Energy Monitoring Dashboard for the Process Industries

1
Energy Monitoring Dashboard for the Process
Industries

• Honeywell Process Solutions
• Brendan Sheehan
• February 2009

2
Two Key Areas Utility Plant Process Plant
Raw materials
Process Plant
Utility Plant
Purchased fuels
Power, steam, water
Products
Power
By-product fuels, waste heat
Water
Energy losses
Energy losses
Energy losses
Equipment types

• Heaters
• Heat exchangers
• Pumps, compressors, fans, blowers, motors
• all process units

• Steam pipes
• Fuel gas pipes
• Transmission lines

• Boilers
• Turbines
• CHP
• Gasification

3
Energy Monitoring Dashboard Key Steps

• Define context for Energy Management
• Hierarchy defined by PRM (Plant Reference Model)
• Energy data to be monitored steam, fuel, power
• Calculate actual energy use
• From process data after applying data
  reconciliation step
• Define unit/equipment energy models
• Off-line process various modeling technologies
  can be used
• PLS, linear regression, local regression (DSS),
  etc.
• Set target energy use
• Using energy models or alternative techniques
  (experts, industry benchmarks, best performance
  achieved in the history, etc.)
• Monitor energy use
• Systematically compare actual to predicted
• Calculate economic () and environmental (tons of
  CO2) impact
• Continuous improvement
• Reporting, deviation analysis, recommended actions

4
Use Energy More Efficiently Optimize the Process
Create energy models Turn data into meaningful
information
Orient
Capture Energy Data Observe Current Situation
Observe
Decide
Determine possible actions based upon analysis.
Act
Take action based upon informed decisions.
5
Energy Management High-Level Scheme
Orient
Energy monitoring
Deviations (planned actual), (predicted
actual)
Planned specific energy consumption
Predicted specific energy consumption
Actual specific energy consumption
VisualizationAnalysisReporting
Decide
Predictive model
Calculating engine
Energy losses
Data reconciliation
Planned conditions
Actual conditions
Sampling actual data
Observe
Process Unit
Act
Energy OUT
Energy IN
Operating conditions
Recommended actions
Production Planning/ Scheduler
Operating Instructions
Production schedule
6
Architecture of EMD Demo
on-demand
Models of KEIs
recommended actions
KEI models
What-if scenario modeling
Models of KEIs
Target models
Calculation service
every 60 minutes
Trends
Deviation analysis
Modeling Environment
OM Reports
PHD tags
reasons
Tree Map
off-line workbench tools
Energy targets (predicted)
Operations Monitoring
Raw data
VisualPHD
Variance Monitoring
Reconciled data
NCDARR engine
MS Access
OM deviations
Energy targets (actual)
VisualPHD
Target types
Uniformance PHD
Reason types
Calculation service
PRM
OM Database
Data reconciliation
every 60 minutes
7
Context for Energy Monitoring Dashboard

• Hierarchical representation (tree) of plants,
  units, pieces of equipment included in the Energy
  Management
• Plant Reference Model (PRM) or its modification
  can be used for this purpose
• Advantage energy consumption can be aggregated
  at arbitrary level of the hierarchy

Fuel
Steam
70 - Preheat
70-Crude Unit
Power
70-Furnace
75-FCCU
Refinery
Fuel
70- Tower
Steam
76-Reformer
Chemicals
Power
Malibu
Alumina
Enterprise
Area
Plant
Unit or equipment
Energy levels
8
Energy Monitoring Demo
Developed in EAF collector. Key functional
features are organized on 8 tabs
Demo is focused on refinery CDU unit. The initial
screen shows its scheme
9
Observe - CDU Overview
Gauges illustrate the current status for each
type of energy that is being consumed on given
unit
Green zone is defined as a target /-5
Live data fields connected to PHD tags
Information about energy being used by various
sub-systEMD
10
Observe - Tree Map
11
Observe - Tree Map
Tree Map visualizes relative energy use in the
context of plant hierarchy (PRM). Rectangles
correspond to individual targets, and are further
grouped by equipment and plant.
Calculated metrics like total economic impact (in
) of deviating from the target are used to
characterize targets
Size of each rectangle corresponds to nominal
energy use, recalculated to the same basis
(MBTU/hr)
Color indicates how well or poorly is the given
energy target met
12
Observer - Tree Map Time Filter
Drill-down window shows details for the currently
selected (clicked) rectangle, incl. numerical
values for given metric
Drill-down window shows details for the currently
selected (clicked) rectangle, incl. numerical
values for given metric
Performance of individual targets (i.e. their
metrics) can be calculated for arbitrary time
interval
13
Observe - Tree Map Aggregated Metrics
All metrics can be aggregated into one number at
the equipment (unit) level. Consequently also the
number of displayed rectangles is reduced.
14
Observe - Tree Map Calculated Metrics
Four types of metrics can be visualized a)
Economic impact in b) Environmental impact in
tons of CO2c) Duration of all deviationsd)
Total number of deviations The metrics are
aggregated (SUM, AVG, MAX) over the selected time
interval
15
Observe - Tree Map Target Type Filter
Energy targets can be filtered according to the
type of energy fuel, steam, power, and cooling.
Consequently, only the reduced number of relevant
targets is displayed.
16
Observe - Tree Map Browsing through Time
It is possible to define a specific time
interval, e.g. 1 day, and then use the VCR
buttons to browse through the history day by day.
The metrics displayed in the Tree Map will be
re-calculated every time, showing how the
performance on individual units changed over time.
17
Orient - Calculation of Actual Energy

• Depends on type of energy
• Power/electricity directly metered
• Steam using steam enthalpy flow rate
• Fuel using calorific content flow rate
• Data Reconciliation
• NCDARR ( Non-Linear Constrained Data
  Reconciliation and Regression) engine can be used
  to
• Reconcile mass and volume flows
• Take advantage of sensor redundancy
• Calculate unmeasured flows
• Spot metering errors (gross errors)
• in general improve calculation of actual
  energy use

