Digital Twins for Power and Cement Plants

Optimizing Efficiency and Reliability

CII Conference on Energy Efficiency in Captive

& Waste Heat Recovery Power Plants
23rd May, 2024
Hyderabad
REALIZING A DIGITAL TWIN
3
Impact Stories

1 2 3
KILN - COMBUSTION BOILER TUBE
COOLER - OPTIMIZATION LEAK
WHRS
Impact Story
Kiln - Cooler - WHRS
WHAT To provide maximum Mid-Air heat to WHRS without compromising
recirculation of SA/TA temperature and quantity to the Kiln

WHY To maintain consistent process, Kiln Specific Heat, maximize

WHRP generation, lower NOX emission, stable clinker quality and
consistent Clinker Production

HOW Learn from your data using AI and provide real-time alerts for
corrective actions
 Process Variation
● Variation in cooler operation leading to variation in kiln operation
● Due to lack of correct guidance, operations of cooler varies from shift to shift.
● Cooler Auto mode PID loops have limitation and do not consider multivariate parameters to control
Grate Stroke/Speed. For example, increases in cooler fan flows that change under-grate pressures
with same bed thickness as opposed to just programmed under-grate compartment pressures.
● Optimizers do not account for delayed responses between Kiln and Cooler operations.

Variation in Kiln parameters due to unstable heat recirculation from cooler. Kiln feed stable at 660-670 TPH:
 Inconsistent WHRP Generation
Balance of heat to Kiln vs. Mid-air duct to WHRS is not maintained due to inconsistent and inadequate
heat exchange taking place to generate enough mid-air temperature and flow.

.
Effective optimization in process control ensures stable secondary air
temperature while minimizing coal consumption.
Root Cause: Control Inappropriate

➢ PID Controller Issues: Delays can cause the PID controller to fail to operate effectively.
➢ Operational Impact: These delays affect overall combustion efficiency and stability.
 Heat Recovery Wastage
● Wastage of heat due to poor heat exchange in cooler leading to higher overall specific heat and
coal consumption.
● Inefficient operation of cooler grate stroke resulting in increases of mid-air temps, lower SA/TA
temps, and high clinker temperatures. Excess mid-air temperatures will be sent as exhaust due
to boiler inlet heat specification.

.
Case Study: Impact of Cooler Grate Speed on Heat Exchange Efficiency

➢ This graph highlights the

impact of incorrect
increases in cooler grate
speed while kiln RPM and
feed remains constant.

➢ The improper heat

exchange results in
elevated clinker
temperatures and
reduced SA/TA
temperatures, and lower
mid-air temperature
leading to lower MW
generation.
AI Driven Optimization Algorithms
Linking Data Science and Cement Domain Expertise

● Predictive models using AI and ML - Neural network based Transformers,

Regressors, Decision Trees/ Forests etc.
● Multivariate analysis - models can accommodate hundreds of parameters
simultaneously
● Real-time predictions and recommendations
● Delayed relationships in process data is accounted for accurate predictions

Kiln - WHRS - Cooler Optimization Parameters

● Cooler: SA/TA temperatures, TA + Mid-air flow from DP (after installing pressure

sensors), Grate Stroke, Bed Thickness, Cooler fan flows + compartment pressures +
distribution ratios, Mid-air temperature + pressure, Cooler ID/ESP Fan RPM +
Pressure + Load, Clinker Temperature
● Pyro Section: Kiln I/L temperature + pressure + CO + O2, PC Outlet temperature + O2
+ CO, PC coal + Kiln coal TPH, PH Fans temperatures + pressures + RPM + Load, TAD
Damper position + Pressure, Burning Zone temperature, Kiln Load + RPM + Torque,
Hood temperature + pressure. Kiln Shell Temperature, Feed Rate etc.
 Real-time Corrective Actions
Recom. Controllables Actual Inputs WHRP Generation Historical Inputs

