Refinery Wastewater Process Modeling with GPS-X

Transcription

1 Refinery Wastewater Process Modeling with Malcolm Fabiyi, PhD Hydromantis USA John Joyce, PhD PE SAGE Environmental February /8/16 Page 1

2 Outline Issues in refinery wastewater Typical Refinery Wastewater systems Overview of Hydromantis Modeling tools and their application Industrial WWTP case studies SimuWorks Conclusion 3/8/16 Page 2

3 Refinery Wastewater Issues Varying crude stocks and where those crude s come from present differing wastewater issues to refineries --- very much more so now than in the past.. Refineries are also now evaluating: End of life API s, DAF s Evaluating higher rate DGF s and DNF s And evaluating how capital cost can be deferred by optimizing existing wastewater process units 3/8/16 Page 3

4 Refinery Wastewater Issues Refineries are also facing tightening metals discharge limits Mercury Selenium Arsenic Many membrane bio-reactors being designed & installed Troublesome Inorganics --- cyanides, salts Toxicity/bioassay compliance Need for surge capacity Reducing treatment costs 3/8/16 Page 4

5 Refinery Wastewater Systems De-Salters Sour Water/ Scrubbers Distillation Blowdown Stormwater Cooling Tower Blowdown Tank Drains 3/8/16 Page 5

6 Refinery Wastewater Issues If you are responsible for managing the wastewater system at a refinery or chemical plant Even the more simplistic. What is my optimal MLSS What is my optimal aeration rate What is my optimal RAS/WAS rate Or the more complex.(before you have to engage a wastewater consultant) What happens if a biological contactor, MBR, or new DGF is installed in this location or that location? 3/8/16 Page 6

7 Refinery Wastewater Issues There are two fantastic modeling tools that refinery and chemical plant wastewater managers and engineers should avail themselves of: Dynamic Wastewater Treatment Modeling Process design Energy models GHG models Controller models Real-time Optimization Operator Training and Plant Analysis System Customized interface platform Operational support Planning Training 3/8/16 Page 7

8 Traditional Design Models Based on simplified mass balance models Assume steady-state Constant influent loading and operating conditions No change in solids storage in clarifiers or aeration basins Equations solved to predict tank sizes, sludge production, and performance given SRT, MLSS, etc. Design parameters from simple models serve as inputs to more comprehensive dynamic models 3/8/16 Page 8

9 Why Use Modeling? Cheaper than building/modifying the real system Easier than carrying out testing on existing systems Risk-free - see the consequences before implementation Get results quickly 3/8/16 Page 9

10 How Is It Used? Create Model Calibrate to Known Performance Simulate Different Scenarios WAS TSS Concentration (mg/l) Simulate Base Case Compare and Evaluate Time (days) Simulated Measured 3/8/16 Page 10

11 Features of Industrial Activated Sludge Can have high soluble COD fraction High temperatures Highly variable influent characterization Nutrient deficiency common Varying sludge settleability Dynamic Can have multiple influent streams of interest (VOCs, Oil & Grease, Hydrocarbons, Phenols, Colloidal materials, Sulfur, Amines, Metals) Can have high strength fractions (e.g., Spent caustic) 3/8/16 Page 11

12 Physical/Chemical Processes? Handles biological treatment and solids separation As well as other processes Oil and grease separators ph management PAC Chemical addition 3/8/16 Page 12

13 Case Studies 3/8/16 Page 13

14 - Features 3/8/16 Page 14

15 - Features Intuitive modeling. and simulation views Quick panel views for all objects Sankey Graphs 3/8/16 Page 15

16 - Features 3/8/16 Page 16

17 - Features Select Control Variables Clarifier behavior Sensitivity analysis Treatment thresholds 3/8/16 Page 17

18 - Largest Number of Unit Objects 3/8/16 Page 18

19 Supports a Broad Range of Custom Projects Selenium removal API/CPI separators Oil & grease removal 1,4 Dioxane removal in WWTP Process water simulator Capacity evaluation Biotransformation of estrogenic hormones* Water reservoir simulation Coupled CFD & process dynamics Calibration of PAC + IFAS treatment process Customized process simulators *Natural estrogenic hormones 17βestradiol (E2) and estrone (E1), and the synthetic estrogen 17α-ethinylestradiol (EE2) 3/8/16 Page 19

20 Case Study 1: Develop a calibrated model of a petroleum refinery WWTP to support plant analysis, decision making and process optimization Plant layout 3/8/16 Page 20

21 Case Study 1: Develop a calibrated model of a petroleum refinery WWTP to support plant analysis, decision making and process optimization Plant layout New API unit developed has the largest process object library of any process modeling software 3/8/16 Page 21

22 Case Study 1: Develop a calibrated model of a petroleum refinery WWTP to support plant analysis, decision making and process optimization provides accurate representation of a facility s process layout 3/8/16 Page 22

23 Case Study 2: Development of model for assessing impacts of loading changes, operational changes, toxic spills, etc. Chemical Manufacturing Facility, WWTP Diversion Neutralization Step-Feed Activated Sludge Clarifiers Influent Grit Basins Equalization Belt-filter Presses Treats waste from plant that manufactures chemicals, plastics, and synthetic fibers 3/8/16 Page 23

