This thesis deals with questions of data quality control based on the principles of mass conservation. The focus is entirely on operational data from wastewater treatment plants. The goal was to provide a practically applicable method for the determination of well-balanced time periods associated with high data quality in historic data. CUSUM charts were found to be an appropriate way to evaluate the error vector of mass balances on a day-to-day basis. This method was called "Continuous mass balancing" and can also be applied for quasi-online monitoring of current operational data. Contrary to static mass balancing as commonly applied in the field of wastewater treatment, continuous mass balancing allows to incorporate the temporal redundancy contained in the data and can therefore detect even minor systematic errors. Flow dynamics (hydraulic retention), leading to delayed output of influent mass flows, have to be considered to achieve good balancing results. Accumulation would normally also need to be considered in short term balances. However, on the time scale relevant for continuous mass balancing its calculation was found to cause too much noise in the balancing error. In addition to continuous mass balancing an algorithm was developed that allows the calculation of all possible balancing equations upon definition of the plant layout and measured and unmeasured variables in all streams. Flow is treated as an individual variable and therefore balancing equations are bilinear. The developed algorithm is based on structural redundancy analysis as known in data reconciliation. There is hope that this thesis may help to close the existing gap between data quality evaluation in wastewater treatment and the powerful methods of data reconciliation developed in the field of process engineering.