Cooler grate ESP Fan speed Cooler Total Kiln Total coal MW MW MW Kiln Feed Total coal
speed/stroke (%) Flow Feed actual Actual Historical Actual Prediction historical historical
(Strokes/s) (cub. nm) actual (TPH) (Current (Diff Cooler (Current actual actual
(TPH) timestamp) setpoints for Timestamp) (for same rec. (for same rec.
same Kiln Inputs) Contr. range) Contr. range)
(TPH) (TPH)

13.49→14.83 85.7→95.7 10737→11811 526.8 31.64 9.93 10.7 10.45 551.6 31.62
(14.63) (92.9) (11739)

9.0→12.50 54.9→64.9 11656→12822 526.7 29.6 10.32 11.4 11.25 515.4 31.18
(11.96) (61.9) (12417)

15.35→16.88 85.9→95.9 11673→12840 515.9 32.0 8.97 9.8 9.96 514.4 29.78
(16.1) (92.9) (12345)

12.15→ 13.36 85.7→95.7 10775→11853 526.5 31.05 9.88 10.75 10.96 521.5 31.69
(13.13) (92.9) (11860)

9.0→12.5 91.9→97.0 11134→12247 531.8 30.62 8.91 10.01 9.60 502.2 31.46
(13.0) (99.7) (11774)

14.18→15.59 91.9→97.0 10943→12037 535.4 31.09 9.3 9.8 10.24 509.4 29.63
(15.27) (92.9) (11845)

12.52→13.77 65.0→75.0 11825→13007 530.7 29.81 9.83 11.68 10.33 550.4 31.50
(13.4) (76.6) (12227)
Expected Improvement in WHRP Generation
Actual Megawatt
Predicted Megawatt
Impact Story
Combustion Optimization
Fuel Path
Air Path
Water Path
HP LP
Steam Path T
IPT
T
Gas Path
Process constraints B
D
Operator controllable parameters COND

IDF IDF S R
1 2 H H
CE
M P
A 1
EC M
FD O B
F LPH
2 M 1
C
M LPH
FD
F D 2
1 Furnace
M
APH E LPH
3
PA
M
F F
BF
2 P
2
HP HP
PA
F H2 H1
BF
1
P
1
For an end-user, the inputs
are simple.

Make a course correction,

Gain in performance
 Combustion optimization module progress
timeline

Dec 26 - 28 Jan 19 2024 Feb 07 Mar 05 Mar 06 - 11 Mar 27

2023

Site visit Coal bunker and Air flow Feedback for boiler 8 Modification in Implementation of
UBC tags received. recommendation received, updated Air parameters recommended air
(GCV tag pending) logic changed to flows selection to find flows in boiler 7
Excess air % recommendation. best week. and 8.
used in best week

18
 Combustion optimization result on boiler - 7

Heat Kcal
Coal PA SA TA
GCV MS Flow UBC consumptio PA Flow SA Flow TA Flow Fan APC
Flow Recomm Recomm Recomm
n

Before
(20/03 - 803.50 3186.33 155.24 1.68 686.80 82.37 — 66.59 — 37.71 — 1057.96
26/03)

After
(27/03 - 783.33 3257.17 156.40 2.36 679.50 85.54 86.20 70.54 69.27 33.51 32.89 1080.54
01/04)

1. All parameters mentioned on the above tables are weekly avg for that week
2. Improvement in heat rate after following recommendation - 7.3 Kcal/kg of steam
3. Coal Saving - 8.41 Tonnes/day
4. Increase in Fan APC - 22.58 kwh/day
5. With approximate costs, the net impact (coal reduction - apc increase) is positive
19
 The boiler is closer to ideal now than at
start of trials

Potential coal savings in BLR 7 in first 20 days of trial. When boiler runs optimally, daily report states so and no
recommendations are made
The potential savings have progressively reduced as
combustion recommendations are followed → boiler is
closer to ideal state in terms of fuel consumption.