24 Case Study 2 Chemical Manufacturing Facility, WWTP Diversion Neutralization Step-Feed Activated Sludge Clarifiers Influent Grit Basins Equalization Belt-filter Presses Q = 27 MGD Organic acids & alcohols (COD = 1700 mg/l) Solids (TSS = 140 mg/l) Ammonia, nitrates, phosphates Some denitrification 4 trains with 3 aeration basins each Step feed strategy varies Mixed liquor sampled from each basin 3/8/16 Page 24

25 Case Study 2: Model Setup - Layout Activated Sludge Elevated temperatures (26 to 36 degrees C) ASM3 default kinetic values allowed better calibration than values of Koch et al. (2000) Decreased: Maximum autotrophic growth rate to match ammonia in basins Added simultaneous nitrification/denitrification to ASM3 to handle observed denitrification in aeration basins TSS Concentration (mg/l) Time (days) WAS TSS Concentration (mg/l) Time (days) Total COD Concentration (mg/l) Time (days) Simulated Measured Simulated Measured Simulated Measured 3/8/16 Page 25

26 Case Study 3: Pharma Facility Plant treats pharmaceutical wastewater using MBBRs (with fungal and bacterial reactors), PAC, and DAFs Developed, calibrated, & verified plant model Model capable of predicting behavior of fungal biomass Model used to assess impact of number of aeration trains in service, influent loading, temperature, airflow, & nutrient addition Hydromantis, Inc. TSS (mg/l) Time (days) Existing One Train Out of Service Soluble COD (mg/l) Time (days) Simulated Measured 3/8/16 Page 26

27 Case Study 4 Developed customized tool that is used to: Provide predictive analyses for plant operation, e.g. spill control assistance Help GE meet the effluent criteria Help optimize plant operation Provide a customized operator training platform Savings to operators of > $36,000 USD per year 3/8/16 Page 27

28 Case Study 5 WWTP treats waste from 3 chemical production facilities Studies conducted using : Capacity evaluation Sensitivity of plant performance to nitrogen loading to the plant Bottleneck evaluation Determined aeration power requirements Dynamic analysis of plant Safety margin for plant capacity Development of management plan for facility Used industrial pollutant library (IP) to track specific chemicals 3/8/16 Page 28

29 Case Study 6 Developed model of WWTP at chemical plant Activated sludge treatment system model used to: Investigate aeration control options Investigate opportunity for MLSS control Investigate impact of loading changes to the plant Determine best approach to handle high-flow events used to model other DOW plants 3/8/16 Page 29

30 Applications of Modeling in Industrial WWT Applications Application areas Modeling focus Software tool Aeration demand O 2 for supplementation O 2 for organic treatment VOC & Odor emissions reporting and analysis Process optimization, planning, training & knowledge mgt. Contaminant degradation mechanism Metals & organics Emissions reduction Contaminant fate Odor stripping Plant optimization Production throughput constraint evaluation Product, by-product, metabolites Speciation, transport and fate Toxchem & SimuWorks & Toxchem Bulking control O 2 demand vs organic loading Aeration efficiency Oxygen transfer efficiency Fouling estimation Spent caustic treatment Incremental O 2, VOC stripping implications, Capacity validation, ph effects & Toxchem Overall process design & optimization Process tank sizing & configuration Capital cost, operating cost Alkalinity, ph CapdetWorks & GPS- X 3/8/16 Page 30

31 Next Generation Uses of Models Engineers are trained in refinery processes It took me about 5 months to understand the activated sludge process Engineer at Major Refinery Facility.Sometimes do a rotation in the WWTP 3/8/16 Page 31

32 SimuWorks Overview Advanced operator training and education Plant optimization and cost saving strategies Project and risk analysis Key decision making support tool 3/8/16 Page 32

33 Simulator Interface Clickable Diagram Simulation control Output graphs / input panel Flyover panel 3/8/16 Page 33

34 Step 1: Step 2: Step 3: Step 4: Click on Unit Process Adjust Parameters Press Start See Results 3/8/16 Page 34

35 Example #1: Troubleshooting Nitrification Initial Case: poor performance not nitrifying, poor COD removal Influent quality: COD = 750 mg/l TKN = 100 mg/l Ammonia = 80 mg/l Soluble P = 12 mg/l ph = 7.0 Temperature = 28C (82F) 3/8/16 Page 35

36 Conclusions Activated sludge models can be applied to industrial WWTPs Modeling is powerful, cost-effective tool for design, process analysis, trouble-shooting, optimization, planning, and training Modeling is becoming a required process skill 3/8/16 Page 36

37 Technical Support 3/8/16 Page 37

38 Material Copyright Hydromantis Environmental Software Solutions, Inc. All rights reserved Wastewater modeling and design solutions Dr. Malcolm Fabiyi, PhD Dr. John M. Joyce, PhD PE 3/8/16 Page 38