Ruel based decision support for the process flow

1
Ruel based decision support for the process flow

• Embedding SIMONE optimisation modules in a
  Knowledge and rule based process

2
Rule based decision support for the process flow
- Contens -

• Introduction
• Process flow of an optimisation ?
• Knowledge based system ?
• Rule based system ?
• Rules for compressor plant configuration
• Pressure rules

3
Introduction
Transport optimisation is a highly combinatorial
Problem
4
Introduction - Compressor Plant -
First level Compressor plant
second level Compressor station
third level Compressor unit
fourth level Compressor Driver
(Cooler)
M
5
Introduction - Network description -
 Compressor plants without crossings and circles
(inline).
Compressor plants with crossings and without
circles (tree)
Compressor plant with crossings and circles
(mesh)
6
Rule based decision support for the process flow
- Process flow of an optimisation -

• Introduction
• Process flow of an optimisation ?
• Knowledge based system ?
• Rule based system ?
• Rules for compressor plant configuration
• Pressure rules

7
Process flow of the optimisation- Overview -
8
Process flow of the optimisation- Loads -

• Inputs and off takes
• Valid for all runs
• Data sources
• SCADA System
• various planning files

9
Process flow of the optimisation- 1.
Pre-processing -

• Read process data from SCADA system
• Create a balanced load scenario
• Calculate flows at the Compressor plants
• Set pressure boundaries
• Set storage pressure
• Set flow dependant pressure boudaris

10
Process flow of the optimisation- 2.
Pre-processing -

• The results of the 1 pre-processing are used as
  input for the rule system
• The user can further reduce the resulting flow
  patterns for the compressor plants
• Maximum of 5 flow patterns per compressor palant

11
Process flow of the optimisation- Permutation -

• Permutation of the flow patterns for the
  compressor plants derived by the 2.
  Pre-processing
• All derived flow patterns of the compressor
  plants are independently combinable with each
  other
• It is not neglectable to reduce the number of
  flow patterns as much as possible
• 10 plants
• 5 flow patterns per station
• 510 different scenarios (N 9.765.625)
• runtime O(15N) ? 4,64 years (N 750 ? 3h7m30s)
• runtime O(1N) ? 113 days (N 750 ? 12m30s)

12
Process flow of the optimisation- configuration
set point optimisation -

• Send data via API to Simone
• Run configuration set point optimisation with all
  Scenarios of the permutation
• Standard machine type has to be configured
• Number of available machines has to be configured
• Mixed integer and discrete optimisation with
  SIMONE (CSO)

13
Process flow of the optimisation- 1. Post
processing-

• Read data via API from SIMONE
• Collect result data of the best results
• Resulting configuration of the compressor
  stations
• Set point
• Decision criteria for the selected runs
• Fuel gas consumption
• Necessary line pack shifting
• Create new variants by manual configuration
• Pre-selection of machine combinations with the
  estimated Power
• Select feasible combinations of aggregates

14
Process flow of the optimisation- set point
optimisation -

• Send data via API to SIMONE
• Set point optimisation with all variants
• SPO Module is used

15
Process flow of the optimisation- 2.
Pre-Processing -

• Read data via API from SIMONE
• Show best results of the scenarios (variants)
• Configuration of the compressor plants
• Set points

16
Rule based decision support for the process flow
- Rule based System -

• Introduction
• Process flow of an optimisation ?
• Knowledge based system ?
• Rule based system ?
• Rules for compressor plant configuration
• Pressure rules

17
Knowledge based system - handled data -

• The knowledge based system contains the database
• Grid export from Simone
• Grid topology
• Static data
• Scenario parameters and configuration
• Simulation results

18
Rule based decision support for the process flow
- Rule based System -

• Introduction
• Process flow of an optimisation ?
• Knowledge based system ?
• Rule based system ?
• Rules for compressor plant configuration
• Pressure rules

19
Rule based system - overview -

• Rule configuration to reduce the maximum number
  of possible flow patterns per Plant
• Set of rules for each compressor plant
• Dependency on the flow in the Branches of the
  compressor plants
• Declaration of pathes and direct connections
• Configuration of rules for pressure bounderies
• Dependency of flow on nodes
• Normal stations
• Bidirectional stations
• Storage pressure
• Formula for pressure boundary

20
Rule based system - Condition for flow pattern
(1. conditions) -
21
Rule based system - Condition for flow pattern
(2. flowpattern) -
22
Rule based system - Pressure rules -
23
END

• Thanks for your attention