18
Orient - Data Reconciliation NCDARR
19
Orient - Data Reconciliation Browsing History
It is possible to pick-up any time and review the
reconciled values
20
Orient - Data Reconciliation Interactive Form
Configuration of any variable can be changed
using the interactive form
21
Orient Data Reconciliation Interactive Form
Any value can be entered manually and
reconciliation restarted with new settings
22
Orient Derive Unit/Equipment Energy Models

• First principle models
• Thermodynamics, mass and energy balances
• Existing product
• UniSim
• Statistical models
• Data-driven models created by fitting historical
  data
• Existing products and technologies
• ProfitSensorPro (PLS, linear regression, robust
  regression)
• DSS (local regression)

23
Orient - Energy Model for Crude Distillation Unit

• Correlates energy consumption on CDU unit (
  energy targets) with production targets that
  uniquely define the mode of operation

• Predicted energy consumption

Fuel energy (preheat, tower, total)Steam energy
(tower, side strippers, total)Power energy
(tower, pump-arounds, total)

• Feed

• Products

Saturated gas flow rate Naphtha flow
rate Kerosene flow rate Diesel flow rate AGO flow
rate Residue flow rate Naphtha distillation D86
90 Rec Kerosene distillation D86 10 Rec Diesel
distillation D86 90 Rec Diesel sulfur wt
Feed flow rateFeed temperatureAPI
densityAssay/blend properties
CDU

• Disturbances

Weather
24
Orient - Energy Target Models
Energy Target Models tab provides details about
the types and performance of individual energy
target models used
This model was created in profit Sensor Pro
25
Orient - Energy Target Models
Specific energy target can be selected from the
combo box
Histogram of all deviations
Scatter plot displays calculated model values
against measured energy values
Time series view on the measurement tag
corresponding target model
26
Decide Setting Target Energy Use

• Target energy use is determined by energy models
• Alternative techniques can be incorporated
• Solomon index, industrial benchmarks, experts,
  best historically achieved performance, etc.
• Two types of calculated targets
• Planning targets
• Updated 1-2 times per day as maximum - for new
  operating modes, changes to planned production
  rates, changes to operating conditions
• Dynamic targets
• Updated regularly hour by hour to reflect actual
  plant conditions
• Calculated targets are
• Written to PHD tags
• Systematically monitored by Operations Monitoring

27
Decide - Trends CDU Energy Targets
Three types of PHD tags (a) raw energy in yellow
color (b) planning target in red color (c)
dynamic target in blue color
28
Decide - Trends Production Flow Rates
29
Decide - Trends Data Reconciliation
Comparison of raw data versus reconciled
30
Decide - Monitoring Energy Use against Targets

• Compare actual to predicted use per
  unit/equipment
• Record deviations
• Adjustable threshold in
• Calculate
• Economic impact using marginal energy costs
  /MBTU
• Environmental impact using emission factors
  kgCO2/MBTU
• Show results to operators and engineers
• Have operator enter reasons and comments about
  inefficiencies each shift

31
Decide - Summary of Energy Targets in OM
32
Operations Monitoring Equipment Picker
It is possible to use the Equipment Picker dialog
to filter targets according to equipment/unit
Oil refinery was added as part of the EMD demo.
CDU unit has a history of two months (June/July
2008) of realistic data, all other units have
just few mock-up time series.
33
Operations Monitoring Filter
Only a subset of all OM targets can be visualized
like only energy targets in this case
34
Operations Monitoring Summary
The two months of history contain deviations on
various targets, some of them are annotated.
35
Operations Monitoring Deviations
Some of the historical deviations are provided
with reasons and verbal descriptions (notes).
E.g. here the deviations is caused by a general
problem with side strippers duties, and
specifically by high kerosene stripper duty.
36
Operations Monitoring Adding Reasons
There is a pre-defined list of reasons for CDU
unit
The users of OM can assign reasons and comments
to any deviation using the Analysis dialog.
37
Energy Reports
Energy reports can be calculated for an arbitrary
time interval
Minus deviations are indicated by blue color
they actually correspond to energy savings
Marginal energy costs (/MBTU and /MWh) are used
for calculation of the economic impact
Environmental impact is calculated using the
emission factors (kg of CO2 per MBTU or per MWh)
38
Energy Reports Time Filter
Energy cost is always summed up for the selected
time interval, like one week
39
Energy Reports Export to PDF
Any report can be exported to a PDF file
40
Energy Reports Export to PDF
Basic version of the PDF report
41
Decide - Deviation Analysis
Color gradient highlights economic impact of
deviations
List of all energy targets defined at the CDU unit
Each deviation is displayed as a bar with
associated specific color
42
Decide - Deviation Analysis Reason Codes
Pre-defined set of reason codes for the CDU unit
Alternatively, colors can be associated with
reason codes
43
Act - Perform Continuous Improvement

• Generating reports to highlight major problems
• deviations having the largest economic impact
• Analysis of deviations
• taking advantage of annotated deviations
• Tracking Key Energy Indicators (KEI)
• KEIs are process parameters, which have
  significant impact on the overall energy
  consumption
• Examples for CDU
• reflux ratio, coil outlet temperature, stripping
  steam rates, draw flow rates, pump-around flow
  rates and duties, etc.
• Aim to provide guidance on possible remedial
  action (TBD)
• Recommended actions

44
Act Links to other Applications

• Recommended actions may include
• Define new set of Operator Instructions
• Enter new set-points for APC application
• Run new production schedule
• (Above applications not currently included in EMD
  package)

45
Energy Monitoring Dashboard