WATER UTILITY ENERGY CHALLENGE

Transcription

1 WATER UTILITY ENERGY CHALLENGE PEPSO II User Manual WATER UTILITY ENERGY CHALLENGE GREAT LAKES BASIN

2 COPYRIGHT BY 2016 ALL RIGHT RESERVED S. Mohsen Sadatiyan A., Carol J. Miller WAYNE STATE UNIVERSITY, DETROIT, MI

3 TABLE OF CONTENTS Table Of Contents... iii List Of Tables...iv List Of Figures... v List of Acronyms and Abbreviations... 7 Introduction... 8 Chapter PEPSO II Interface Main Form Project Configuration Form Electricity Form Pollution Emission Form Constraints Form Optimization Options Form Reporting Option Form Chapter Step by step optimization Guide How to install Introduction of the sample problem Optimization Procedure Appendix A - Glossary Appendix B - Sample Optimization Run Appendix C - Sample Optimization Outputs References iii

4 LIST OF TABLES Table 2- Emission factor values that are used for optimization of Monroe WDS Table 3- Constraints on water level in tanks of the Monroe WDS Table 4- Constraints on water pressure at strategic junctions of the Monroe WDS Table 5- Optimization options that are used for optimization of Monroe WDS Table 6- A sample result of Monroe WDS optimization with PEPSO II iv

5 LIST OF FIGURES Figure 1- PEPSO II Flowchart... 9 Figure 2- Main form of PEPSO II Figure 3- Project Configuration form of PEPSO II Figure 4- Electricity tariff (Top) and electricity meter (Bottom) tabs of the electricity form of PEPSO II Figure 5- Pollution emission form of PEPSO II Figure 6- Pump constraint (top), tank level constraint (bottom) tabs of constraint form of PEPSO II Figure 7- Junction pressure constraint (top) and water velocity constraint (bottom) tabs of constraints form of PEPSO II Figure 8- Optimization option (top) and initial pump schedule (bottom) tabs of the optimization options form of PEPSO II Figure 9- Report option form of PEPSO II Figure 19- First group of four screens that you see during installation of PEPSO II (starts from top left and ends at bottom right) Figure 20- Second group of four screens that you see during installation of PEPSO II (starts from top left and ends at bottom right) Figure 21- Model schematic WDS of Monroe Figure 22- Main form of PEPSO II after loading the Monroe WDS project Figure 23- Project configuration from after loading the sample Monroe WDS project Figure 24- Tariff tab of electricity from of Monroe WDS project Figure 25- Meter tab of electricity form of Monroe WDS project after selecting the main pump station Figure 26- Pollution emission form of Monroe WDS project Figure 27- Pump tab of constraints form of Monroe WDS project v

6 Figure 28- Tank tab of constraints form of Monroe WDS project Figure 29- Junction tab of constraints form of Monroe WDS project Figure 30- Pipe tab of constraints form of Monroe WDS project Figure 31- Optimization option form of Monroe WDS project Figure 32- Report options form of Monroe WDS vi

7 LIST OF ACRONYMS AND ABBREVIATIONS GA: Genetic Algorithm GHG: Green House Gas GUI: Graphical User Interface LEEM: Locational Emission Estimation Methodology NSGA II: Non-dominated Sorting Genetic Algorithm II PEPSO: Pollution Emission Pump Station Optimization RRS: Relative Rotational Speed VB.NET: Visual Basic.NET WDS: Water Distribution System 7

8 INTRODUCTION Pollutant Emission Pump Station Optimization 2 (PEPSO II) is the second generation of a computer program that has been developed for optimizing pump operation of WDSs (Sadatiyan 2016). PEPSO II uses code libraries of EPANET Toolkit V (Rossman 1999, Rossman 2008) for hydraulic simulation. In the first chapter of this document, different component of the graphical interface of PEPSO II is explained. In the second chapter, an optimization run example is introduced and explained step by step. If you are interested in the technical detail of the process, you can read the technical manual of PEPSO II, and if you want to be familiar with the practical usage of PEPSO II as soon as possible, you can skip the first chapters of this document and go directly to the last chapter. 8

9 CHAPTER 1 1. PEPSO II Interface PEPSO II was aimed to work as a user-friendly software that WDS designers and operators with an average knowledge of hydraulic and WDS operation can use. Therefore, the inclusion of a strong graphical user interface (GUI) was a key element in PEPSO II s development. PEPSO II has seven major forms that allow the user to define an optimization project and execute it. Figure 1 illustrates the process flow enabled by PEPSO II s interface. Each of the steps and related forms is explained in the following section. Figure 1- PEPSO II Flowchart 9

10 1.1. Main Form The Main form is the first form that appears on the user s screen upon execution of PEPSO II. It provides access to all forms via menus and tool strip. It also shows a summary of all defined or loaded project information. During the optimization run, the main form provides run-time information and statistics of the optimization process. Figure 2 provides a screenshot of the main form that is displayed at one point for a test simulation. Figure 2- Main form of PEPSO II 1.2. Project Configuration Form The project configuration form is the initial point for defining a new project. It also can be used for changing some basic information of a loaded project, such as name, project folder address, and hydraulic model file address. Through buttons of this form, users have access to all other forms for adjusting project parameters before running the optimization process (see Figure 3). 10

