When addressing the steps that need to be undertaken when assessing and selecting a generator for your critical power needs, there are two elements which must be made in the first instance:
It is essential that the generating set rating is optimised, ensuring the set with the running load is closely matched. By doing this, the capital cost can be minimised, not just from the set itself but all the ancillary equipment, such as acoustic container, plant room attenuation, fuel storage and flues, all of which can often equate to 50% of the project cost.
As part of assessing generator selection requirements, the following will be detailed:
An essential element of defining your site parameters is through system voltage. Within the United Kingdom most critical power works undertaken will be at 400 Volts or 11,000-Volts, three phase, 50 cycles.
There are some other applications within the UK market which require different voltages and different frequencies, but they are a few and far between and should be treated as specific entities in themselves. An example would be powering a ship with land-based power source.
Most of the UK will have a nominal ambient temperature of 35°c, although in certain areas, there may be differences. Examples include:
In this case the focus is on BSENISO 8528 (noting the documentation has its own glossary of definitions). Three of the most common that are used to rate the capacity of a generating set in the UK are defined as:
Continuous Power means ‘the average power output of a generating set over a 24-hour period’. COP generating sets can deliver 100% of its power rating all the time, the load doesn’t need to be varied. Further to this, you can run the generating set for unlimited hours in a 24-hour period. However, the peak demand cannot exceed 100% of its rating.
As an example, if your generator is rated at 1000KW, your set will be able to provide you with 1000KW of power on a continuous basis.
The predominant applications in which a continuous power generating set would be used in based load generation, co-generation and utilities, who often use this rating of generating sets on treatment works, for example where they’re doing it for energy recovery.
Prime Power is the most common rating of any generating set in the UK market. There are some very specific criteria that we need to meet in relation to the definition of a prime power rated set. The first of these is that the average power output used over 24-hour period should not exceed 70% of its rating. However, this is the minimum standard and so does vary from manufacturer to manufacturer and engine to engine. Therefore, it’s of paramount importance that this should be validated as part of any tender appraisal process.
The generating set can be run at a 10% overload from one hour every 12. However, the load must vary, but that average should never exceed 70% of its rating. The 70% is the minimum set out in the ISO standard and some engine manufacturers and product manufacturers will vary that level, but that 70% rating is known as the ‘load factor’.
To summarise with prime sets, you can:
Standby power rating (ESP) is a term which often gets confused because generating sets can be referred to as being ‘stand by” generating sets. However, it is important to differentiate between the application and its rating.
A standby power rated generating set carries most of the standard characteristics of a prime rating set, but with one major difference; with this rating you lose the ability of the to carry a 10% overload for one hour in 12. This maximum rating cannot be exceeded.
The load must be a varying load; it cannot be a static and you will see limitations in the warranty available on the engine. If you wish for your engine to run for a number of years, the generating set should not be run for more than typically 200 hours per year.
The running time does vary from engine to engine and from manufacturer to manufacturer, but that is the absolute rating for that generating set.
There are other definitions within the ISO standard, but these are specific applications and carry different criteria with them. Examples include:
The following example demonstrates how different ratings translate into a generating set, using a Kohler-SDMO KD1100. This generating set carries a standby rating of 1,100kVA, which is the maximum output power that the set can deliver. It’s prime rating is 1,000 kVA, so here we see that the prime rating can deliver a 10% overload and that takes us up to the 1100 kVA output level. If you want to run the set on a fixed non varying load i.e., as a base load or for peak lopping then set it cannot run at more than 750 kVA.
These ratings are for fundamentally the same generating set. There may on occasion be minor nuances, but it is fundamentally the same generating set, which carries different levels of rating depending on the application. Using the set outside of its design parameters can cause damage. So, when selecting a generator, understanding the application is critical in ensuring the generating set aligns with your specific loading requirements.
One of the key considerations when selecting an appropriate rating of a generating set is “first step load acceptance”. This is predicated against the brake mean effective pressure that the engine can deliver – so the amount of torque that the engine can put into the shaft to which the alternator is connected. As this is not a prescribed amount it will vary from engine to engine and from manufacturer to manufacturer.
