Control possibilities in wastewater treatment

Transcription

1 Control possibilities in wastewater treatment Gustaf Olsson Lund University, Sweden Università Degli Studi di Palermo 13 March 2014 Content Introduction - why ICA (Instrumentation, Control and Automation) A basic activated sludge plant The energy issue Disturbances The role of control and automation Examples of control applications Monitoring Plant wide control 2 ICA Instrumentation, Control and Automation Why monitoring and control? The system is dynamic Using real time sensors to measure is to know Feedback drive the process towards high performance all the time Maximize efficiency Variable influent (disturbances) Effluent requirements Consistent operation 1. Keep the plant running control equipment 2. Satisfy the limits 3. Minimize the costs Economic gains in design Economical gains in operation Saving energy and chemicals 3 4 Gustaf Olsson, Lund University, Sweden 1

2 Driving forces demand pull Driving forces - technology Regulatory requirements Energy efficiency Nutrient recovery Heat recovery Biogas production Disturbance resilience ICA impact on design Instrumentation Computer revolution From kilobytes to gigabytes Data acquisition Model representation Generations of SCADA systems Networking and Internet Actuators and power electronics 5 6 Saving resources and energy Content Saving energy Dissolved oxygen control Pumping control Coordination of many unit processes Optimal sequencing in sequential batch reactors Saving chemicals Dosage control Producing more biogas energy Monitoring and control of AD Introduction - why ICA (Instrumentation, Control and Automation) A basic activated sludge plant The energy issue Disturbances The role of control and automation Examples of control applications Monitoring Plant wide control 7 8 Gustaf Olsson, Lund University, Sweden 2

3 Towards more ICA in wastewater treatment Pre-denitrification plant Only C removal Less restrictive legislation Few instruments Little economic incentive Biological N and P removal More elaborate effluent rules On-line nutrient measurements Advanced model knowledge Increasing economic pressure Education Chemical P removal Gradually stricter legislation More computing power Better model understanding In flu e n t A n o x ic r e a c t o r A e r o b ic r e a c t o r Internal recirculation S l u d g e r e c i r c u l a t i o n S l u d g e o u t t a k e E fflu en t 9 10 Ammonia concentration along the aerator - oxidation Nitrate concentration along the anoxic zone - reduction Ammonia conc mg/l Low load Ammonia removal Nominal load High load NH4 + NO3 - Nominal High load Low load Measure the ammonium conc. Denitrification rate Increase nitrate recirc. Decrease nitrate recirc Along anoxic zone Measure the nitrate conc. Desired DN rate Too slow DN Too fast DN NO3 - N 2 Aeration zone Gustaf Olsson, Lund University, Sweden 3

4 Plant design for bio-p removal In flu e n t B io - P re a c to r External carbon A n o x ic r e a c t o r S l u d g e Air supply A e ro b ic re a c to r In tern a l recirc u la tio n r e c i r c u l a t i o n Chemicals S l u d g e o u t t a k e Control actions have different costs E f f l u e n t Rya Plant Sweden Aerator Sludge in microscope Gustaf Olsson, Lund University, Sweden 4

5 Content Introduction - why ICA (Instrumentation, Control and Automation) A basic activated sludge plant The energy issue Disturbances The role of control and automation Examples of control applications Monitoring Plant wide control 17 Swedish averages Water distribution 0.22 Drinking water treatment Source: Lingsten et al Swedish Water Water use > 50 kwh/m Pumping to waterworks Pumping sewage Wastewater treatment kwh / kg BOD xx kwh/ kg N xx kwh / pe / yr 18 Clean Water Requires Energy! Pumping Having efficient pumps Increase efficiency! for adequate flows Operating at dynamically changing flows and pressures Aeration in wastewater treatment Minimize air flow! Adequate compressors Controlling the air flow for variable loads Recirculation nitrate 3% Mixing anoxic zone 11% Electrical energy for the Biological Plant Some German data ( pe) Aeration 83% Return sludge pumping 3% Gustaf Olsson, Lund University, Sweden 5

6 Future water supplies. Content will be more energy intensive Readily accessible fresh water supplies are limited and have been fully allocated in some areas Increased energy for pumping (deeper longer) New technologies to access/treat water will use more energy Impaired, reused, brackish, sea water Introduction - why ICA (Instrumentation, Control and Automation) A basic activated sludge plant The energy issue Disturbances The role of control and automation Examples of control applications Monitoring Plant wide control Disturbances influent small WWTP Disturbances influent large WWTP Flow N tot COD tot P tot Gustaf Olsson, Lund University, Sweden 6

