Power quality : implications at the point of connection

There is a growing interest in de quality of voltage and current. This can be noticed most clearly in reports of the regulators and in discussions within national and international commissions for standardisation about the quality of voltage. Manufacturers insist on a tighter description of the voltage characteristics to optimise the design of their devices, which generally increases the sensitivity for voltage deviations. The grid operators resist against these adjustments because risk exists that the costs for the management and the operation of the grid will increase. Discussions concerning the quality of the voltage are also conducted by the idea that the actual power quality might decrease because of the liberalisation and the need to reduce costs. Furthermore, there is the question to what extent making the energy supply more sustainable by implementing a lot of dispersed generation (DG) can influence the voltage quality. To deal with these problems, more insight in the actual quality of the voltage is needed and possibilities must be there to steer the management and operation of the grid in this respect. Insight in the actual power quality is insufficient. The national measuring campaigns only deliver global information. Intensification of the measurements shall produce a huge amount of data which must be analysed and transferred into useful information for customers and regulators. For incorporating the quality of the voltage in the management of the grid the measurements must be made more specific and the data have to be analysed faster and presented more conveniently. The classification method described in this thesis, is very suitable for this purpose. The advantage of the method is the simple presentation of the results with the "ABCDEF" classification directly showing the quality of the voltage. Therefore in the example described the quality of the voltage is classified from very high quality to very poor quality. To obtain this classification the data can be processed locally with as result the classification and a reduction of the amount of data which has to be communicated to a central location with a factor 300. Therefore, data transmission can be minimised and online PQ monitoring becomes possible The influence of dispersed generation on the voltage is unmistakable present. A simple example is the increase of voltage level at the end of the feeders in the low and medium voltage grid. The combination of load and generation in the medium and low voltage grid will lead to higher voltage variation in the grid. In several studies the possibilities are analysed to deal with this problem. Solutions which are considered are on demand control of the loads and the dispersed generators connected to the grids. In this thesis is concluded that in the Dutch situation several favourable circumstances occur that supports a large scale application of DG without the necessity of making use of complex control strategies. These circumstances are the adaptation of the upper limit of 230 V +10% for voltage level, the voltage regulation in the HV/MV substation and the relatively short length of the cables in the grids. If voltage regulating will be introduced on the MV/LV transformer, implementation of DG with a power equal or less then 70% of the nominal power of the transformer will be acceptable without additional controllers if only the voltage level is considered. No requirements are described in the national Dutch grid code about dips. Since the year 2005, measurement devices are installed on points of connection in the Dutch high voltage grid to count the amount of dips. This information however is not enough to enable the regulator to develop good requirements on this aspect. In this thesis is analysed the average amount of dips, with depth and duration in low and medium voltage grids which can be used as basis to reach edge conditions for dips. Furthermore, also for dips a classification method based on the "ABCDEF" classification is made. Another aspect is the problem with harmonic distortion. On this moment it is only a local problem because in general the level of harmonic distortion in the grid is rather small. However, it has been observed that on locations where a lot of DG is installed, connected to the grid by inverters, the level of harmonic distortion is significant higher. On some places the level of harmonic distortion is even higher than the acceptable level with disconnection of DG as consequence. In this thesis it has been analysed what the causes of this problem are, where the occurrence of oscillation is an important factor. Also the interaction of the inverters current with the voltage is of substantial importance. The "harmonic fingerprint" is developed to make this harmonic behaviour visible. This method has been further applied to other devices and is suitable for type test of devices and test results can give input for improving the standards for harmonic current limits of devices. By using this method also the capacitive behaviour of the devices is measured. This is important to predict the possibility of resonances in the grid. The capacitance of devices should be limited to prevent the occurrence of resonance frequencies in a domain where still a considerable harmonic background voltage exists. From grid operators point of view it is desirable to have edge conditions for the current on the connection point (POC). On this moment only some requirements for -vflicker are incorporated in the national Dutch grid code. For harmonic distortion and flicker there is an urgent need for additional requirements. In this thesis a proposal for such requirements is given based on work which is going on in several standardisation commissions and national research of the Dutch grid operators. These developed and in this thesis described requirements are depending on the current capacity of the connection and are related to the grid impedance. Since the flicker phenomena is of more importance for the grid design than the harmonic phenomena, in first instance the relation between the capacity of the point of connection, the flicker requirements and the grid impedance is determined. These new flicker requirements are simple and univocal to make easy grid design rules possible. Using the same grid impedances as determined for the flicker phenomena, requirements for the harmonic currents on the point of connection are developed. In the final chapter the results of measurements on the location "Bronsbergen" will be presented. This is a Holiday Park where a micro grid situation will be realised in the coming years. Measurements in the existing situation with a considerable amount of PV-systems installed confirm the conclusions drawn before about the voltage level in relation to DG. Also the possible problems with harmonic distortion due to grid connected DG using inverters are underlined. A development of DG systems, which improve or at least does not deteriorate the voltage quality, is desirable.