Logo: to the web site of Uppsala University

The dynamic behaviour of hydropower plants participating in primary frequency control is investigated in this paper through frequency response, step response and setpoint change tests on three Swedish hydropower plants. Grey-box system identification is used to estimate the parameters of simple linear models suitable for power system analysis and the major shortcomings of the linear models are discussed. It is found that frequency response tests with sinusoidal input signals give more reliable information about the dynamics of the plants than step response tests. It is also shown that backlash in the runner and guide vane regulating mechanisms are of great importance for the dynamic behaviour of the plants, and that the incremental gain from guide vane opening to power varies considerably with the operation point.

The grid frequency quality in the Nordic power system has been deteriorating during the last decade. To improve the situation, a better understanding of the system is needed. In this paper, a model of the Nordic power system dynamics with respect to normal operation frequency control is set up and compared with full-scale measurements on the system. The "60 s oscillation" of the grid frequency is measured and explained by the system model.

This paper presents a mathematical model of hydro power units, especially the governor system model for different operating conditions, based on the basic version of the software TOPSYS. The mathematical model consists of eight turbine equations, one generator equation, and one governor equation, which are solved for ten unknown variables. The generator and governor equations, which are different under various operating conditions, are presented and discussed in detail. All the essential non-linear factors in the governor system (dead-zone, saturation, rate limiting, and backlash) are also considered. Case studies are conducted based on one Swedish hydro power plant (HPP) and three Chinese plants. The simulation and on-site measurements are compared for start-up, no-load operation, normal operation, and load rejection in different control modes (frequency, opening, and power feedback). The main error in each simulation is also discussed in detail. As a result, the model application is proved trustworthy for simulating different physical quantities of the unit (e.g., guide vane opening, active power, rotation speed, and pressures at volute and draft tube). The model has already been applied effectively in consultant analyses and scientific studies.

Nowadays the importance and need of primary frequency control of hydro power units are significantly increasing, because of the greater proportion of intermittent renewable energy sources and more complex structure of power systems. It brings a problem of increasing wear and tear of turbines. This paper studies this problem by applying numerical simulation and concise theoretical derivation, from the point view of regulation and control. Governor models under opening and power feedback mode are built and validated by measurement data. The core index, guide vane movement, is analyzed based on ideal sinusoidal frequency input and real frequency records. The results show the influences on wear and tear of different factors, e.g. governor parameters, power feedback mode and nonlinear governor factors.

A deterioration of the grid frequency quality has been observed in the Nordic synchronous system, and an oscillation with period 40-90 seconds is often clearly visible in the frequency deviation. A working group, consisting of e.g. transmission system operators and power producers, has been initiated to improve and harmonise the technical requirements for primary frequency control in the Nordic countries. In this paper, a method to optimise the tuning of hydropower governors providing primary frequency control to the system is suggested. A model of the Nordic power system is set up, nominal values of the system parameters and their typical ranges of variation are presented, and a controller structure is defined. The objectives of primary frequency control are discussed and interpreted as requirements on the gain of the closed loop transfer functions of the system. These requirements are then used to define tuning goals for an optimisation of the controller parameters. The optimisation is carried out as a weighted minimisation of the closed loop system transfer functions in frequency domain. The result is evaluated in frequency domain and by time domain simulations of a system with added actuator non-linearities. The sensitivity to system parameter variation is analysed in terms of the performance of the optimised controllers in a system with changing parameters, but also in terms of how the optimisation result changes if the nominal system parameters are changed. The results show that compared to the governor tuning currently used in many hydropower plants in the system, retuned governors could reduce the amplitude of the 40-90 second oscillation considerably. A small, floating deadband on the controller output signal is discussed as a means to reduce the number of small movements in the actuators and turbines. The advantage of the presented method is that many different aspects and requirements on primary frequency control are taken into account and the trade-off between different aspects of the performance is visualised and can be controlled directly.

Nowadays, the wear and tear of hydropower turbines is increasing, due to more regulation movements caused by the increasing integration of intermittent renewable energy sources. In this paper, a controller filter is proposed as a solution to the tradeoff between reducing the wear of turbines and maintaining the regulation performance and thereby the frequency quality of the power systems. The widely used dead zone is compared with a floating dead zone and a linear filter, by time-domain simulation and frequency-domain analysis. Simulink models are built and compared with onsite measurement. Then, the time-domain simulation is used to investigate the guide vane movement, the load disturbance and the power system frequency, based on a one-day grid frequency datameasured in this study. In the theoretical analysis, the describing functions method and the Nyquist criterion are adopted to examine the stability of the system with different filters. The results show that the floating dead zone, especially the one after the controller, has a better performance than the dead zone on both the wear reduction and frequency quality. The linear filter has a relatively weak impact on both guide vane movements and the frequency quality. Other related conclusion and understandings are also obtained.

Power systems are making a transition from purely technical, centrally planned systems to market based, decentralized systems. The need for balancing power and frequency control reserves are increasing, partially due to variable renewable production, which gives an opportunity for new incomes but also a challenge in terms of changed modes of operation with risk for reduced lifetime for controllable power plants. This paper investigates how the allocation of a sold volume of frequency control reserves within a large hydropower production fleet can affect the costs of providing primary and secondary reserves, in terms of its impact on wear and fatigue, production losses, and the quality of the delivered frequency control. The results show that for primary control, low static gain in the governors results in poor quality and a large amount of load cycles of the units. High static gain, on the other hand, increases the production losses. The control work of the fleet can be reduced by using a proper balance of primary and secondary control gain on each unit, although the intuitive results from linear models exaggerate this effect. Automatic secondary control improves the system frequency quality but also increases the wear.

This article describes a method to assess the needed production flexibility to adapt the power system to the production from variable renewable energy sources such as wind power and photovoltaics over time horizons of 1-14 days. Load and production data from the German power system is used to quantify the flexibility need in terms of power and energy storage requirement due to higher shares of renewable energy (20-80%). It is found that with an 80% variable renewable energy share in the German system, the average power need from flexible sources decreases by 31 GW (59%) while the peak power need only decreases by 3 GW (4%). In terms of energy, the storage need over a two week horizon increases by 2.6 TWh, which is 14% of the average load per day. If the European plans for 100 GW wind power in the North Sea region are realised, this would mean an increase of the energy storage need in the region with 2.2 TWh over a two week horizon.

Standstill Frequency Response (SSFR) test data from a salient-pole synchronous machine with reconfigurable damper winding is presented. In addition to the regular measurements, the damper bar currents are measured and used to obtain the stator-to-damper transfer functions. The test is performed three times with physically different damper winding configurations. An extension to the standard SSFR test analysis scheme is suggested where the stator-to-damper transfer functions are included. The validity of the identified models is substantiated by comparison of the simulated and measured machine response to a drive torque step disturbance. It is found that the damper winding measurements can be incorporated in the analysis scheme to isolate the effect of the damper circuits. However, for a machine of the type studied, also the standard SSFR test produce yields models that are accurate enough for power system studies.