11 Figure 3- Project Configuration form of PEPSO II After initializing a project using the configuration form, a suite of additional forms can be accessed to define the project further. These additional forms include the electricity, pollution emission, constraints, optimization, and report options form. All of these forms have been designed with the same logic to create a consistent user experience. This ensures that multiple scenarios of electricity tariff, pollution emission, pump, tank, junction and pipe constraints and optimization options can be defined, saved, loaded and selected as an active scenario by using the same logic Electricity Form Most of the industrial electricity tariffs have two parts: a) energy consumption charge and b) power demand charge. The energy consumption charge ($/kwh) should be multiplied by the amount of consumed energy (kwh) to calculate energy consumption cost ($). Similarly, power demand charge ($/kw) should be multiplied by peak power demand to calculate the power demand cost ($). The peak power demand of an electricity 11

12 meter can be calculated as a maximum power demand of the electricity meter during a defined billing period (e.g. one month) that is measured in a defined time intervals (e.g. 30 minutes intervals). Total electricity cost is electricity consumption cost and power demand cost of all electricity meters. The electricity form has two tabs.users can input various types of electricity tariffs in the first tab. In this tab, electricity tariffs that have a constant rate energy consumption charge, as well as time-variant rates ($/kwh) can be defined. Also, power demand charge ($/kw) and duration and intervals of calculating peak power demand can be defined via this tab. Note that it is possible to define and use multiple electricity tariffs in an optimization scenario for different electricity meters. However, each electricity meter can have only one electricity tariff. After defining, at least, one tariff, the second tab can be accessed to define electricity meters and assign the defined tariff to them. Most of the time, a pump station has one electricity meter. However, it is possible to define multiple electricity meters for pumps that are physically located in one pump station. Each electricity meter should have a list of the connected pumps. Peak power demand and energy consumption of pumps that are connected to an electricity meter will be added up before calculating the electricity cost. Note that each pump can be connected to only one meter. Figure 4 shows tariff (top) and meter (bottom) tabs of the electricity form. Latitude and longitude of electricity meter are necessary input parameters if the user plans to use the pollution emission calculation or optimization routines. Location of the electricity meter will be used to retrieve the emission factor values from the LEEM subroutines. LEEM which stands for Locational Emission Estimation Methodology is a technology that is developed for using publicly available data of electricity grid to predict pollution emission that is associated with energy consumption at each location and time. 12

13 Figure 4- Electricity tariff (Top) and electricity meter (Bottom) tabs of the electricity form of PEPSO II 13

14 1.4. Pollution Emission Form One of the unique characteristics of PEPSO II in comparison with other pump operation optimization tools is its ability to use the emission factor report of LEEM to enable real-time spatially-explicit emission reduction optimization. The pollution emission form is the interface for user-defined pollution emission calculation scenarios. Each scenario may include one pollutant or a user-defined pollution index that is a linear combination of multiple pollutants. Users can elect to receive emission factor values from the LEEM server via internet or use an offline LEEM report. The offline LEEM report option is useful when the user wants to compare results of different optimization runs and wants to prevent unwanted changes due to receiving different reports from LEEM during different optimization runs (due to the time-sensitive nature of the emission factors). Figure 5 shows a screenshot of the pollution emission form. Figure 5- Pollution emission form of PEPSO II 14

15 It should be noted that only those pollutants that their information can be obtained from the LEEM server or an offline report can be selected in this form. Currently, LEEM 2.5 server provides emission factor (lb/kwh) of five pollutants (CO2, NOx, SO2, Hg, and Pb). PEPSO II uses the user-specified location of each electricity meter (from the electricity meter tab of electricity form) and time of optimization, to query emission factors from LEEM server. LEEM 2.5 has information of power generators in the Great Lakes region. Therefore currently, PEPSO II is able to obtain emission factor of pump stations that are within the Great Lakes region from the LEEM server Constraints Form The constraints form has four tabs that allow users to define customized constraint scenarios for pumps, tanks, junctions, and pipes. It also is possible to select the default constraint scenario that PEPSO II automatically defines based on characteristics of the WDS model. Although it is not recommended, it is possible to turn off constraint scenarios, allowing network optimization in the absence of any constraint on the operation of pumps, the water level in tanks, pressures of junctions or water velocity in pipes. As shown in Figure 6 (top), the first pump tab of the constraints form allows the user to define whether a pump is a variable speed or fixed speed pump. For a variable speed pump, the user can input the pump s minimum possible relative rotational speed (RRS). RRS is a number between 0 and one that 0 means the pump is off, and 1 means the pump is working with its maximum rotational speed. Based on the pump affinity law, the power demand of a pump is directly proportional to the cube of the RRS (Pelikan 2009). When RRS of a pump is 0.5, it only can push water with (0.5 3 ) = 12.5% of its nominal power, so it is not practical to use a number less than 0.5 as the minimum RRS of the pump. Note that the maximum RRS of all pumps is considered 1 (100% of the maximum rotational speed of the pump). Other constraints for pump operation include a) a maximum number of switches in a day, b) a minimum duration of time between pump shut-down and start-up, and c) maximum continuous period of operation for the pump. When a pump operation schedule violates these limits, a penalty will be calculated and added to the total penalty of the pump schedule. Users can define these limits as hard constraints or not. This means that in addition to calculating penalties, these limits can be used for defining an acceptable or 15

