Transcript Sigmafine

Sigmafine:
Providing Reconciled Data to the Business
Tom Hosea
OSIsoft, Houston, TX
Confidential and Proprietary © Copyright OSIsoft, Inc. 2007
11
Production Management/ Loss Control
A Simple Problem,
Complicated by Reality
Confidential and Proprietary © Copyright OSIsoft, Inc. 2007
2
The Fog of Data


Typical large petrochemical complex or refinery can
be polling 100,000 points, once per minute.
This corresponds to about 200 Gbytes per year.
(It will be 200,000 points or more in five years)
– Data does not balance
– Used (Misused) by multiple groups
– Inconsistent and incompatible conclusions
– How do you find the data you need?
– How do you analyze it?
Confidential and Proprietary © Copyright OSIsoft, Inc. 2007
3
Typical Refinery

A refinery is consistently shows an average mass
balance of 97.5%, i.e. the products plus fuel
consumed plus known losses is only 97.5% of the
crude plus intermediates purchased
– Is the refinery paying for crude not received?
– Is the refinery not being paid for all products?
– Is there theft of product or leakage or evaporation?
– Is more fuel being burned than estimated? Flared?
– Is there excessive off spec product being recycled?

2.5% losses on a 200,000 BPD refinery are worth
$300,000 per DAY or $110 million per year.
Confidential and Proprietary © Copyright OSIsoft, Inc. 2007
4
What is Data Reconciliation?

A statistical method of resolving detected
errors according to pre-specified rules and
tolerances

Distributes errors across a system

Reports on and explains errors

Includes:
– Data Validation
– Systematic Detection of Gross Errors (e.g. missing
measurements, mis-specified movement, etc)
Confidential and Proprietary © Copyright OSIsoft, Inc. 2007
5
The Issues with Data Validation

Too much data
– Thousands of data points

Too many sources
– Lab systems, DCS, manual entry

Too many interactions
– Transfers, flows, measurements

Not enough time…
Confidential and Proprietary © Copyright OSIsoft, Inc. 2007
6
Sigmafine
A product that
enables data reconciliation
and validation
for any industrial process.
Confidential and Proprietary © Copyright OSIsoft, Inc. 2007
7
Typical Scenario Without Validation

Some sort of local balance

Some arbitrary and subjective corrections

No agreement on data

Difficult to detect measurement errors
Fog
Data
Information
Confidential and Proprietary © Copyright OSIsoft, Inc. 2007
?
8
Validation with Sigmafine

A unique balance, valid for the whole
operation

Systematic and objective corrections

Agreement on balanced data

Easier to detect measurement problems
Data
Information
Confidential and Proprietary © Copyright OSIsoft, Inc. 2007

9
Data Reconciliation Challenges
Refining
Many Products
X
Lineup Changes
X
Flows & Transfers
X
Metals & Mining
Chemicals
X
X
Model Size (elements)
5000
1000
1000
Model Complexity
High
High
Low
Redundancy
High
Low
Low
Analyzer counts
Low
High
Medium
Unmeasurable Feedstocks
X
Material Acct per Element
X
Componenet Balances
X
Stoichiometric Balance
X
Confidential and Proprietary © Copyright OSIsoft, Inc. 2007
10
How to solve these problems
Use Sigmafine to…
1.
Build and configure a model (once)
2.
Run the model using the appropriate
analysis rules (frequently)
3.
Analyze results (frequently)
Confidential and Proprietary © Copyright OSIsoft, Inc. 2007
11
Sigmafine Tools

Data References
– A component that reads, writes and executes
calculations

Analysis Rules
– Provides model analysis for balances, composition
tracking or gross error detection

Data Loader
– Imports data elements of many formats to create
cases or transfers

Visualization/Analysis tools
– ProcessBook, AF Excel Add-in, RtReports
Confidential and Proprietary © Copyright OSIsoft, Inc. 2007
12
Benefits by Industry

Refining
–
–
–
–

Chemical
–
–
–
–

Transfers provide the basis of model receipts, shipments and movements
Automatic Inventory calculations
Composition Tracking of stored products
Refining specific calculations – gross to net conversion
Mass and Component balance
Configurable reaction constraints
Meter Compensation – gas and liquid
Inventory calculations
Metals and Mining
–
–
–
–
Component balance of materials not typically measured
Independent solvability of components
Independent accuracies of measurements
Efficient system management of sparse measurements
Confidential and Proprietary © Copyright OSIsoft, Inc. 2007
13
Conclusions

Sigmafine can be applied to any industry

Validated data is available to make better
business decisions

No process model is required to derive value
from Sigmafine

The use of data references does not require a
model
Confidential and Proprietary © Copyright OSIsoft, Inc. 2007
14
Good Data for Good Business Decisions
“You can't manage what you can't control, and
you can't control what you don't measure.”
Tom DeMarco
Sigmafine increases confidence in what you
measure and provides estimates of what you
don’t measure, helping you to make better
business decisions
Confidential and Proprietary © Copyright OSIsoft, Inc. 2007
15
Thank You
Questions?
Confidential and Proprietary © Copyright OSIsoft, Inc. 2007
16
16