guitart-towards multi-objective efficiency water ... · pdf filetowards multi-objective...

Post on 07-Feb-2018

216 Views

Category:

Documents

1 Downloads

Preview:

Click to see full reader

TRANSCRIPT

Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision

Support System

G. Anzaldi, A. Corchero, X. Domingo, F. Guitart, J. Pijuan, E. Rubion, R. Sanfeliu

Bologna, 20th October 2016

Water-Ideas Conference

2Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

Outline

Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

1. Introduction

2. Framework Overview

3. Decision Support System Implementation

4. Example of Use

3Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

Outline

Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

1. Introduction

2. Framework Overview

3. Decision Support System Implementation

4. Example of Use

4Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

Introduction

• Decision-making over the Water Distribution Networks (WDN)

• Use hydraulic simulation tools (i.e. EPANET)

• Future operational status of a WDN using Hydraulic

simulation tools

• The resulting report is a big raw text

• Resulting report is very useful for specific analysis BUT

• Quite difficult to derive general conclusions relevant for

decision making in WDN

5Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

Introduction

• Decision Support framework

• Capable of providing recommendations for the multi-

objective efficiency

Energy Quality

Demand Maintenance

Multi-objective

Optimization

Results of the

simulation tools

Expert

Knowledge of

Water Managers among others

Decision Support Framework Overview

6Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

Outline

Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

1. Introduction

2. Framework Overview

3. Decision Support System Implementation

4. Example of Use

7Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

Framework Overview

• Split Information and Knowledge:

• FACTS: status and infrastructural elements (EPANET information and

network configuration)

o RULES: collect all the Water Managers’ knowledge referring the

WDN (known as expert knowledge)

Main advantages of using a rule based system in the DSS are:

(i) separation between decisional process and domain information

(ii) decisional process is guided by the rules (what to do) instead of lead by

them (how to do it)

(iii) good performance parameters in terms of execution speed and

scalability achieved by the Inference Engine;

(iv) the knowledge is centralised in the form of rules within the Rules

Repository

8Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

Framework Overview

9Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

Framework Overview

1. A Semantic Parser to extract the data from the hydraulic simulation

tool results.

2. The Water Management Ontology (WMOnt) provides a common

vocabulary abstracting the water infrastructural elements.

3. A Web Based Rule Editor to transform Expert Knowledge to a

Declarative Rule Language that the Inference Engine can interpret.

4. A Water Based Standard and a Standard Interface to structure

information and the way results are delivered.

10Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

Outline

Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

1. Introduction

2. Framework Overview

3. Decision Support System Implementation

4. Example of Use

11Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

Introduction

Decision

Support

Framework

• Applied• Implemented • Validated

• Implementation of a Decision SupportSystem (DSS)

• Real cases• Reuse of Expert

Knowledge• Integration in a IWMS• Use of specific tools

• DSS runs after each simulation of the WDN • Specific recommendations (warnings and alarms) • Increase the efficiency (multi-objective)• Custom configurable parameters

Decision Support System Overview

12Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

Decision Support System Implementation

For the implementation of the DSS the following decisions have been

taken:

(i) the simulation tool is EPANET which is imposed by the CBR and the real-

scenario;

(ii) the applied inference engine is Drools due to their robustness and

reasoning efficiency;

(iii) the ontology corresponds to the WMOnt fully aligned with existent

hydrological standards, data models and semantics;

(iv) the standard used to return the results is OGC®-WaterML2.0 widely

adopted for hydrologic information exchange;

(v) the open interface is OGC®-WPS, a standardized web service definition

for geospatial processing services.

13Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

Decision Support System Implementation

14Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

Outline

Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

1. Introduction

2. Framework Overview

3. Decision Support System Implementation

4. Example of Use

15Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

Example of Use

Example:

Let’s suppose that there’s a big city simulation case, and particular focus the use of the DSS to a reduced set of reservoirs (e.g. a tank).

# Rule Level Rule

1 Warning The minimum level of water should be 2m at any moment

to assure water quality.