7 Temperature and influent flow rate A rainy period with sludge escape Temp Flow l/s Celsius t = 15 min Temperature Meausurements (week 50-52) Flow to line t = 15 min Meausurements (week 50-52) 3 weeks (Nov Dec) 3 weeks (Nov Dec) Data from Källby, Lund, Sweden Rain intensity Flow rate Susp solids Interaction - recycles Water supply - pipe failure Oxygen uptake rate Reject water flowrate mg/l/h m3/h days days (from Nielsen et al.) Burst an event Leak a condition Gustaf Olsson, Lund University, Sweden 7

8 Content Introduction - why ICA (Instrumentation, Control and Automation) A basic activated sludge plant The energy issue Disturbances The role of control and automation Examples of control applications Monitoring Plant wide control Actuators Computer implementation Disturbances Processes - Plant The human Operation and control Sensors Supervision Time Scales of Control in WWTP Simple feedback control Days - months strategies Days - weeks biomass growth Hours to days concentration dynamics nutrient removal Minutes to hours flow dynamics dissolved oxygen Management Supervisory control Advanced process control Basic control Setpoint = reference value + Error Controller Control signal -1 External disturbances Process Output = measurement Gustaf Olsson, Lund University, Sweden 8

9 Levels of control and automation example water systems Aspects of automation Keep the plant running Control of machinery: motors, pumps, valves, etc. Satisfy the product quality Satisfy effluent quality regulations Minimize the cost and maximize the efficiency Minimize electrical power use, chemicals, etc. Planning/ Optimization Modelling/ Simulation Measurement Process Supervision Communication Control Added controllability (1) Added controllability (2) Bioreactors anaerobic, anoxic, aerobic zones More sophisticated air supply Separate control of zones Variable pressure control More intermittent systems (SBR) Aerated tank settling operation More recirculations e.g. nitrate recirculation Chemicals added enhanced primary clarification chemical P removal Volatile fatty acids enhance Bio-P External carbon control denitrification Gustaf Olsson, Lund University, Sweden 9

10 Content DO control using variable speed compressors Introduction - why ICA (Instrumentation, Control and Automation) A basic activated sludge plant The energy issue Disturbances The role of control and automation Examples of control applications Monitoring Plant wide control Without speed control DO DO (Bo Brink, Laholm) With speed control Growth rate as function of substrate & DO concentrations DO sensors Growth rate % Substrate concentration DO concentration 4 K o =1 K s =20 Matlab: monodcurve3d Gustaf Olsson, Lund University, Sweden 10

11 DO control = cascade control Dissolved oxygen (DO) control Airflow controller (slave) Airflow setpoint Actuator Air valve Air flow Air supply DO setpoint DO Aerator DO controller (master) Valve opening Dissolved oxygen conc Influent flow rate Ronneby WWTP Sweden Pre-denitrification plant Ammonia concentration along the aerator Compromize in DO requirements In flu en t E f f l u e n t A n o x ic r e a c t o r A e r o b ic r e a c t o r Internal recirculation S l u d g e r e c i r c u l a t i o n S l u d g e o u t t a k e No oxygen Sufficient oxygen Correct airflow rate Concentrations Nitrification Too high airflow rate Too low airflow rate time (hr) We can calculate the best DO conc. setpoint! Gustaf Olsson, Lund University, Sweden 11

12 Aeration closed loop control DO setpoint control Variable speed drive Determine DO setp. Compressor PLC Aerator Air flow Communication DO sensors Ammonia sensor Performance of NH 4 based DO controller Ammonia conc. (Setpoint=3) DO setpoint value Variable DO setpoint further 10-15% savings Ingildsen Variable DO setpoint control Unit process control DO NH 4 setpoint setpoint DO setp. controller DO controller Airflow setpoint Airflow controller Valve opening Air flow system Air flow Aerator DO NH 4 DO control Hundreds of papers since the 1970s How complex controller is needed? Controlling the DO profile DO setpoint control (ammonia sensor) Nitrate recirculation Return sludge, sludge age Dosage control (Ext. C, chemicals) DO setpoint control DO control Air flow control Gustaf Olsson, Lund University, Sweden 12

