جستجوی مقالات مرتبط با کلیدواژه « differential evolution (de) algorithm » در نشریات گروه « عمران »

In recent years, with the expansion of urban planning and the larger and more complex water supply networks, the need for energy consumption to carry out the process of water supply and distribution has increased. On the other hand, the increase in energy costs nowadays, due to the limited resources and power plants that produce electricity, has made it necessary to quota available energy for various purposes and activities. Water pumping requires a lot of energy, which is a significant part of the network's operating cost. In this research, intelligent operation of these infrastructures using optimization tools and mathematical models has been considered to be useful by increasing the efficiency of pumping stations in urban areas and reducing costs. Over the past decade, many studies have proposed various automated systems for optimal planning of pump operation with the aim of saving energy and reducing operating and maintenance costs, but a small number of these studies have investigated the optimization of the rotation speed of the pumps, the parameter that has been investigated in this research. The main purpose of this study is to achieve the best energy efficiency in water distribution networks (reducing energy consumption and its costs) while achieving other operating goals (providing standard limits of node pressure, etc.). Optimum adjustment of pumps rotation speed is a topic that has not been investigated in previous researches by considering the historical series of water consumption data.

The hydraulic model of the network, pumping station and its control equipment are simulated by EPANET hydraulic solver. Hydraulic solver coupling with differential evolution (DE) algorithm as a random search algorithm is used to adjust the motor speed of water distribution network pumps optimally. In the DE algorithm, a string of numbers with the range [0.5-2], which is the network pump's rotation speed coefficient, is randomly generated. The objective function evaluates these random values to obtain the optimal answer finally. Constraints monitored in the optimization problem are continuity, energy conservation, and minimum node pressure constraints. In the static operating policy developed in this research, the operational variables of the rotational speed of the pumps are in 15-minute time steps for one day. The DE algorithm optimizes the rotational speed values for one day in 15-minute time steps, and then the optimal pattern is applied uniformly for all days. Hassanabad water distribution network in Tehran has been selected for a case study, and the optimal pumps speed pattern has been extracted as an operation policy.

We received water demands of the entire network for 219 days in 15-minute time steps as the initial research data. The amount of energy consumed by the pumps for these 219 days in the current state of the network is equal to 491496.97 kWh. In the current state, the pumps rotate at a constant speed of 1450 rpm. According to the number of decision variables in this issue (192 variables), the initial population is equal to 400, and the number of iterations of the optimization algorithm is defined as 100. The completion condition of the optimization process is completing 100 iterations. After optimizing and determining the optimal rotation pattern of the pumps, by statically considering this optimal pattern for these 219 days and simulating the network by EPANET software, the energy consumption of the pumps for these 219 days was obtained to be 444212.54 kWh. This optimization-simulation process took about 27 days on a system with "8 core, 2.3 GHz" CPU and "10 GB" RAM. The obtained model shows a significant saving in energy consumption compared to the current state of the network while determining this optimal model by the operator through work experience alone is practically impossible. The obtained energy savings show the efficiency and proper performance of the proposed approach.

In this research, an optimization-simulation model was presented to determine how to adjust the speed of variable speed pumps in water distribution networks with the aim of minimizing energy consumption. The conclusion obtained from the comparison of the results of the optimal model developed for the water distribution network of Hassanabad town using the differential evolution algorithm against the current state of the network indicates that the approach adopted in this study has reduced energy consumption by 47284.42 kilowatt-hours equivalent to 9.6%. Also, in the peak hour of water consumption, the presented approach has been able to reduce the energy consumption by 19914.4 kilowatt-hours, equivalent to 74.3%. This is an important advantage over traditional methods and can be effective in saving energy. The main advantage of this operation approach is simplicity, comprehensibility for the operator and its operationality. In addition to this advantage, perhaps the most important limitation of this method is the lack of consideration of special events and possible accidental conditions in the network and non-compliance with it in the way of operation.