Autonomous control systems have been requested recently for large-scale real systems. Distributed reinforcement learning is attracting attention specifically in control of physical flow systems such as lifeline systems. In this paper, we will introduce a model of Multi-Stage Flow System (MSFS) as a new problem class. MSFS is a framework which can describe various physical flow systems. Furthermore, it is effective in handling balance between a purpose of system and constraints, constraints under uncertainty and so on that are difficult to solve in conventional methods because of its features. We propose a new bi-directional decision making algorithm with feasible action sets based on a least commitment strategy. We apply our method to controlling of real sewerage systems. The simulation results show that only our method satisfies permissible levels and attains the performance within an acceptance level.
This paper presents a stormwater inflow prediction method based on a rainfall-runoff model. The rainfall runoff model is derived by using system identification method with a nonlinear model called Hammerstein model, which consists of a static nonlinearity followed by a linear time invariant (LTI) system. It is shown that the Hammerstein model can represent nonlinear phenomena of a rainfall-runoff process effectively, thus the introduction of the Hammerstein model improves the accuracy of inflow prediction results. The effective ness of the proposed method is illustrated through numerical simulations by using the rainfall-runoff data which is acquired in actual urban catchment and a sewerage treatment plant.
To reduce the CSO (Combined Sewer Overflow) pollutant discharge, one of the effective options is cleaning of sewer pipes before rainfall events. To maximize the efficiency, identification of pipes to be cleaned is necessary. In this study, we discussed the location of pipe deposit in dry weather in a combined sewer system using a distributed model and investigated the effect of pipe cleaning to reduce the pollutant load from the CSO. First we simulated the dry weather flow in a combined sewer system. The pipe deposit distribution in the network was estimated after 3 days of dry weather period. Several specific pipes with structural defect and upper end pipes tend to have an accumulation of deposit. Wet weather simulations were conducted with and without pipe cleaning in rainfall events with different patterns. The SS loads in CSO with and without the pipe cleaning were compared. The difference in the estimated loads was interpreted as the contribution of wash-off in the cleaned pipe. The effect of pipe cleaning on reduction of the CSO pollutant load was quantitatively evaluated (e.g. the cleaning of one specific pipe could reduce 22% of total CSO load). The CSO simulations containing pipe cleaning options revealed that identification of pipes with accumulated deposit using the distributed model is very useful and informative to evaluate the applicability of pipe cleaning option for CSO pollutant reduction.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.