Materials Use Reduction via Advanced Process Control
A graphic depiction of how Honeywell Profit Controller and Profit Optimizer were implemented in Nirma’s overall control structure to achieve plantwide global dynamic real-time optimization.
281x96x2 = the number of controlled variables, manupulated variables, and disturbance variables, respectively
Pre-frac = a tower that flashes off the lightest materials to prevent flooding in the downstream fractionation tower(s)
UF = union fining, UOP technology; Molex, Pacol, PEP, and Detal also refer to UOP processes
FE = front end
BE = back end
Par column = paraffin column
Based in India, Nirma manufactures detergents, bath soaps, salt, industrial products, and fertilizer. Its products have a retail reach of more than two million retail outlets and more than 300 million loyal consumers throughout India. The company started as a one-man operation in 1969, and today employs about 15,000 people.
Nirma’s linear alkyl benzene (LAB) complex is considered one of the most efficiently run LAB plants in India. The LAB production process starts with the processing of straight run kerosene fraction. During processing, normal paraffin is separated from other kerosene fraction.
The normal paraffin separation process is called the front end, in which the normal paraffin is olefinated and eventually reacted with benzene, which results predominantly in LAB. It also produces some heavy alkyl benzene (HAB).
The LAB production unit is called the back end. For providing heat to the distillation column, reboilers, and other heaters, hot oil is used in both front-end and back-end processing. Hot oil return is sent to a furnace for maintaining its supply temperature.
Nirma’s main goals for its LAB plant were to improve energy efficiency and to optimize normal paraffin recovery from kerosene. Because of the high level of interaction between the variables and large amount of dead time and settling time in its processes, Nirma strove to overcome inconsistencies and waste to provide dependable control.
On the control front, there was difficulty in maintaining paraffin molecular weight and maximizing n-paraffin recovery from feed kerosene simultaneously. Paraffin molecular weight control was a challenge, as the control handles and property measurement were in different sections of the plant (in addition to the previously mentioned issues of control dynamics with lengthy dead time and settling time). Other front-end challenges included return kerosene flash point control and hydro-treater reactor pressure control, which needed regular adjustments. In the back end, compound ratio control was the primary concern for operation.
On the energy front, Nirma looked to minimize the hot oil consumption in the distillation columns and attain power savings in hot oil circulation pumps.
The company chose Honeywell to help achieve these goals based on earlier success with the company’s distributed control system. This project also involved Honeywell’s UOP division, a part of Honeywell’s Specialty Materials business. UOP provided engineering services to design and implement advanced process control on Nirma’s LAB plant.
To manage N-paraffin molecular weight control, Honeywell’s inferential property prediction tool, Sensor Pro, was used to develop inferentials for predicting molecular weight. To attain control in the individual columns and gain global level optimization, Honeywell’s distributed quadratic program tool, Profit Optimizer, was used. This tool is used to control the return kerosene flash point.
Feed suppliers require the return kerosene flash point to be above a specified value. Since Nirma’s return kerosene streams originate from four different places, there is difficulty in controlling the flash point.
To control it from a multivariable process controller (MVPC), all four columns (source for return kerosene) must be connected to one MVPC. Since these columns are positioned from feed entry point to product out point, they have vast differences in dynamics and dead time, which leads to poor manipulative variable move planning and thus poor control. Profit Optimizer allows the streams to reside in a different controller and unites them with bridge models and source-clone connection for good control on the final flash point. Since the program looks and directs all four controllers, it can also move the units towards optimal operation.
The back end offered more challenges owing to the non-linear nature of the process. Predicting the kinetic properties like conversion, selectivity, and recycled gas composition was a challenging task. Inferentials were engineered to predict the catalyst selectivity, conversion, and recycled gas hydrogen composition. These inferential inputs were linearized based on first principle equations before being given to inferential property prediction.
A total of 11 Profit controllers, Honeywell’s advanced process control technology, were implemented along with Profit Optimizer. Five Profit controllers each were implemented in the front end and back end, and one for the complete hot oil system. Profit Optimizer, which provides the interconnection between the front-end and back-end processing, was implemented to achieve the plant-wide real-time dynamic optimization.
Overall, the project included implementation of the Honeywell advanced process control system, which included the Profit Controller, Optimizer, and SensorPro, as well UOP technology. UOP enabled the production of normal paraffins and linear alkyl benzene via a supply of catalysts and adsorbents to achieve desired product and product quality objectives.
Initial benefits from the project include:
N-paraffin recovery improvement from feed kerosene of 1.031%;
Reduction in fuel consumption, including front-end fuel savings of 0.7% and back-end fuel savings of 1.48%;
Because the Profit controllers adjust the column parameters every minute to meet the product specification, energy consumption reduction includes a front-end power savings of 1.1% and back-end savings of 1.1%;
By sustaining maximum production and consistent control of plant parameters, overall standard deviation has been reduced by more than 30%;
Hydrogen savings have been in excess of 31%; and
Annual monetary savings of more than $750,000 USD have been realized.
|Dilip Kannan is engineering manager for Honeywell Automation India Ltd. More advanced process control information can be found at|