Reducing Costs and CO2 Emissions by BTG Optimizing Control Service

Introduction

This document discusses the implementation of Yokogawa's Advanced Process Control (APC) system at the Takasago plant of Kaneka Corporation to optimize its Boiler, Steam Turbine, and Generator (BTG) facilities. These facilities are critical for supplying power and steam to the large-scale production plant, and their CO2 emissions are a significant concern. The adoption of APC, combined with Yokogawa's Exapilot operation efficiency improvement package, has successfully automated manual operations, leading to reduced fuel costs, lower CO2 emissions, and improved labor efficiency.

BTG facilities, particularly coal-fired ones, face challenges due to rising coal import prices and environmental regulations. Optimizing plant operation is crucial for managing fuel costs and CO2 emissions. BTG facilities supply electricity and steam, requiring control to meet fluctuating production demands. Unlike power company facilities, BTG systems often have complex steam pipelines, fuel-dependent boiler dynamics, and various types of steam turbines (condensing, back-pressure, extraction), placing a heavy workload on operators to adjust operations for changing demands and production processes.

This paper details a solution implemented at Kaneka Corporation using Yokogawa's Distributed Control System (DCS), APC, and Exapilot procedural automation package.

Cost Reduction and CO2 Minimization Activities at Kaneka Corporation

Background

The Takasago plant initiated a cost reduction program called ADVANCE17, focusing on energy saving and CO2 emission reduction as part of its environmental initiatives. The BTG facilities, which consume a substantial amount of energy (fuel and purchased electricity), were a primary focus due to increasing crude oil prices.

Introduction of the BTG Optimizing Control Service

Kaneka Corporation's Takasago plant aimed to optimize BTG operations by applying Exapilot and APC to its existing DCS. APC was enhanced with a model adapting function to simulate various operation patterns of boilers and turbines, ensuring more stable operation.

Yokogawa's APC

Despite previous energy-saving efforts and hardware modifications, the Takasago plant identified further opportunities with Yokogawa's APC. While initial analysis suggested limited gains from automating manual operations, the plant proceeded with introducing APC in collaboration with Yokogawa, recognizing the urgent need for cost reduction and environmental management improvements.

Parameter Fitting by Exapilot Operation Efficiency Improvement Package

To address the potential increase in operator workload due to APC's requirement for parameter adjustments with changing operation patterns, Exapilot was used to automate manual operations like changing model parameters, PID loop configurations, and control parameters. APC and Exapilot were designed to work together, sharing DCS memory for synchronization.

Figure 1: System Configuration illustrates the system configuration for Yokogawa's APC and Exapilot. It shows inputs like city gas and coal feeding into a boiler, which produces steam for turbines (G, GD) generating electricity. Commercial power is also an input. The system includes a DCS, Exapilot for switching operation patterns, and APC for optimization, managing stable operation and the optimizing engine.

Yokogawa's Advanced Process Control (APC)

Yokogawa's APC products, including Exasmoc, are designed for advanced process control. Exasmoc is a core product that uses multivariable model predictive control. It simulates target process dynamics with multiple inputs and outputs and controls the process by predicting its behavior steps ahead.

Figure 2: Example of the configuration of Yokogawa's APC products displays a configuration example of Yokogawa's APC products. It outlines various components such as optimization models, control models, estimation models, process data, setpoints, control logic, and their interaction with the DCS and OPC server.

Figure 3: Conceptual diagram of a dynamic model provides a conceptual diagram of a dynamic model used in Exasmoc. It depicts multiple manipulated variables (MVs) influencing multiple control variables (CVs) through process dynamics, along with unmeasured disturbances (DV). The model predicts future process behavior.

Outline of BTG Facilities and Purpose of APC

BTG facilities comprise boilers with various fuels and turbines, often supplemented by purchased electricity. The primary goal of optimizing these facilities is to minimize overall utility costs and CO2 emissions. This is achieved by strategically managing purchased power, buying more during off-peak hours (nighttime, holidays) and less during peak hours (daytime), thereby minimizing the gap to operational constraints.

Introducing APC

The implementation of APC at the Takasago plant took approximately one year, with the APC introduction itself taking about six months. The remaining time was dedicated to plant testing (response characteristics) and commissioning (performance verification) across various operation patterns.

