Case Studies Multivariate Statistical Analysis Finds the Bad Actors in Light Component Losses

Edit This Case Study Record

	Multivariate Statistical Analysis Finds the Bad Actors in Light Component Losses

Multivariate Statistical Analysis Finds the Bad Actors in Light Component Losses

Technology Category	Analytics & Modeling - Predictive Analytics Analytics & Modeling - Process Analytics
Applicable Industries	Chemicals
Applicable Functions	Process Manufacturing
Use Cases	Predictive Maintenance Process Control & Optimization
Services	Data Science Services
Challenge	The petrochemical company was facing significant losses due to light component losses that go to the bottom of a fractionation column and pressurize the downstream column. This pressure resulted in downstream column production being lost to the flare. This stream contained a very valuable product, and the loss represented more than $1M USD. The company was looking for a solution to understand and resolve this production problem faster to limit the losses. Read More
About Customer	The customer is a large petrochemical company from Latin America, producing products like ethylene and propene, mostly from naphtha. These compounds are key products in the manufacturing of thermoplastic resins. The company has a successful history of innovation and advanced technologies use, including simulation for conceptual design, revamps and process monitoring, as well as several profitable advanced process control implementations and real-time optimization. Currently, the company is evaluating potential solutions to support operation decisions based on advanced analytics technologies. Among these solutions there is a specific initiative on process analytics to make profitable use of the rich historical data information for process knowledge improvement, troubleshooting, optimization and process monitoring. This effort could also complement their current use of technologies like simulation using Aspen Plus. Read More
Solution	The company launched an Aspen ProMV pilot project to investigate the light component losses. Using Aspen ProMV for continuous processes, a model was developed, and bad actors that are highly correlated to the light product loss were identified. Aspen ProMV’s optimization tool was also utilized to provide better operating conditions to reduce losses. The software was able to highlight a few process variables that are highly correlated with the light product losses and optimize the fractionation column operating conditions to limit the losses. The software was able to demonstrate that by changing the operation conditions, the column product losses can be reduced. Read More Log in to view content
Contents

Technology Category

Analytics & Modeling - Predictive Analytics

Analytics & Modeling - Process Analytics

Applicable Industries

Chemicals

Applicable Functions

Process Manufacturing

Use Cases

Predictive Maintenance

Process Control & Optimization

Services

Data Science Services

Challenge

The petrochemical company was facing significant losses due to light component losses that go to the bottom of a fractionation column and pressurize the downstream column. This pressure resulted in downstream column production being lost to the flare. This stream contained a very valuable product, and the loss represented more than $1M USD. The company was looking for a solution to understand and resolve this production problem faster to limit the losses.

About Customer

The customer is a large petrochemical company from Latin America, producing products like ethylene and propene, mostly from naphtha. These compounds are key products in the manufacturing of thermoplastic resins. The company has a successful history of innovation and advanced technologies use, including simulation for conceptual design, revamps and process monitoring, as well as several profitable advanced process control implementations and real-time optimization. Currently, the company is evaluating potential solutions to support operation decisions based on advanced analytics technologies. Among these solutions there is a specific initiative on process analytics to make profitable use of the rich historical data information for process knowledge improvement, troubleshooting, optimization and process monitoring. This effort could also complement their current use of technologies like simulation using Aspen Plus.

Solution

The company launched an Aspen ProMV pilot project to investigate the light component losses. Using Aspen ProMV for continuous processes, a model was developed, and bad actors that are highly correlated to the light product loss were identified. Aspen ProMV’s optimization tool was also utilized to provide better operating conditions to reduce losses. The software was able to highlight a few process variables that are highly correlated with the light product losses and optimize the fractionation column operating conditions to limit the losses. The software was able to demonstrate that by changing the operation conditions, the column product losses can be reduced.

Impact #1	Aspen ProMV was able to highlight the most important process variables to light product losses at the bottom of the fractionation column.
Impact #2	The software was able to demonstrate that by changing the operation conditions, the column product losses can be reduced.
Impact #3	Aspen ProMV enabled the users to explore a model that maps the operating space of the process represented in the historical data.

Impact #1

Aspen ProMV was able to highlight the most important process variables to light product losses at the bottom of the fractionation column.

Impact #2

The software was able to demonstrate that by changing the operation conditions, the column product losses can be reduced.

Impact #3

Aspen ProMV enabled the users to explore a model that maps the operating space of the process represented in the historical data.

Benefit #1	The reduction in light product losses at the bottom of the column was seen less than two days after the change.
Benefit #2	The value and variability in pressure of the downstream column was also reduced.

Benefit #1

The reduction in light product losses at the bottom of the column was seen less than two days after the change.

Benefit #2

The value and variability in pressure of the downstream column was also reduced.

Download PDF Version

Overview

Multivariate Statistical Analysis Finds the Bad Actors in Light Component Losses

Operational Impact

Quantitative Benefit