Tim Baldock Anna K. Portela 29 January 2019 | Birmingham Target of - - PowerPoint PPT Presentation

Tim Baldock Anna K. Portela

29 January 2019 | Birmingham

▪ Target of data collection, processing & analysis for Smart Site development

▪ Pla lant-wid ide vis isib ibili lity (via Dashboards) ▪ Key Perf erformance In Indic icators (via Real Time Control) ▪ Oper erati tional In Intell elligen ence ▪ Process optimisation

– 10% of sites enabled RTC (upfront costs are high; need for benefit case assessment →

• ffline model)

– 90% of sites are dashboarded (to fill gap in RTC installed base; most recent development of 2nd Generation Dashboard)

2

3

10 % Incremental implementation

• f RTC

Modelling for development of business case to install online RTC Achieve maturity in building

• ffline/online dynamic models

for Dashboards and RTC applications

RTC

90 % Offl fline da dashboards, mod

• dels

100 %

Fully Fully con

• nverged

“Smart Site”

1-2 yrs 2-5 yrs 5-8 yrs 8-10 yrs

• Past & present- “black-box” models (averaged historical data

and rule-of-thumb empirical process efficiencies)

• Present & future- dynamic, transparent plant-wide models

(telemetry + field data) for RTC

4

Source Field Data Telemetry / PI System Telemetry + Field Data

• Spot samples
• Real-time process data for back-end and

inter-stage monitoring

• Historical events analysis
• PI tags

Data Points Dashboards Models

• Telemetry + Field Data- enabled to track KPI

Intelligent Data Management

▪ Spot samples- SAP BI Sample Reports ▪ Telemetry data- PI Datalink ▪ Dynamic visualisation the data-time series ▪ PI Process Book & Prism

5

▪ Telemetry + Field Data used to build KPI-driven dashboards ▪ Red/ Amber/ Green Status for KPIs (Performance Management)

▪ Enables Operations to achieve Operating Plan targets ▪ Continuous process improvement

6

STC 1 STC 2 STC 3 STC 4 STC 5 STC 6 STC 7 STC 8 STC 9 STC 10 STC 11 STC 12 STC 13 STC 14 STC 15 STC 16

▪ Pre-Dashboard

▪ look-up target sheet

▪ 1st generation Dashboard

▪ Involves threshold data analysis (against the target value) ▪ Data sources: telemetry data; operating plan

▪ 2nd generation Dashboard

▪ involves linear and non-linear regression analysis ▪ data sources: black-box model outputs; telemetry & field data; design parameters

▪ 3rd generation Dashboard

▪ as 2nd generation Dashboard + dynamic model outputs (future Dashboard)

▪ Ultimate target – rob

• bust RTC via interlin

inked Da

Dash shboards

7

8

▪ Generation of the improved MLSS Dashboard ▪ Determine costs of running Activated Sludge plants outside the operating MLSS range set in the operating plan ▪ Identify benefits/spend comparing to the baseline

500 1000 1500 2000 2500 3000 3500 4000 4500 11/12/18 16/12/18 21/12/18 26/12/18 31/12/18 05/01/19 10/01/19 15/01/19 20/01/19

Morestead Road Winchester

Lane 1 Lane 2 Lane 3 Minimum Maximum

MLSS

Telemetry data Operating plan targets

Existing 1st Generation MLSS Dashboard

9

Inputs Black-box model Output analysis

Physical design parameters Operating plan targets £/kWh = f (MLSS) Telemetry / PI Datalink Sample reports

Set the MLSS baseline

Operating plan targets Telemetry / PI Datalink Unit energy costs

Outputs 2nd Generation Dashboard

Linear regression Physical design parameters Analytical formulation Method verification

10

▪ Site selection

▪ Size, process type (carbonaceous, nitrifying, BNR,...) and asset type (diffused air system, mechanical/surface aeration)

▪ Model output

▪ Energy demand on aeration as a function of MLSS concentration ▪ Data sources: PI Datalink, spot sample results, O&M manuals, operating plan ▪ Calculations performed for over 30 sites ▪ Slope-intercept form of each E=f(MLSS) curve were consecutively analysed and compared against process design parameters

11

▪ Linear regression analysis- employed to determine possible correlations of the data sets with process design parameters ▪ The accuracy of the correlation developed for energy demand computation was validated against the values originated from the oxygen demand calculators

12

▪ Energy expenditures determined for the actual MLSS concentration- identification of potential benefits/costs comparing to the baseline ▪ Up to date- dash boarded 56 sites ▪ £300k/year benefit identified

2nd Generation MLSS Dashboard

Telemetry data Operating plan targets Baseline

Total variance over month

▪ Plant-wide modelling WTW & STC ▪ SW Team: Optimisation and Process Capacity & Modelling ▪ University of Portsmouth

13

WTW & STC model- identify quick win concepts & verify benefits from “black-box” mass balance models ▪ BioWin

▪ Evaluation of the unit process efficiencies & overall treatment performance ▪ Costs analysis for different operating scenarios (flows and loadings, choice of alternative C source for BNR)

▪ BW Controller

▪ Evaluation of the existing control philosophy & its optimisation towards more sustainable control strategies (e.g. DO, N-DN and SRT control, recycle rates, chemical dose rates, etc.)

14

Data points sources ▪ Physical details & operating parameters

▪ O&M Manuals, Drawings

▪ PI Datalink ▪ Prism ▪ Field data

▪ Dedicated sampling campaign ▪ Spot sample database

Goo

• od qualit

lity data required for model calibration and validation!

15

16

Hourly power cost

17

▪ Risk-free evaluation of different operating scenarios for effective optimisation

▪ Process performance under variable flow and loading conditions ▪ Changes in process parameters, configuration and control philosophy

▪ Re-assessment of the potential benefits & additional costs

▪ Biogas production ▪ Energy demand on aeration ▪ Optimised chemical dosing and tankering ▪ Capex required for optimised treatment performance

▪ Key innovations

▪ Use of plant-wide dynamic models for Dashboards and RTC for “Smart Sites”

18

19