16 unacceptable pump schedule. If the user decided to set these limits as hard constraints, violation from them completely discredits the pump schedule from being selected as the optimum final result. PEPSO II will not use these hard constraints during the optimization process. However, at the end when a solution should be selected among the final qualified solutions (Pareto frontier) as the optimum solution, these hard limits will help to filter out unacceptable solutions. Using a hard constraint during the optimization may restrict the ability of PEPSO II to explore the solution space for the optimum solution. In the constraints form for the tank (Figure 6 bottom), users can define allowed and desired minimum and maximum level of water in the tanks. Note that the minimum and maximum tank levels that are set out in the EPANET model file are physical limits and EPANET does not let water level to go beyond these limits. However, the desired minimum and maximum levels that are defined by users are soft constraints. Water level can go beyond desired limits, but this violation causes some penalties. By default, the minimum and maximum desired water level constraints of a tank are 15% higher and lower than the bottom and top level of the tank, respectively. The minimum and maximum allowed water level that can be defined by users are hard constraints and like hard constraints of the pump operation tab, will not be used during the optimization process. However, at the end of the process, they will help to filter out all unacceptable solutions from the final list of solutions (Pareto frontier). By default, the minimum and maximum allowed tank levels are equal to the minimum and maximum tank levels of the EPANET hydraulic model, respectively. Five last columns of the tank constraint table can be used for constraining water level in the tank at a given moment. For instance, if the operational requirement of a WDS dictates that a tank should be 50% full at 7:00 AM, this part of the table can be used to define constraint water level and time. Like desired minimum and maximum level, this is a soft constraint and will be utilized only for calculating penalties. However, if users select the strict water level control at specific time option, it will be used as a hard constraint for filtering out the unacceptable solution at the end. 16

17 Figure 6- Pump constraint (top), tank level constraint (bottom) tabs of constraint form of PEPSO II 17

18 Figure 7- Junction pressure constraint (top) and water velocity constraint (bottom) tabs of constraints form of PEPSO II 18

19 The junction and pipe tabs of the constraints form that are shown in Figure 7 (top and bottom respectively) allow the user to select strategic junctions and pipes from the list of all junctions and pipes of the WDS and assign the minimum and maximum allowed and desired pressure or velocity limits to each of them. It also is possible to indicate the relative importance of each junction or pipe with respect to others by defining the constraint importance multiplier. By default, these multipliers are equal to one for all junction and pipes, resulting in equivalent penalty associated with the violation of pressure or velocity limits of all selected junction and pipes. However, changing the constraint importance multiplier of a junction increases the penalty associated with pressure violation of that junction with respect to others. Like the tank level constraints, the desired limits define soft constraints. The pressure or velocity violation from these limits increase the calculated penalty. It is important to know that violation from each of the minimum and maximum limits of water level in the tank, water pressure at a junction or water velocity has different meaning and PEPSO II stores these violations separately. PEPSO II will use them separately to discover promising ways of changing the pump schedule for improved results. The pressure and velocity allowable limits are stricter hard constraints and will just be used for filtering out acceptable solutions from the final Pareto frontier at the end of the optimization process Optimization Options Form Users can open the first tab of optimization options form to define optimization algorithm parameters and objective functions (see Figure 8, top). In the upper part of this tab, three objectives of optimization (electricity cost, pollution emission, and penalties) can be selected. Electricity cost is composed of energy consumption cost and power demand cost ($). Pollution emission is the weight of a single emitted pollutant (lb) or values of the user-defined pollution emission index. Lastly, the penalty value is total penalty formed from pump operation constraint violations, water level violation of tanks, pressure violation, and velocity violation. Here users also can define relative weights of each selected objective. This weight will not be used during multi-objective optimization process of PEPSO II that optimizes each objective independently. However, at the end of the process and before reporting the final optimum solution, it will be used to select the 19

