The trend towards the decentralised production of renewable electrical energy entails new challenges for the network operators, since the fluctuating active power infeed may lead to exceedances of the maximum permissible current and voltage values. In particular, the legally stipulated voltage limitations are often the limiting factor for the integration of fluctuating producers into distribution networks. To counteract the voltage rise also without grid reinforcements, local Q(U)-, P(U)- and cos'(P)-controls are often implemented into the inverters which couple decentralised producers, like photovoltaic arrays for instance, with the distribution grid. These controls influence the active or reactive power exchange between the producers and the grid in order to improve the grids voltage profile. Furthermore, the active power injection of a producer could be completely prevented during overvoltage conditions by an overvoltage protection.^ Another conceivable, but so far barely considered voltage control concept is to operate locally controlled reactive power sinks with voltage set-points close to the maximum permissible voltage limitation at the ends of the affected feeders, and to simultaneously use the customers inverters to cover their own reactive power demand. This concept is called L(U)-control with Q-autarkic customers in this thesis. Based on load flow simulations, the impacts of the above-mentioned control concepts on the performance of a theoretical low voltage network are analysed under different load/production scenarios and compared with each other. The performance of the low voltage grid is evaluated by means of the network losses, the transformer loading, the reactive power exchange between the low and the medium voltage grid, the possibly remaining voltage limit violations and the possibly curtailed active power.^ In addition, it is assessed which of these control concepts leads to a discrimination of individual customers and which concepts require a data exchange between the customers and the network operators in order to enable a coordinated reactive power control of the low voltage network. Furthermore, the L(U)-control with and without Q-autarkic customers is simulated under different real network conditions. The results of the active power based control concepts show that the partial curtailment of the producers active power injection by P(U)- controls waste less active power than the disconnection of the producers from the grid by their overvoltage protections, while it still reliably prevents upper voltage limit violations. For each of the evaluation criteria, the results of the reactive power based control concepts show that the cos'(P)-controls averagely lead to the worst grid performance, while the L(U)-controls with Q-autarkic customers averagely lead to the best grid performance.^ Also under realistic network conditions, the L(U)- controls with and without Q-autarkic customers reliably prevent violations of the upper voltage limitation and thus increase the networks hosting capacities for distributed electricity production. Furthermore it turned out that exclusively those voltage control concepts, which require customer-owned inverters for voltage control, lead to a discrimination of individual customers and requires a data exchange between the customers and the DSO in order to enable a coordinated reactive power control of the low voltage grid.