Steps for Introducing APC

Feasibility Study

A feasibility study was conducted to estimate the benefits of APC, focusing on:

• Minimizing the boiler load during nighttime, analyzing operations and defining constraints.
• Maximizing the boiler load during daytime, analyzing operations and defining constraints.

Basic Design

Following project scoping, the basic design for APC implementation was developed. This included defining:

• Configuration of controllers and control periods.
• Control variables (CVs).
• Manipulated variables (MVs).

Several operation patterns were selected for control loop design, including switching pressure control configurations between day and night, and managing the number of on-stream boilers. APC for BTG facilities typically assumes fixed process configurations to avoid overly complex models, excluding start-stop sequences or major control loop switches. However, it can handle multiple operation patterns, including varying boiler numbers and pressure control loops, for enhanced energy saving.

Retuning of PID Controllers

Before APC implementation, PID control parameters in the DCS were retuned. This step ensured that input signals to APC showed no significant noise or fluctuation, and that the targeted PID controllers responded accurately to setpoint changes. Quantitative analysis confirmed a reduction in fluctuations after retuning.

Modifications to existing control loop configurations were also proposed to ensure better controllability at the regulatory control level.

Plant Test

Plant tests were conducted to build process dynamic models for APC. Many BTG facility components exhibit quick responses and stable mass balancing, allowing for accurate model configuration. Tests focused on slower responses (e.g., temperature) and heat balance confirmation. Repeated tests were crucial for identifying dynamic parameter differences across various operation patterns. A universal model was developed to accommodate different patterns with only parameter adjustments, avoiding model reconfiguration.

Figure 5: Example of a control model presents an example of a control model, likely used for fitting process dynamic models. It shows inputs such as 'From Plant,' 'Measured Disturbance Responses,' and 'Unmeasured Disturbance Responses,' and outputs related to 'Start,' 'Scenario Setting,' and 'From Plant'.

Detailed Design: Simulation Model Fitting

Control models were fitted using plant test data. This allowed for the review of process responses under APC control on a PC before online implementation, shortening the commissioning period.

Commissioning

During commissioning, the APC system was tested in the actual BTG facilities across various operation patterns, confirming the expected benefits. Beyond the economic advantages, the implementation resulted in an 85% reduction in manual DCS operations by operators and a decrease in process alarms.

Effects of APC Introduction to BTG Plant

An example of APC's impact is the minimization of boiler load during nighttime. The operation is subject to constraints:

• Differential pressure of the boiler furnace chamber (Lower load limit).
• Coil outlet temperature of the boiler (Upper load limit).

Figure 6: Constraints for APC illustrates these relationships. Figure 7: Example of the effect of APC introduction (Trend data) shows that after APC implementation, the boiler operation at minimum load achieved closer control to the lower differential pressure limit, while remaining well below the upper coil outlet temperature limit.

Conclusion

This report summarized the optimization of BTG operations at Kaneka Corporation's Takasago plant using Yokogawa's Exasmoc multivariable model predictive control package and Exapilot operation efficiency improvement package. The project led to a significant reduction in overall energy consumption, with an equivalent CO2 emission reduction exceeding 1000 tons per year. The integrated control system (DCS, Exapilot, Exasmoc) also reduced operator workload and process alarms. These achievements were attributed to the plant staff's dedication and Yokogawa's expert guidance. Future work will focus on further improving controllability and minimizing boiler load change times.

References

• (1) Hiroshi Wakasugi, Kazuo Sueyoshi, "Energy-saving Solutions by Advanced Process Control (APC) Technology," Yokogawa Technical Report English Edition, Vol. 53, No. 1, 2010, pp. 11-14
• (2) Yoshinori Urasawa, Yukio Innami, Masao Mizuuchi, "Energy-saving Solution by BTG Optimal Control," Yokogawa Technical Report English Edition, Vol. 53, No. 1, 2010, pp. 15-18
• (3) Yasunori Kobayashi, “Exapilot, Assistant Package for Improvement of Operational Efficiency," Yokogawa Technical Report, Vol. 43, No. 3, 1999, pp. 17-20 (in Japanese)

VigilantPlant Services, Exasmoc, Exarqe, Exaspot, Exacoast, CENTUM, Exaopc, and Exapilot are registered trademarks of Yokogawa Electric Corporation.