20 optimum pump schedule among all the acceptable solutions among the final group of solutions (Pareto frontier). The middle section of the options tab defines stopping criteria. Optimization can be stopped under any of 5 user-defined conditions: (1) elapsed computation time, (2) the maximum number of iterations, (3) the maximum number of solution evaluations, (4) when a set of predefined objectives is reached, or (5) a maximum number of stagnant iterations. The bottom section of this tab gives users some options to adjust the optimization algorithm options. For instance, for NSGA II optimization method, users can define the number of solutions in each population, crossover and mutation percentage and rate, and the number of elite solutions of each population. The crossover and mutation percentages define the portion of the population that should be used in crossover (reproduction) process or should be mutated, respectively. The crossover and mutation rates determine the portion of a selected pump schedule which should be changed during crossover and mutation processes, respectively. By default, both the crossover percentage and rate are 50%. The mutation percentage and rate by default are 5% and 10%, respectively. High mutation and crossover rates may change the selected pump schedule drastically that may aid PEPSO II s exploration of the solution space, but may also decrease the efficiency of exploitation process and fine tuning the near-optimum solution. The second tab of the optimization options form that is shown in Figure 8 (bottom) helps users to customize start point of optimization and define an initial population of solutions. By default, PEPSO II forms the initial population by a group of randomly created pump schedules. It also adds two extreme pump schedules to the population to catch two extreme points of solution space. In one of those two extreme pump schedules, all pumps are off and in the other one, all pump are on. However, in addition to the default initial population, users can define some initial pump schedules and use them to replace all or part of the initial random population. This option is especially useful when it is desirable to initiate an optimization run from the result of a previous optimization run. It also can be used for comparing different optimization scenarios when users want to keep the initial population of all scenarios the same. 20

21 Figure 8- Optimization option (top) and initial pump schedule (bottom) tabs of the optimization options form of PEPSO II 21

22 1.7. Reporting Option Form The reporting options form provides all options that users need to customize reports of PEPSO II. The top section of this form allows users to select different types of reports that should be included in the text output. A field in the middle section of the form is provided to define the file name for the optimized EPANET model. The bottom section of the form shows all options for customizing the graphical report. Users can select different types of graphical reports, their updating frequency, and detailed adjustments about label or scale of axes of the graphs. Figure 9 shows a screenshot of this form. The text report, Richmond Test_Optimized.inp, with optimized model and graphs, will be saved in the project folder. Users can select to show (during optimization) and save a) the best practical pump schedule and b) the optimization objectives trends. Figure 9- Report option form of PEPSO II 22

23 CHAPTER 2 2. Step by step optimization Guide 2.1. How to install PEPSO II package can be installed by running the PEPSO II Setup.exe file. These are minimum system requirements of PEPSO II: Windows Vista SP2 or newer versions CPU: 1 GHz RAM: 512 MB Available HDD space: 15 MB (plus 110 MB for results of a sample optimization run) Display resolution: 800x600.Net framework 4.5 If all the prerequisites are installed in advance, PEPSO II will be installed in eight straightforward and self-explanatory steps that are shown in Figure 10 and Figure 11. PEPSO II installer creates start menu and desktop shortcuts. It also creates the Sample Project folder in the installation directory which contains the required files for running the sample optimization that is explained here. 23

24 Figure 10- First group of four screens that you see during installation of PEPSO II (starts from top left and ends at bottom right) This user manual will be copied to the installation directory of PEPSO II during the installation process. Therefore, you can have access to this from the installation directory but at the end of installation process, you can open it by clicking on Read me button of the fourth screen that is displayed in Figure

25 Figure 11- Second group of four screens that you see during installation of PEPSO II (starts from top left and ends at bottom right) 2.2. Introduction to the sample problem In this chapter, we are going to show you how to find an optimum pump operation schedule for modified water distribution system of the city of Monroe, MI. We are going to use PEPSO II to reduce electricity cost and associated CO2 emission of the both pump stations of this WDS while the water level in tanks and pressure at strategic junctions are within the desired ranges. In this example, WDS of the city of Monroe serves about customers, and its water demand is (m 3 /day). The minimum and maximum hourly demand multipliers 25

26 of the system are 0.67 and 1.19 respectively. EPANET hydraulic model of WDS of Monroe that is used in this study has 1531 junctions, 1945 pipes, 11 constant speed pumps, two variable speed pumps, one reservoir, three tanks, and one 24-hour water demand pattern with one-hour time step. The hydraulic simulation period of the model is also 24 hours with a one-hour time step. Figure 12 displays a model schematic of WDS of Monroe. Figure 12- Model schematic WDS of Monroe Finally, it should be noted that in this test, emission factors just change in time, and we did not include any special variation for emission factors. However, as it was mentioned previously, emission factors that are reported by LEEM may vary due to change in location of the energy consumption points. The area that is covered by Monroe 26