From here, optimization module will keep an eye on any sub-optimal operations and immediately
notify operations accordingly to avoid fuel wastage. 20
 Combustion Mill APC Saving Report

COAL COAL COAL COAL COAL

COAL FEEDER FEEDER FEEDER FEEDER FEEDER
Total
Current Predicted Power Duration Change FEEDER B C D E F
time Savings
Power Power Saved of saving in mills A FEED FEEDRAT FEEDRAT FEEDRAT FEEDRAT FEEDRAT
(kWh)
RATE E FEED E FEED E FEED E FEED E FEED
BACK BACK BACK BACK BACK

20.755 --> 23.802 --> 23.947 --> 20.205 --> 14.222 --> 0.043 -->
2023-07-11 12:38:18 1111.509 992.945 118.564 469 5 --> 4 6487
25.099 26.033 26.147 25.614 0.067 0.015

15.621 --> 23.157 --> 22.501 --> 21.404 --> 16.598 --> 0.04 -->
2023-07-12 12:28:18 1097.082 974.127 122.955 477 5 --> 4 6842
23.713 25.294 25.456 24.772 0.069 0.016

23.426 --> 24.843 --> 0.002 --> 23.839 --> 21.694 --> 13.409 -->
2023-07-13 12:33:18 1164.638 1015.488 149.15 434 5 --> 4 7551
26.211 26.965 27.043 26.911 0.069 0.015

17.868 --> 24.116 --> 23.53 --> 21.788 --> 20.873 --> 0.034 -->
2023-07-14 12:12:18 1150.582 1021.023 129.559 514 5 --> 4 7769
26.573 27.257 27.188 27.109 0.067 0.015

16.424 --> 23.722 --> 23.659 --> 24.001 --> 23.732 --> 0.028 -->
2023-07-15 12:26:18 1167.253 1044.297 122.956 404 5 --> 4 5795
27.276 28.1 28.374 27.746 0.056 0.014

19.672 --> 21.219 --> 21.938 --> 19.234 --> 12.74 --> 0.035 -->
2023-07-16 13:48:18 1082.13 948.195 133.935 398 5 --> 4 6219
0.023 24.544 24.45 23.832 21.968 0.021
Impact Story
Boiler Tube Leak
 Timeline of events - advanced warnings with
increasing risk scores

Apr 25, 2024 Apr 30, 2024 May 01, 2024 May 02 May 03 2024
2:05 PM 8:40 PM 8:14 AM 5:00 pm 1:11 PM

LTSH & RH PCFB Very high ID Fan current Informed site Unit shutdown
Metal Temp Furnace draft high and had due to BTL
High discussion
Boiler Risk
score
Deviation Summary

In March, boiler operations are largely ok as April 15 onwards, more alarms generated for
indicated by the presence of very few alarms longer durations
Change in Thermal Profile
R D
i r
s o
e p

I I
n n

T T
e e
m m
p p
Between April 25 and April 30 - sudden flip of metal temperatures
● 18 sensor values increased by about 40°C
● 41 sensor values decreased by 75°C
● Maximum temperatures have also breached the 600 degC mark (not observed commonly before)
 High variations in furnace draft

The fluctuation in furnace pressure, ranging from +13 to -23, could indicates a Boiler Tube Leakage scenario,
as it deviates from the normal expected range of -1 to -8.
ID fan overloading indicates higher boiler
operations risk

ID fan current showing overloading from This period coincides with furnace draft
April 28th onwards variations
 How the same approach identified previous
leaks

23 Oct 2023 BTL Increase in deviations Constant deviations observed in APH, ID Fan, PA Fan, FD Fan,
● APH and ID Fan: 20th Oct Flame Intensity, Platen,Panel, RH Metal Temperature, Economiser,
● LTSH, Platen, Panel Metal Temp: 18th Oct Rh Desuperheater, Final Superheater since start of January.
● RH Metal Temp,FD fan,Flame intensity, Economiser :
16th Oct
● PA Fan: 13th Oct
How can we help you build value from an AI program?
Safety
Identify problem / use case
Maintenance

Process / Performance

Quality
Data availability + integration

Align team + stakeholders

Deploy + feedback

Iterate + course correct

Thank you!
Partner with us to solve your toughest operational
challenges!

www.exactspace.co
[email protected]