fault-ride-through capability requirements for type B (up to 50 MW) synchronous power-generating modules are defined. The duration for the deepest voltage trough is recommended to be 150 ms and only if system protection and secure operation considerations require a longer duration should the assumed duration be increased to 250 ms.
The 150 ms is a number that MTU Onsite Energy knows very well from the German BDEW medium voltage directive. But to prepare for meeting the requirements of countries that determine that the 250 ms duration assumption is necessary, simulations are being carried out to help optimise the control system philosophy. With faster response control additional oversizing of the generator can be avoided. This is also beneficial because the more oversized the generator the less is the efficiency of the system as a whole.
These fault-ride-through provisions represent a part of NC RfG that is “non-exhaustive”, ie require regulatory decisions at the national level.
Another issue currently under discussion is the setting of banding thresholds for the Type A, B, C and D power-generating modules.
A UK National Grid working group has set out reasoning in some detail as to the merits of the thresholds contained in the final version of NC RfG:
However in Germany the network operator view is that the banding thresholds should be significantly reduced, without detailed explanation of the reasoning and without documentation comparable to that in support of the UK case.
The German proposal is as follows:
There is therefore potential for significant divergence between national requirements, with different design approaches being required in different countries for modules having the same genset size.
Thus genset manufacturers have a good deal of work to do between now and 17 May 2018 to keep on top of developments within each EU country and to ensure their products meet the requirements of their target markets.