27 WDS was not wide enough to change emission factor values based on the location of the pump stations. However, one can use PEPSO II to optimize a WDS that its pump stations are far from each other. In this case, PEPSO II can take advantage of the change in emission factors at different locations and find a better solution by shifting the location of energy consumption from one pump station to another one Optimization Procedure After running PEPSO II using its shortcut on the desktop or start menu or directly by double clicking on PEPSO II.exe you can see the main window (see Figure 13). Here you should click on Open icon ( ) or click on Open item of File menu to open the file browse dialogue. Now you need to locate the project file of Monroe WDS example and open it. The Monroe WDS.prj file can be found in the Sample Project folder in the installation directory of PEPSO II. If you did not change the default installation directory during installation of PEPSO II, the sample project file could be found at this location: C:\Program Files (x86)\wsu\pepso II\Sample Project. If you run PEPSO II as an administrator, you can keep the Sample Project folder in its default place and open the project file. However, if you cannot run PEPSO II.exe as an administrator you need to copy the Sample Project folder to a drive that is not protected, and you have complete access to it. Then you need to open the Monroe WDS.prj file with a text editor (e.g. notepad) and change the address in front of Project Folder: phrase to the address of newly copied Sample Project folder and save the project file. In this case, PEPSO II can read the project file without having the administrator permission and write the result of optimization in the Sample Project folder. Appendix shows the content of the project file that, in addition to PEPSO II, can be accessed and edited with any text editor. Figure 13 displays the main form of PEPSO II after loading the Monroe WDS project. 27

28 Figure 13- Main form of PEPSO II after loading the Monroe WDS project Now all required information for running the optimization process is loaded but before starting the optimization process, it is good to take a look at the predefined aspects of this project and be more familiar with different forms and menus of PEPSO II. By clicking on Configuration icon ( ) or selecting Configuration item of Tools menu, you can open the project configuration form that is displayed in Figure 14. When you want to define a new project, you should click on New icon ( ) or select New item of File menu. This action also opens a project configuration form. Here you should pay attention to the address of project folder and EPANET input file. If you installed PEPSO II in the default directory ( C:\Program Files (x86)\wsu\pepso II\Sample 28

29 Project ), these values should be good. However, if you changed the installation directory during the PEPSO II installation, you need to change these two address accordingly. After doing required changes, save the form by clicking on the Save button at the bottom of the form. Other forms of PEPSO II can be accessed via the button of the project configuration form, or you can close this form and go to the main form to have access to the rest of forms. Figure 14- Project configuration from after loading the sample Monroe WDS project By clicking on Electricity icon ( ) (or selecting this item: Tools>Electricity>Tariff) you can open the Tariff tab of the Electricity form. As it is shown in Figure 15, you can edit the defined electricity tariff or add a new electricity tariff. The electricity tariff that is already defined and named as Tariff 1 is based on the D6-220 electricity tariff of DTE Energy which is providing power for Monroe WDS. This electricity tariff includes the energy consumption charge and power demand charge. The energy consumption charge for on-peak hours (11:00 to 18:59) is ($/kwh) and for off-peak hours (19:00 to 10:59) is ($/kwh). The power demand charge is ($/kw) that should be multiplied by the 30 minutes peak power demand during 30 29

30 days period to calculate the power demand cost. So considering the similar peak power demand for all days of a month, daily power demand charge is 14.34/30=0.48 ($/kw). Figure 15- Tariff tab of electricity from of Monroe WDS project Each one of the two pump stations of Monroe WDS has an electricity meter. Their information inputted in Meter tab (see Figure 16). By selecting a defined electricity meter from the drop-down menu, you can see all pumps that are connected to the meter. Figure 16 displays location and list of pumps of the electricity meter of the main pump station. 30

31 Figure 16- Meter tab of electricity form of Monroe WDS project after selecting the main pump station Pollution emission form can be accessed by clicking on Pollution Emission icon ( ) or selecting Pollution Emission item of Tools menu. Before opening the pollution emission form, PEPSO II connect to the LEEM server to download an updated list of available pollutants. This might takes a couple of seconds and then you can see the form as it shown in Figure 17. For this sample optimization run, PEPSO II uses an offline report of LEEM server. So you do not need to connect to the LEEM server. LEEM report location field shows the address of the offline LEEM report file. If you installed PEPSO II in the default directory ( C:\Program Files (x86)\wsu\pepso II\Sample Project ) this value is good. However, if you changed the installation directory during the 31

32 PEPSO II installation, you need to modify this address accordingly. Based on the offline LEEM report values, the CO2 emission factors are presented in Table 1. Table 1- Emission factor values that are used for optimization of Monroe WDS Time CO 2 Emission Factor (kg/mwh) Time CO 2 Emission Factor (kg/mwh) 00: : : : : : : : : : : : : : : : : : : : : : : : Figure 17- Pollution emission form of Monroe WDS project The next step is controlling the pump, tank, junction and pipe constraints. You need to click on Constraints icon ( ) (or select Tools>Constraints>Pump) to open Pump 32

33 tab of Constraints form. If you look at Figure 18, you can see that both booster pumps PMP-544 and PMP-9 are defined as variable speed pumps. The minimum relative rotational speed of both variable speed pumps is 60%. For all pumps, the maximum allowed number of pump switches in a day is 24 and the minimum duration of time between pump shut-down and start-up is 15 minutes. The maximum allowed a continuous period of operation for all pump is 24 hours. As you can see, all check boxes of the last three columns of pump constraint table are checked. It means that if a pump of a nondominated solution from the final Pareto frontier has any of the three conditions that are defined by the header of these three columns, that solution will not be selected as the final best solution. Figure 18- Pump tab of constraints form of Monroe WDS project Clicking on Tank or Junction tabs of Constraints form displays the defined constraints on the water level in a tank or pressure of strategic junction (see Figure 19 and Figure 20). The constraints on the water level in tanks and water pressure at strategic junctions are also presented in tabular format in Table 2 and Table 3. 33

