1.3.7 Automatic run-up and loading systems
Automatic run-up and loading systems are provided for two main reasons:
- To assist the operator in making the complex sequence of checks necessary prior to and during start-up.
- To run-up and subsequently load the turbine in a safe and consistent manner. Sophisticated schemes can be provided where the rate of run-up or loading is faster and less stressful for the machine than a human operator could reasonably achieve.
The primary output of a typical system is to the basic speed governor, either to increment the speeder gear or to the governor in such a way that failure of the output signal gives a similarly rate-limited change in demand. Some outputs of digital state signals provide interlock or initiation signals to the governor and auto-synchroniser, whilst others provide indications to the operator. The inputs and outputs are shown on Fig 2.10. The inputs to system depend on the degree of sophistication, in the most comprehensive case they would be as shown.
The scope of the system described here assumes that separate operator actions or sequencing systems bring all the auxiliary systems necessary for the safe operation of the turbine to a satisfactory state of readiness. Thus lubricating oil, main and auxiliary CW systems, etc., will have previously been brought into service. A digital state signal will indicate that each of these conditions or prestart interlocks has been satisfied.
Because of the large number of plant-mounted transducers providing inputs, careful consideration is given at the design stage to provide enough redundancy to permit fully automatic run-up and loading with high availability. Most inputs are either duplicated by 'master' and 'check' transducers at the same measuring point, or there is functional duplication built into the system because of measurements made at entirely different measuring points (e.g., separate measurement of inlet steam pressure in left-hand and right-hand steam chests). Thus, if one of the minor input signals is lost, fully automatic operation may still be possible. The failed transducers are identified to the operator.
If several minor input signals are lost or one of a duplicate pair of major signals, automatic control may be restricted such that there is only a fixed safe acceleration rate (if in run-up mode) or fixed safe loading rate (if on load control).
If further signals fail, only manual control may be possible. The turbine governor will provide a limitation on the rate-of-change of speed or load, but this will not necessarily be safe, particularly if a cold start is in progress. The operator must then decide whether there are sufficent desk indications to permit progress, or to hold at a safe condition whilst a repair is effected, or to shut down the turbne.
The main requirement of the automatic run-up and loading system is to limit the thermal stresses within the turbine rotor and valve chests during changes of speed
and load. The method of controlling these is dependent on the type and size of the turbine and the steam conditions for which it is designed. Thus large turbines working at high main and reheat temperatures, when specified for two-shifting duty, require the most sophistication if optimum run-up and loading rates are to be achieved. A large wet-steam turbine for base load operation requires control of stress in the HP chests and rotor, but the LP chests and rotors do not need any separate control at all. Direct measurement of stress, by thermocouples measuring the difference between inner and midwall metal temperatures at suitable measuring points, provides a means ofcontrolling run-up or load changes to the optimum values. Control is of a closed-loop proportional form, acting into the governor as shown in Fig 2.10, so that stress is held constant at the reference value throughout most of the run-up.
One exception to this concerns the critical speed bands of the turbine, which are determined by rotor dynamics. Here the strategy is to ensure that there is a large stress margin before entering the critical speed band and then to provide a rapid acceleration through it. Any 'hold' signals from the operator (or any other source) are vetoed whilst within the band.
Provision is often made to 'hold' run-up and loading by vibration, eccentricity or differential expansion signals sensed by the separate Turbine Supervisory Equipment (TSE). These will also be vetoed within critical speed bands or during block loading; outside these bands they act first to hold and then to advise the operator if any of the TSE signals exceeds a predetermined limit. The 'hold' will be released if the controlling parameter reduces to 80% of the application level. If the TSE signal continues to increase above the 'hold' limit, a second limit is reached at which the operator is advised to trip the machine.
When the turbine reaches synchronous speed, a number of pre-synchronising checks are made (e.g., the Automatic Voltage Regulator must be on 'Auto') and then a signal is given to the autosynchronising system that the turbine is ready and in a safe state to accept load. The governor speed demand is then increased or decreased automatically by pulses from the autosyn-chroniser until synchronism is achieved, the circuit-breaker being automatically closed by the synchroniser following the operator's command to autosynchronise. Alternatively, the synchronisation and circuit-breaker closure is carried out manually.
Following synchronisation, a block load of about 5% is applied automatically. This is to ensure that changing network frequency does not act through the governing speed loop to unload the turbine completely and produce undesirable motoring of the generator.
Thereafter loading takes place as required by the operator or by the unit control system. As described previously, the rates-of-change are controlled by limiting stress levels, default rate limits or by TSE 'hold' signals.
The operator interface is an important factor taken into account in auto run-up and loading schemes. The essentials are that the operator has the ability to initiate the sequence of run-up and the synchronising action. He must also have overriding control to hold or to reduce speed at all points in the sequence if required, apart from critical speed bands and when block loading. Information must be presented to the operator so that he knows the state of the turbine, the limitation being applied at the time and the status of the control inputs. Failed signals must be brought to his attention and abnormal turbine operating conditions alarmed.
In a wider context than auto run-up and loading schemes, it is vitally important to keep the operator aware of any abnormal control action, such as the operation of unloading gear or any governor fault which is causing, or likely to cause, a load reduction if not rectified. Typically an 'unloading gear operated' alarm and a 'governor fault' alarm would be provided. If a governor fault developed into a situation where the governor could no longer control the machine, a further alarm 'governor tripped' would be generated concurrently with the signal to trip the turbine.
The component parts of governing systems have now been discussed fully in functional terms. The subsidiary functions described may be provided in total or in part for any particular application. Irrespective of the method of implementation, it is necessary to understand how the subsidiary functions are connected into the main governor. Figure 2.11 shows the most complex scheme, incorporating all the described functions. The speeder gear provides the basic setpoint to the governor. The acceleration feedback is normally not effective in the steady state but comes into operation transiently when required. Unloading gear can reduce the steam demand when required. Other inputs, such as the overspeed test speed demand or valve testing controls, are switched into service by the operator when required. Auto run-up and loading rates apply the demands of operator or the unit control system for a change in turbine output to the speeder gear in the optimum manner. Note that the governor can be split into two parts — a common processing system and a number of individual valve controllers.