Optimization of wastewater and sludge treatment plants regarding efficiency, flexibility and reduction of CO2 emissions

Abstract

After the 2015 Paris Agreement, most of the countries have taken actions to decrease their greenhouse gas (GHG) emissions. Some of them such as the European Union have even aimed to be climate-neutral by 2050. To achieve these results, the share of renewables in electricity has started to be increased. However, this increase may result in fluctuations in the grid depending on the nature of the renewables. This situation must be handled cleverly. Wastewater treatment plants (WWTPs) have a share of 56% of the GHG emissions in the water industry. Direct carbon dioxide (CO2) emissions of WWTPs are considered to be biogenic and not included in the GHG calculations. However, indirect CO2 emissions, which are produced by usage of electricity, contribute to the GHG emissions. WWTPs need 3-5% of global electricity to operate and this is predicted to increase in the near future. Demand side management is a suitable method to deal with the future electricity grid with a high share of renewables. WWTPs are good candidates to take part within this approach. They have flexible processes to produce and/or consume energy and storage options. If they can be operated flexibly based on the CO2 emission factor of the grid, the share of the renewables in the grid can be increased and the CO2 emissions can be reduced. In this thesis, an optimization model of the WWTP of Krefeld city (in Germany) was built and investigated to decrease its indirect CO2 emissions by flexible operation. This WWTP applies a two-stage biological treatment, named A-B process, to the wastewater. As a result of this process, A- and B-sludge are produced. These sludge streams are dewatered and fed to digesters, which are additionally supplied with co-substrates, to produce biogas. The digested sludge leaving digesters is further treated in centrifuges and dryers to obtain dried sludge. The WWTP is integrated to a waste incineration plant on the same site, which burns the dried sludge and the biogas produced in the WWTP along with the waste, and supplies steam to the WWTP. The units of the WWTP that need power and/or electricity were determined. Only the units with more than 30 kW demand were included in the optimization model. After a detailed analysis of the units, it was found that only 29.5% share of the total energy consumption of the WWTP could be flexibly operated due to operational reasons. A software program called TOP-Energy, which applies mixed integer linear optimization, was used to build and optimize the WWTP. This software has some common units such as water pumps, mixers etc. in its library, but the rest of the units, especially WWTP-specific units such as aeration/activation tanks, sedimentation tanks, filters etc. are missing, so they were created by using the Template Editor of the software within this thesis. GUROBI was used as a solver. The WWTP model was optimized to minimize its indirect CO2 emissions that are generated by usage of electricity from grid. To do this the CO2 emission factor of the grid was used. This factor shows the total amount of the CO2 emissions generated during the production of electricity. It is dependent on the source that is used for the electricity production. As the share of the renewables in the electricity grid increases, the CO2 emission factor decreases. The idea is to shift the operation of the WWTP to the times when the CO2 emission factor is relatively low. Inlet and outlet temperature and pressure values of the units were fixed initially. Solid mass fraction of the sludge and organic mass fraction of the total solid within the sludge were defined for each fuel stream as fixed values ("fuel" represents sludge, co-substrate and biogas.). The minimum and the maximum mass flow rates and power consumption values of the units were given as boundary conditions. All these data were taken from the internal documents of the WWTP. The inlet mass flow rates of the wastewater and the co-substrate into the plant were defined as hourly time series for the whole year 2017. The CO2 emission factor of the German electricity grid and the outside temperature values of Krefeld were also used in the model as hourly time series. The optimization model could not be run for the whole year since the model was very detailed and would take extremely long time to solve.