34 As you can see in Figure 19, the default water level constraint in tanks are used in this sample optimization run. Minimum and maximum allowed water levels act as hard constraints and are defined based on the level of bottom and top of the tanks. However, desired water levels are acting as soft constraints. Minimum and maximum desired water level defined based on the tank that is 15% and 85% full respectively. Although it has not been used in this sample optimization run, you can see that it is possible to define a minimum water level at a given moment and ask PEPSO II to find a solution that also satisfies this constraint. Allowed and desired minimum and maximum water pressures that are displayed in Figure 20 also act as hard and soft constraints respectively. These four strategic nodes that are added to the constraint table and their pressure limits defined based on regular pressure control nodes of Monroe WDS that its operators use. The constraint importance multipliers of all four strategic junctions are one; that indicates that the water pressure violations at all strategic junctions have the same emphasis. Based on your need, you can add a junction to or remove them from the strategic junction table by using the arrow buttons in the middle of the form. Tank ID Elevation (m) Table 2- Constraints on water level in tanks of the Monroe WDS Water Capacity (m 3 ) Minimum Allowed Water Level (m) Minimum Desired Water Level (m) Maximum Desired Water Level (m) Maximum Allowed Water Level (m) T T T Strategic Junctions ID Table 3- Constraints on water pressure at strategic junctions of the Monroe WDS Minimum Allowed Pressure (psi) Minimum Desired Pressure (psi) Maximum Desired Pressure (psi) Maximum Allowed Pressure (psi) Constraint Importance Multiplier J J J J

35 Figure 19- Tank tab of constraints form of Monroe WDS project Figure 20- Junction tab of constraints form of Monroe WDS project 35

36 The pipe tab that is showed in Figure 21 can be used for defining a constraint on water velocity in strategic pipes of WDS. In this sample optimization run, we did not define any limitation for water velocity in pipes so from the drop-down menu None is selected as the pipe constraint. Figure 21- Pipe tab of constraints form of Monroe WDS project After defining electricity, pollution emission and constraints information we just need to adjust optimization parameter and required reporting options and run the optimization process. Different parameters that are used as optimization options of the WDS of Monroe are listed in Table 4. Clicking on Optimization Options icon ( ) or selecting Optimization Option of Tools menu open the optimization option form which is displayed in Figure 22. The information of Table 4 is used for filling the fields of Figure 36

37 22. For more information about these parameters, please refer to chapter one of this document and PEPSO II technical manual. Table 4- Optimization options that are used for optimization of Monroe WDS Parameter Value Optimization Duration (hr) 24 Optimization Time Step (min) 60 Maximum Number of Iterations 300 Maximum Number of Solution Evaluations Maximum Optimization Time (min) 500 Minimum Optimization Rate 1% During 100 Iterations Electricity Cost Goal ($) 0 Pollution Emission Goal (kg) 0 Total Penalty Goal 0 Population Size 100 Percentage of Elite Solution 20% Crossover Percentage 50% Crossover Rate 50% Mutation Percentage 5% Mutation Rate 10% 37

38 Figure 22- Optimization option form of Monroe WDS project PEPSO II provides both tabular and graphical reports. Report Option form can be opened by clicking on Report Option icon ( ) of selecting Report Options item of Report menu. The top part of the report option form in Figure 23 shows all different section of text report that you can select to see in the final report of PEPSO II. By checking the EPANET file checkbox, you can receive the final optimum pump schedule integrated into an EPANET file of the WDS model. Both best pump schedule and optimization trend graph can be shown and saved during the optimization process. To store all available report of the Monroe WDS sample optimization run you need to have more than 110 MB available hard disk space. 38

39 Figure 23- Report options form of Monroe WDS After looking at all inputs, you can start optimization process by clicking on Run icon ( ) or selecting Run item from Optimization menu. The optimization process has three phases. The pre-optimization phase takes less than a couple of minutes and then the iterative optimization phase will start. During the pre-optimization phase, you can see an initial 1% progress on the progress bar in the lower part of the main form. At this phase, PEPSO II evaluate inputs and available resources of the computer system. Therefore, when the iterative optimization process starts, you can see an estimate of required time for finishing the process in the result section of the main form. You should note that this time is an estimate and as the optimization is not a deterministic process that time will change based on the status of solutions and available system resources at different stages. During the optimization process, the result table that is placed in the result section of the main form shows results of the best solution and average of the final list of qualified solutions (Pareto frontier) at the current iteration. After finishing iterative- 39