The reason for this is purely a commercial one. To make economic sense of generator ratings, manufacturers need to use the same engine on several different power nodes. For example, the engine that is used on a 1000 kVA generating could also be used on the 1250 kVA generating set, and so on.
The ability of that engine to respond to first step load acceptance is contingent on where it is used in relation it is maximum capacity and to the overall loading on the generating set. For instance, the 1000 kVA set would be able to accept a larger first step load than the 1250kVA would be able to accept. This is because it is fundamentally the same engine, although it has been mapped in a different way.
To better understand this, it is important to look at the way that the generating set reacts to the application of that load. When you apply a load to the alternator, that electrical load effectively acts as a break on the engine, trying to slow the engine down.
This has two major mechanical impacts on the generating set. These are:
There is then a recovery time in which the engine responds. The engine returns to a normal frequency, even over speeding slightly as it comes back to a nominal position.
There are also electrical impacts on the generating set:
Those are perfectly normal responses of the generating set. To that first step load acceptance, section five of 8528 acknowledges that but does prescribe criteria against which the generating set should perform.
We call these the governing classes. Within the ISO standard there are four performance classes, G1 (being the least onerous) through to G4 (being the most onerous).
When creating a specification, the specific level of first step load acceptance that you wish the generating set to achieve (which is typically in the order of around 60%), without a G-classification, means you leave your specification open to misinterpretation or a lack of initial clarity.
If no G-classification is given in the specification, we would automatically assume that that the classification is G2. This is because G2 governing performance most closely represents the transient excursions that you would expect naturally within the grid and other electrical equipment that is connected to it.
For example, if you have for example a substation transformer; when you apply load to the transformer, you will see an initial voltage reduction 5then a time to recover. The classification that the transformer usually works to is broadly in-line with class G2.
This data has been quantified that into a tabular form, so you will see in this example the level of deviation on first step load acceptance in terms of frequencies minus 10%. We are allowed a minus 20% variation in voltage.
Then on load rejection we are allowed a 12% increase in frequency and a 25% increase in voltage. However, the most important element of that is that on the bottom line you can see that in both cases we are required by the standard to get back to a steady state condition within five seconds of that load being applied or rejected.
You can see in the class G3 that those criteria are much closer; we moved from a minus 10% variation to a minus 7% on load application, and the recovery time is three seconds.
Class G4 is the most onerous. This is where the consultant or designer defines the operating criteria that he needs this generating set to meet.
From what we have already defined, there are numerous ways to rate a generating set, which can leave a lot of scope for oversizing the generating set.
If we get the rating definition wrong and if we get our class definition wrong, or we have too high load acceptance with a type class definition, all we do is increase the size of the generating set, which may likely be disproportionate to the actual loading requirement of the building.
To mitigate this risk, what we are endeavouring to do here is narrow down and explain the criteria against which the generator is sized, but we must lose sight of the objective that we are trying to achieve.
This example shows us a practical graphical representation of what that translates into in terms of frequency and voltage error. For example, we see 60% load acceptance data on a 2000 kVA generating set. You can see that this generating set clearly performs within the criteria that are set out for G2 performance. We have got a 12.4% voltage reduction and just on 10% frequency reduction, but we are back well to steady state conditions within the prescribed five second period.
Now we are looking at the application to which the generating set is being placed. Typically, one of the unknown factors with some projects is what that load will look like at the project definition stage. This can be quite difficult to identify in the very early stages of project development. To help remedy this, we will seek to understand the methodologies with which we apply the load to the generating set. This slide gives examples of the input current/starting current of four different types of building services loads.
One of our criteria is the first step load acceptance. In a lot of cases the perceived wisdom is that the moment the generator closes the entirety of the building load will be presented to the generating set. If the running load is above 60% of the generating set capacity, which is an ideal scenario then the generating set won’t be able to take on its first step load acceptance which is limited to around 60%.
However, good load analysis will indicate that depending on the mix of the equipment that has been placed on to the generating set, this may not be necessary anyway. With the utilisation of a good building management system, we can phase in the load on a staged basis rather than coming online the moment the generator contactor closes.