2 Alarm Whenever the level is above 5m, all pumps filling the tank

must be switched off.

These rules are introduced by the WaterManager and transformed to a DRL rules file

16Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

Example of Use

Model

variables

Entities

from

Logical

Model

(tank,

pipes,

etc…)

Rules

Body and

Head of

the rules

User

defined

variables

Functions

17Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

Example of Use

rule "Tank"

when

$WR1: WaterResource(id == "FillingTank")

$FoI1: FeatureOfInterest(waterResource == $WR1)

$OF: Observation(feature == $FoI1, phenomenon.id == "Flow")

$OWT: Observation(feature == $FoI1, phenomenon.id == "WaterTable")

$TSOF: TimeSeriesObservation(observation == $OF)

$tvp1: TimeValuePair(value>0) from $ TSOF.values

$TSOWT: TimeSeriesObservation(observation == $OWT)

$tvp2: TimeValuePair(value>5) from $ TSOWT.values

$tvp1.getPosition()!=$tvp2.getPosition()

then

MESSAGES.add(new RuleMessage(

RuleMessage.MessageType.ERROR, “Tank ", String.format("Pumps from Tank are

running at %s when the water table is over 5m ",$tvp1.getPosition())));

End

18Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

Example of Use

<WaterResources>

<WaterResource id="FillingTank">

<FeaturesOfInterest>

<FeatureOfInterest id="Tank">

<Observation id="T_WLevel" phenomenon="WaterLevel" procedure="Simulation">

<TimeSeries table="Node" elementId="Tank" column="Head"/>

</Observation>

</FeatureOfInterest>

<FeatureOfInterest id="Pump1">

<Observation id="P1_Flow" p="Flow" procedure="Simulation">

<TimeSeries table="Node" elementId="P102783henomenon" column="Flow"/>

</Observation>

</FeatureOfInterest>

<FeatureOfInterest id="Pump2">

<Observation id="P2_Flow" phenomenon="Flow" procedure="Simulation">

<TimeSeries table="Node" elementId=" P102784" column="Flow"/>

</Observation>

</FeatureOfInterest>

</WaterResource>

</WaterResources>

19Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

Example of Use

The process begins when the OGC®-WPS send the execution order to the

DSS.

At this time, this execution function is accompanied with the required

execution information as:

(i) EPANET simulation result;

(ii) a DRL file with the Water Manager knowledge; and

(iii) an XML file with the model including the WDN assets to be considered

by the DSS.

20Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

Example of Use

The final output if pumps are running when the tank level is above 5m:

<wml2:Collection

xmlns:wml2="http://www.opengis.net/waterml/2.0"

xmlns:gml="http://www.opengis.net/gml"

xmlns:om="http://www.opengis.net/om/2.0"

xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"

xmlns:xlink="http://www.w3.org/1999/xlink"

xsi:schemaLocation="http://www.opengis.net/waterml/2.0 http://www.opengis.net//waterml//waterml2.xsd"

gml:id="alert_waterlevel_tank">

<gml:description>Pumps from Tank are running at 16:00 when the water table is over 5m </gml:description>

<wml2:metadata>

<wml2:DocumentMetadata gml:id="WatERP.DocMD.1">

<wml2:generationDate>2016-10-15T15:10:00+02:00</wml2:generationDate>

<wml2:generationSystem>DSS</wml2:generationSystem>

</wml2:DocumentMetadata>

</wml2:metadata>

<wml2:temporalExtent>

21Towards Multi-Objective Efficiency in Water Distribution Networks through a Decision Support System

Conclusions

1. Minimization/maximization of WDN parameters through a

DSS

2. Automatic analyisis of simulation results

3. Apply the same reasoning process that a Water Manager

would have followed

4. Enhancement of the functionality of EPANET results

5. Provide semantic capabilities to EPANET outputs

6. Centralization of Water Manager knowledge

7. Use of standards and ontologies for Interoperability

achievement

Thank you!

Francesc Guitart, PhDfrancesc.guitart@eurecat.org

www.eurecat.org

top related