Cost reduction of the wastewater treatment plant operation by MPC based on modified ASM1 with two-step nitrification/denitrification model
ABSTRACT The Activated Sludge Model No. 1 (ASM1) considers that nitrification and denitrification are single step processes and nitrite nitrogen (NO2–N), which is an intermediate for the two processes, is not accounted for. The first part of this paper presents the development of an enhanced ASM1 with two step nitrification/denitrification processes and its implementation in the Benchmark Simulation Model No.1 wastewater treatment plant (WWTP). The secondary settler was considered to be reactive in order to achieve a better fit between the simulation model and the behavior of the real WWTP. The second part presents the investigation of Model Predictive Control approach for the advanced control of the WWTP. Two control strategies are implemented for the wastewater treatment plant and they are analyzed from the perspective of the benefits brought to the WWTP operation. The proposed control strategy shows a reduction of the operational costs and the improvement of the effluent quality index.
SourceAvailable from: Sy najib bin Sy salim[Show abstract] [Hide abstract]
ABSTRACT: Wastewater treatment plant is a large-scale system and highly known with the nonlinearity of the parameters, making them a challenge to be controlled. In this paper, enhanced nonlinear PI (EN-PI) controller is developed for activated sludge process where a sector-bounded nonlinear gain with automatic gain adjustment is cascaded to conventional static-gain PI. The importance in controlling the dissolved oxygen concentration and the improvement of nitrogen removal process are discussed. The effectiveness of the proposed EN-PI controller is validated by comparing the performance of local control loops and the activated sludge process to the benchmark PI under three different weathers. The EN-PI controller is effectively applied in improving the performances of the static-gain PI, hence controlling the dynamic natures of the plant. Itwas proved by significant improvement in effluent violations, effluent quality index and energy saving of the Benchmark Simulation Model No.1.ARABIAN JOURNAL FOR SCIENCE AND ENGINEERING 07/2014; 39(8):6575-6586. DOI:10.1007/s13369-014-1285-2 · 0.37 Impact Factor
[Show abstract] [Hide abstract]
ABSTRACT: The shortage of easily degradable carbon source in the denitrification process limits the biological nitrogen removal efficiency. Partial nitrification has shown to be an attractive technology due its savings in aeration and external carbon source addition cost in the biological nitrogen removal. In this article, a model predictive control method was proposed to optimize the aeration and the external carbon source addition under the disturbance of influent flow rate and nitrogen load. The results indicated that 56% saving in external carbon source addition and 26% saving in aeration cost can be achieved by the model predictive control method applied in the partial nitrification, in comparison to the complete nitrification.Journal of Environmental Chemical Engineering 12/2014; 2(4):1899–1906. DOI:10.1016/j.jece.2014.08.007
[Show abstract] [Hide abstract]
ABSTRACT: This paper makes a critical review of the available techniques for analysing, completing and generating influent data for WWTP modelling. The solutions found in literature are classified according to three different situations from engineering practice: 1) completing an incomplete dataset about the quantity and quality of the influent wastewater; 2) translating the common quality measurements (COD, TSS, TKN, etc.) into the ASM family components (fractionation problem); 3) characterising the uncertainty in the quality and quantity of the influent wastewater. In the first case (Situation 1), generators based on Fourier models are very useful to describe the daily and weekly wastewater patterns. Another specially promising solution is related to the construction of phenomenological models that provide wastewater influent profiles in accordance with data about the catchment properties (number of inhabitant equivalents, sewer network, type of industries, rainfall and temperature profiles, etc.). This option has the advantage that using hypothetical catchment characteristics (other climate, sewer network, etc.) the modeller is able to extrapolate and generate influent data for WWTPs in other scenarios. With a much lower modelling effort, the generators based on the use of databases can provide realistic influent profiles based on the patterns observed. With regard to the influent characterisation (Situation 2), the WWTP modelling protocols summarise well established methodologies to translate the common measurements (COD, TSS, TKN, etc.) into ASM family components. Finally, some statistical models based on autoregressive functions are suitable to represent the uncertainty involved in influent data profiles (Situation 3). However, more fundamental research should be carried out to model the uncertainty involved in the underlying mechanisms related to the wastewater generation (rainfall profiles, household and industries pollutant discharges, assumed daily and weekly patterns, etc.). (C) 2014 Published by Elsevier Ltd.Environmental Modelling and Software 10/2014; 60:188–201. DOI:10.1016/j.envsoft.2014.05.008 · 4.54 Impact Factor