An example of where we often see an element of over specification would be in the data centre sector. Most of the data centres are what we would term “black buildings”. There is very little lighting load, most of the data centre load is either IT load fed via a UPS or it is cooling load.
Within a lot of specifications, we see requirement for high levels of first step load acceptance which exceed 60%. But we also see heavier requirements on the governing standard as well, so it is not unusual for G2 to be replaced by G3. In this scenario we have got a high level of load acceptance and very tight governing standards. Both require an increase in the capacity of the set by at least one frame size of the set, yet the running load remains the same.
If you observe the current curve voltage in the bottom right-hand corner, this is a typical input characteristic to the UPS system. A UPS system does not know or care whether it is running on mains a generator, it behaves in the same way. What this current power waveform shows is that on the application of a voltage from the generator of mains, which is zero, the most important thing for the for the UPS is that it has a steady income in voltage and frequency source.
Again, the UPS does not care whether it is fed by generator or mains but what it does looks for stability of Voltage and frequency and within its design criteria. On this graph, we are seeing is input power returning, the UPS waiting between four and six seconds for that the stability is present. When this happens, the rectifier is switched on, which will then initiate a steady ramp up of power being drawn from the input source after which there is a seamless transfer from battery supply to either mains or generator. This is what we call a ‘ramp’. Generators typically love this style of load application because it is a gentle increase, it does not represent a specific level of load step, which is not a problem in any way shape or form to the generator. It is a perfect way of applying a load and hence does not need to be considered a part of that first step load.
The other key load is the cooling plant. By its nature is very lumpy and it really does depend on the type of input and switching methodology that is used which will determine how much of a load step this represents. However, the cooling can be staged on. In an ideal scenario, with a data centre cooling we typically have between one and two minutes to get the cooling back on. So, you allow the generating set to accept the UPS load. As the engine is now under load then cooling load, it is then no longer “the first step”, so the set is in a much stronger position to accept larger loads.
In this data centre example what we see is the client looking for a 70% of first step load acceptance requirement. None of the loads falls within that “first step” category. But there is a classic case of over specifying the requirement for a generating set. How important it is then that we clearly understand the type of loads we are applying to the set and how they are working within the building “system”.
Additional considerations that we need to factor in when we are looking at generator selection are the types of load that we are going to be putting onto the generating set and the subsequent impact.
One classic category of equipment, which is growing in prevalence in general commercial buildings, are what we call non-linear loads. These are represented by what we call switched electronics, examples of these include:
These all have a large impact on a generating set deals with those types of load. In terms of complying with other ISO standards and harmonized standards, those levels of harmonic current created by the electric switching devices should not exceed five percent THDI levels, which the generating set is able to deal with. However, care does need to be taken particularly in the sizing of neutrals, but the size of the alternator.
The ability of the generating set to handle those types of loads, with higher levels of harmonic current can and do lead to degradation of the output wave on a generating set. This is because it has a limited capability of being able to deal with harmonics.
Therefore we would always recommend that some form of alternator excitation augmentation be added.
General excitation is very limited in its ability to be able to manage higher levels of harmonic current generally. But the addition of a rep excitation should be added, in one of two ways:
Different companies have different methods, but all solutions still offer the same advantages of permanent excitation with fewer disadvantages. This is because these methods are built into the alternator rather than added, and overall reduce the footprint of the alternator. This is linked with the ability of the generating set to dissipate the heat of the losses generated in the higher orders of harmonic content in the load.
In summary, what we have looked at is:
We have focused on WB Power Service’s methodologies when selecting and sizing generators systems and some of the other ancillary parts accompanying them.
This has meant looking at defining the duty and size and how we analyse the load and running requirements of the generating set. Following this, we have investigated how to match those as closely as we can to the power nodes that we have available.
If you have got a complex load application situation, we do have software available that finds a solution to optimizing loads the as they come on. An excellent example of this would be water treatment pumping stations. We can look at optimizing the sequence those pumps come on to make sure that we are defining the size of the generator in the most appropriate way.
Finally, we’ve detailed how to get the right generator and application to come together. For example, our standby rated generating set can cover several different application types such as hospitals, communications and data centres.