Creating smarter condition management systems with Trip Multiply functionality

By Dr. Ryszard Nowicki

Every hydro turbine generator has varying vibration frequencies, which can be impacted by specific processes and environmental variables. Vibration measurements – including mechanical, electrical and hydraulic – provide the most important data about the health and condition a turbine generator and can also reveal what other variables may be impacting the operating environment.

For example, an increase in vibrations could indicate an oversized guide bearing clearance or a change in power. Changes in vibration measurements can also result from water flow conditions. The causes for water pulsation are divided into two categories – forced vibrations and self-excited vibrations – which can be observed for:

– Steady conditions when there is no change in flow over a period of time;

– Periodic flow when the variations in flow are repeated at a fixed time interval; and

– Unsteady flow when flow conditions can vary with time – particularly with regard to velocity and pressure.

By adjusting their condition monitoring and protection systems, project operators can decrease instances in which alarms are falsely triggered by equipment operating under normal operating conditions.

Asset condition alarm management

While an overload of false alarms can be managed in ways ranging from bypassing or inhibiting them, this is generally not accepted as best practice. Instead, another approach to prevent alarm overloading is the use of Trip Multiply functionality.

Trip Multiply (TM) functionality is defined in the monitoring system by the maximal multiplier (TMMax) that can be applicable for alarm values, and by a resolution (TMRes) with which it can execute for values that are lower than the maximum values.

Improving the TMRes gives operators a more precise monitoring configuration, and consequently, allows more sensitive system reaction in the case of technical condition changes in a turbine generator.

A model for a Francis-type turbine unit shows the relationship between head and power.

The functionality, indicated as ALERTNORM and DANGERNORM, is designed to temporarily elevate alarm set points by a preconfigured multiple. The TM has been previously limited to rotor resonance effects and radial vibration while allowing the asset to ramp through phases of operation where a resonance exists, defined as the frequency band pass from RPMLOW up to RPMHIGH.

This works well for mid- and high-speed rotating machines that produce vibration at a frequency equal to shaft and rotational speed. The unbalanced force produces serious vibrations when the rotor passes through is balance resonance, but it is not a frequent scenario for turbine generators.

Many low-speed turbines operate in normal technical conditions well below the first rotor resonance frequency (RPMRES). However, it can still be useful for them to use the TM functionality because they present another type of resonance behavior for partial load operation.

Level of vibration change in RPM transient conditions

Vibration is influenced by the speed of rotating equipment during operation, and many machines operate at running speeds above the rotor system’s critical speeds if they are in a good technical condition. Therefore, under transient operating conditions, operators can observe increasing and decreasing vibration levels.

This typically occurs when the speed of the machine is increased from a stopped condition to its operating speed (RPMNOM). This transitional condition is not the same for all generators because many of them operate with nominal rotor speed RPMNOM below rotor RPMRES. However, due to some technical condition changes, such as coupling bolt cracking with the rotor system, a turbine can slowly change its operating mode. Additionally, some generators increase vibration levels during the transient operation condition due to other environmental variables.

The occurrence of resonances and the consequent effects on vibration magnitudes is one of the most common problem sources for turbine bearings and foundation. This is largely due to a faulty model design based on improper estimation of design parameters, such as model stiffness, modal mass of supporting structures and excitation frequency of forces in the machine. For example, a high groundwater table in barrage powerhouses could be responsible for excessive vibration transmission to the entire powerhouse building and structure.

Changes of turbine generator dynamics during loading

Using a Francis-type turbine as a model, Figure 1 shows the relationship between head (vertical axis) and power (horizontal axis), and indicates four zones of possible operation. Only Zone 1 and 3 are for operation. In the optimum operation area, Zone 3, there would be no developed vortex belt in the draft tube. Consequently, the pressure pulsation is small. Zone 2 should be crossed as quickly as possible because it lies between the two operational zones, but it is not approved for operation of its own.

Some vibration components of turbine assembly structures result from the hydrodynamic properties of the flow. Therefore, this destructive zone is determined by intense pressure fluctuations caused by the growth and collapse of bubbles during intensive cavitation conditions. In addition, a correlation can be observed between the excitations resulting from cavitation phenomena and the natural frequencies of the turbine structure and concrete foundation for some turbine types.

For this particular turbine, the smallest level of rotor vibrations are observed for nominal head, and the guide bearing vibrations are positively correlated with head on average. This signifies that the smallest seismic vibrations were observed for the minimal head and the highest ones for the maximum head. Additionally, the lowest vibration levels are around nominal power, and for this unit, they present similar magnitudes.

Data from one-of-six 120 MW turbines in a hydro plant shows the velocity trend of guide bearing vibrations over a week-long period.

In these conditions, the entire structure of a plant would begin to vibrate whenever a turbine operated in a certain load range, even if the turbine were in good technical condition from a mechanical integrity standpoint. This vibration – known as “Rheingan’s Influence” – is caused by spiral vortex filaments that rotate at a speed lower than the rotational speed of the turbine, and can vary with changes of the upstream and downstream water levels.

A vibration test may be carried out during unit commissioning to determine the frequencies of the real dominant excitations, answering questions about the turbine-supporting structure and resonances of the water flow system. It can be treated as a signature analysis allowing better monitoring of dynamic unit behaviors, and easier recognition of some technical condition changes. Such changes are usually visible as a decrease in mechanical resonance frequencies.

Analyzing the data

Figure 2 presents a velocity trend of guide bearing vibrations over a period of one week. The data is from one of six turbines running in a hydro plant – each around 120 MW. The figure also shows two thresholds for “Alert” and “Danger”, and the data reveals the unit runs approximately two times per day.

During steady state operation, vibrations measure around 6 mm/s, which is about 25% below an alert. During both transient operation and steady state operation the unit experiences vibration levels higher than those configured for the danger setup.

A quick analysis leads to the following conclusions:

– Vibration increases are observed systematically during each turbine startup and shutdown; and

– Short-time irregularities during the unit’s stop conditions and in its steady state operation.

The problem of control monitoring system sensitivity and its proper operation can be solved for this plant using the TM functionality. Having the danger – defined here as a HIGH-HIGH level -set up to 12 mm/s, it would be enough to put the TM coefficient as 3.5 and then the danger changes to 42 mm/s.

For only one of six units at this plant, the operator is seeing the HIGH-HIGH alarm 20 times per week. This means that operators will see 120 alarms per week and nearly 7,000 alarms per year with all units having a similar behavior – even though all the alarms generated come from units running in normal technical condition.

Conclusions

Trip multiply functionality is an important feature of condition monitoring and protection systems used for turbine monitoring, and it can significantly decrease operator alarm overloading. It can be important to have condition monitoring systems for units with TMMax levels greater than three.

There are available on the market control systems that offer better TM functionality for hydro applications with a TMMAX of five and TMRES 0.1 and 0.25. Improving of the TMRES allows for more precise monitoring configurations, and consequently, assures more sensitive system reactions in the case of technical condition changes in a hydro turbine.

Using control systems with better TMRes ensures the monitoring system has higher exposure to real changes of an asset’s condition. The elimination of false alarms generated by typical asset condition vibrations and symptoms greatly improves an operator’s actions during real problems with a turbine generator’s mechanical integrity.


Ryszard Nowicki is a mechanical engineer for General Electric’s Measurement and Control division.

More HRW Current Issue Articles
More HRW Archives Issue Articles

Previous articlePumped Storage Hydropower Round-up
Next articleFrom the Board Room: Maryse François-Xausa, Alstom

No posts to display