40 optimization phase, PEPSO II enters the post-optimization process. At this phase, the progress bar shows 99% progress and in a couple of minutes PEPSO II saves and shows the final results. You can temporarily pause the optimization process or stopping it completely by clicking on Pause and Stop icons ( and ) respectively. These actions can also be done by selecting Pause and Stop items of Optimization menu. Please note that when you click on stop button, optimization will continue to the end of the current iteration and save the result of the current iteration as the final results. A sample result of Monroe WDS optimization is presented in Table 5. Please note that, as the optimization process has random components and is not a deterministic process, you might see slightly different results when you run the provided sample project. A sample pump schedule of final results of Monroe WDS optimization run are shown in Figure 24. Table 5- A sample result of Monroe WDS optimization with PEPSO II Item Value Electricity Cost ($) Energy Consumption Cost ($) Power Demand Cost ($) CO 2 Emission (kg) Total Penalty Water Level Penalty at Tank Pressure Penalty at Junction Peak Power Demand (kw) Total Unadjusted Energy Consumption (kwh) Stored Water Volume Change -3.2% 40

41 Figure 24- A sample final pump schedule of Monroe WDS optimization run Figure 25 displays a screenshot of optimization trend of a sample Monroe WDS optimization run. You can choose to show optimization trend of the best solution and/or average results of final qualified solutions (Pareto frontier). Also, you can customize the graph by selecting show or hide trends of different objectives and choose the linear or logarithmic scale for the vertical axis. 41

42 Figure 25- A sample optimization trend of Monroe WDS optimization run If you selected to save the text, graphs, and the optimized EPANET model, you should be able to find them in the project folder. Graphics will be saved as JPEG files. Text and tabular reports are in TXT format, and the EPANET model is in INP format. A sample text report is presented and explained in Appendix. 42

43 APPENDIX A - GLOSSARY Average Energy Consumption Charge: The weighted average of on-peak and off-peak energy consumption charge based on the length of on-peak and off-peak periods of an electricity tariff. Constraint Importance Multiplier: Is a user-defined factor which will be multiplied by calculated penalty value that is corresponding to a component of the WDS to increase or decrease its effect on the total penalty value of a solution. For instance, is water pressures at two strategic junctions of a WDS show the same amount of violation but constraint importance multiplier of the first junction is two times more than the second junction, penalty value that is associated with pressure violation of the first junction is twice more that penalty value of the second junction Emission Factor (Emission Rate): a number with the pollution weight over energy consumption dimension (e.g. lb/kwh) that if multiplied by energy consumption results in pollution emission associated with energy consumption Energy Consumption Charge: Is cost of consuming one unit of energy (e.g. $/kwh). Multiplying energy consumption charge by the amount of consumed energy by a pump results in total energy consumption cost of the pump. EPANET input file: is a *.inp file that has all information of the hydraulic model of a WDS. PEPSO II needs this file to optimize a WDS. For more details, please refer to EPANET user manual (Rossman 2000). Exploitation: Fine tuning good solution to improve their quality and get closer to the optimum point or visiting surrounding area of the current solution to find a slightly better solution that is located around them. The crossover operator of GA is mostly used for exploitation process (CrepinˇSEK, Liu et al. 2011). Exploration: Searching for new solutions by visiting new areas of the solution space that have not been discovered. It helps algorithm to prevent getting stuck in a local optimum and increase the chance of finding the global optimum in non-convex problems. The mutation operator of GA can be used to help exploration process (CrepinˇSEK, Liu et al. 2011). LEEM report file: Is a comma separated value (*.CSV) file which has emission factors of the current, past and future times of the requested location. Duration of data in 43

44 the past and future that are reported depend on the location and time of the query. For instance LEEM 2.5 is able to report between 6 to 37 hours of emission factor prediction based on latitude and longitude of the query. Optimized EPANET file: is the final output of optimization process of PEPSO II in the form of a *.inp file which is similar the initial EPANET input file but its pump control section is filled based on the pump schedule of the optimum solution of PEPSO II Optimum Solution: usually a local optimum and occasionally a global optimum solution of an optimization problem. In this specific case, the optimum solution is an optimum pump schedule that satisfies the hard and soft constraint of the problem (e.g. tank level controls, pressure limits, etc.) and minimizes the other objectives (e.g. electricity cost, pollution emission,etc.). Pareto Frontier: Pareto frontier is a set of Pareto optimal solutions that are better than other solutions with respect to all objectives but cannot dominate each other in respect to all different objectives. All solutions that are members of a Pareto frontier are better that other solutions with respect to at least one objective value. Peak Power Demand: peak power demand of an electricity meter can be calculated as a maximum power demand of the electricity meter during a defined billing period (e.g. one month) that is measured in a defined time intervals (e.g. 30 minutes intervals). For calculating peak power demand of an electricity meter at a time block, required the power of all pumps that are connected to the electricity meter at that time block will be added up. Penalty: a numeric value that is calculated based on the amount of violation from a defined constraint. For instance, if maximum allowed pressure of a junction is 25 meter of water head, a junction pressure equal to 30 meter shows 5 meters violation and when the violation raised to the power of 1.5 (or any other defined arbitrary number as a penalty power) final amount of pressure violation penalty is =11.18 Population: collection of a group of solutions Power Demand Charge: Is cost of demanding one unit of power (e.g. $/kw). Multiplying power demand charge by the peak power demand of a pump results in total peak power demand cost of the pump. 44