13 P removal using chemical precipitation principle Chemical dosage based on hydraulic load Precipitation chemicals Phosphates 1. Dosing of precipitation 2. Precipitation chemicals 3. Coagulation 4. Flocculation Flocs/particles sedimentates 5. Sedimentation Effluent PO 4 -P conc. Days Ingildsen Chemical dosage based on P sensor feedback Content Effluent PO 4 -P conc. Saving 36% chemicals Days Ingildsen 51 Introduction - why ICA (Instrumentation, Control and Automation) A basic activated sludge plant The energy issue Disturbances The role of control and automation Examples of control applications Monitoring Plant wide control 52 Gustaf Olsson, Lund University, Sweden 13

14 Instrumentation today Process monitoring (1) Almost 100 sensor companies working with water Sensor networks Data fusion Internet the ubiquitous control room All measurements need to be tested and verified! Cross-check Automatic actions in alarm situations Often thousands of simple measurements Binary sensors Simple physical measurements Process monitoring (2) Example - variability sensor problem Track important variables and states in the process and to relate their values to a reference or baseline Updated knowledge on the process state Detect deviations in the performance of the process Isolate the source of the deviation threshold 55 (from C. Rosen) 56 Gustaf Olsson, Lund University, Sweden 14

15 Process monitoring (3) PCA Principal Component Analysis Univariate monitoring Analysis of an individual signal or measurement. Multivariate monitoring Analysis of several signals or measurements and their mutual relationships Look for abnormal behaviour x 2 PC 2 PC 3 x 3 PC 1 57 x 1 58 Monitoring process signals deviation Principal components (from C. Rosen) Gustaf Olsson, Lund University, Sweden 15

16 Detecting deviations deviation Water Supply Steady-state analysis-based failure monitoring Permanent installation for continuous break monitoring Applicable in the DMA setup Flow at the entry and pressure at selected locations measured Detection and location of nodal and non-nodal bursts Condition demand information is available and the model of the system is calibrated Flow rate Pressures Water Supply Unsteady-state analysis-based failure monitoring Permanent installation for continuous burst monitoring Condition the burst induces a transient wave Immediate detection and location of nodal and nonnodal bursts Location error less than 10 m Can be combined with steadystate approach Pressures H, m M1 M3 M2 M1 M3 M2 Content Introduction - why ICA (Instrumentation, Control and Automation) A basic activated sludge plant The energy issue Disturbances The role of control and automation Examples of control applications Monitoring Plant wide control Time, s 64 Gustaf Olsson, Lund University, Sweden 16

17 Plant-wide control System wide control Influent wastewater Primary clarifier Activated sludge Secondary reactors clarifier Mixing We need a structured way to Thickener Gas Anaerobic Plant-wide digester control! Dewatering Additional loads Bypass 2-10 controllers 2-10 controllers 2-5 controllers 4-40 controllers 2-4 controllers Controllable flow rate Valve Sludge removal Tertiary clarifier 2-10 controllers coordinate all control actions! pumps s valves and gates controllers 2-5 controllers (from C. Rosen) 65 Effluent water Kukudis, 1973: We must speak of automation in the entire system -- the network of sewers and the plants. Sewer control was applied in Cleveland in the early 1970s 66 Why plant wide? Approaches to plant wide Energy efficiency Where to use the carbon Influence of recirculations Relate sludge production to desired gas production Handle varying loads disturbance rejection strategies Minimize influence on receiving waters Development towards integrated modelling Decision support systems (DSS) to deal with complexities of decision making Math models Control algorithms Knowledge-based techniques Experience from operators Existing databases Gustaf Olsson, Lund University, Sweden 17

18 Compromize Sewer operation: Minimize the combined sewer overflow (CSO) WWTP operation: Satisfy the effluent requirements Minimize the operational cost 69 Integrate design and control Computers: Take advantage of the huge storage capacity and processing capacity Make use of new instruments: bioreactors, settlers, sewers, anaerobic processes Signal treatment & monitoring: Exhaust the information, operator support, automatic detection 70 Control: which complexity is needed? System wide control: How to formulate criteria? Handle the complexity of couplings. Making user friendly systems. ICA in the whole water cycle: Non-conventional water systems. The automated decentralized WWTP Drinking water treatment. Smart water grids. Water reuse Gustaf Olsson, Lund University, Sweden 18

19 Automation because The control challenge Disturbances are everywhere Compensate for them On-line sensors no longer the main limitation for on-line control Data acquisition monitoring Early warning Control Saving energy and resources Producing energy Consistent operation Understanding the process Making the best use of sensors Control structure Having actuators with sufficient control authority Exhaust the measurement information Understanding the control criteria Contact Gustaf Olsson Gustaf Olsson, Lund University, Sweden 19