45 Project file: Is a file that is created by PEPSO II based on project definition which is provided by the user via the user interface. This file can be manually edited by text editors. The project file has required information for running an optimization simulation by PEPSO II and includes, electricity tariffs, electricity meter data, pollution emission scenarios, optimization options, reporting options, initial population, WDS component constraints, etc. Relative Rotational Speed: Rotational speed of a variable speed pump with respect to its maximum rotational speed. It can be a number between 0 to 100% which 100% is maximum rotational speed of the variable speed pump Solution: a pump operation schedule that defines on or off status of fix speed pumps and rotational speed of variable speed pumps Solution Space: The solution space of pump operation optimization problem is a collection of all possible combination of the operational status of pumps of a system. For instance solution space of a pair of constant speed pump and variable speed pump that the variable speed pump can work at 0%, 75% and 100% of its nominal rotational speed is: [(off,0%), (off,75%), (off,100%), (on,0%), (on,75%), (on,100%)] Strategic Junction / Strategic Pipe: strategic junction or pipe is an important component of a WDS which can act as an indicator of the status of surrounding component or the whole WDS. It means that for instance, by adjusting the pressure of a strategic junction within the desired range, we can make sure that pressures of other surrounding junction or even all junction in WDS are within acceptable range. 45

46 APPENDIX B - SAMPLE OPTIMIZATION RUN The content of the project file of sample optimization run on Monroe WDS. The content of the project file is organized into seven sections. The first section is Main Section which includes required data for filling the constraint form of PEPSO II. Following that, the Water Network Section has a list of all pumps and tanks of the EPANET hydraulic model. The next section ( Energy ) store information of both tariff and meter tabs of the electricity form. The Pollution Emission section stores information of the pollution emission form of PEPSO II. Similarly, the Constraints section contains all required information to fill pump, tank, junction and pipe tabs of the constraint form. Finally, Optimization Options and Report Options sections store information of optimization and report option forms of PEPSO II. These formatting rules are used for creating the project file content: Start indicator of section XXX : [XXX Start] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~ End indicator of section XXX : ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~ [Main Section End] >xxx <xxx Start indicator of subsection xxx : End indicator of subsection xxx : Section divider: 46

47 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%% Subsection divider: Content of the Monroe WDS.prj file: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%% [Main Section Start] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Project Name: Sample Project Project Folder: C:\Program Files (x86)\wsu\pepso II\Sample Project Create Date/Time: 1/21/ :00:00 PM Last Save Date/Time: 1/21/ :00:00 PM Last Run Date/Time: 4/4/ :29:03 PM Project Note: Start of Note " PEPSO II Sample Optimization Run - Monroe WDS " End of Note EPANET Input File Location C:\Program Files (x86)\wsu\pepso II\Sample Project \160424_Stable_System_VFD_Reviewed_MS_ inp Unit System: SI ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ [Main Section End] %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%% [Water Network Section Start] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >Pumps PMP-544 PMP-9 W-12 W-11 W-10 W-9 W-8 E-2 E-3 E-4 E-5 E-6 E-7 47

48 <Pumps >Tanks T-5 T-2 T-3 <Tanks ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ [Water Network Section End] %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%% [Energy Section Start] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >Electricity Tariff... >>List Tariff 1 <<List... >>Tariff 1 Tariff Name: Tariff 1 Tariff Note: Start of Note " D6-220 DTE Energy " End of Note Start Time of Tariff: 0:0 Power Demand Charge ($/kw): Power Demand Calculation Period (Minute): 30 Electricity Billing Period (Day): 30 Energy Usage Charge: Time (hr) Energy Usage Charge ($/kwh) <<Tariff 1... <Electricity Tariff >Electricity Meter... >>List Main Booster <<List... >>Pumps Without Meter <<Pumps Without Meter 48

49 ... >>Main Meter Name: Main Meter Note: Start of Note " Treatment Plant Pump Station of Monroe " End of Note Meter Latitude: Meter Longitude: Applied Tariff: Tariff 1 Pump List: W-12 W-11 W-10 W-9 W-8 E-2 E-3 E-4 E-5 E-6 E-7 <<Main... >>Booster Meter Name: Booster Meter Note: Start of Note " Booster Pump Station of Monroe " End of Note Meter Latitude: Meter Longitude: Applied Tariff: Tariff 1 Pump List: PMP-544 PMP-9 <<Booster... <Electricity Meter ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ [Energy Section End] %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%% [Pollution Emission Section Start] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >Pollution Emission Scenario Active Pollution Emission Scenario: Offline LEEM... 49