Saudi Polyolefins Co. (TASNEE Petrochemicals) operates two world-scale polypropylene (PP) units in Jubail, Saudi Arabia. The units include polymerization and extrusion sections. The polymerization section for each line has two reactors operating in parallel mode for homopolymer production. Hydrogen is added continuously in a fixed ratio to the propylene feed to control the PPs molecular weight. The molecular weight is expressed as the melt flowrate (MFR). Each reactor train has its own degassing section, intermediate product storage and extruder. The total capacity of the plant is 720,000 tpy.
In early 2009, TASNEE decided to deploy an advanced process control (APC) solution for its recently debottlenecked PP plant in Jubail. Selected APC vendors were supplied with an extensive set of plant operating data, allowing them to perform a preliminary benefits analysis. The project scope included a reactor composition controller and virtual online analyzer for MFR and xylene solubility (XS). Quality control information and automated transition records for the four reactors were also included.
To estimate the potential benefits related to implementing an APC solution, a detailed benefits study was carried out in 2009 for the two PP lines. The benefits study methodology was agreed to by both TASNEE and the APC vendor. The benefits estimates were based on actual process data and the APC vendors experience with similar projects. The benefits were seen in three key areas: Measurable improvements in process stability; the ability to accurately predict key process variables (such as MFR and XS), which can be difficult to measure online; and improvements in process operation consistency.
Another key benefit discovered is that APC will push the process closer to its constraints without violating them, thereby increasing the plant production rate. This benefits study focused on three primary benefit areas; including increased production, reduced off-spec product and lower operating costs.
Results of the initial benefits study showed that less than six months return on investment (ROI) was achievable. In other words, production increased by over 2%, while the process capability index (Cpk ) soared by over 50%.
The TASNEE project team invested significant time and effort for vendor and technology selection. Factors such as technology capability, polymer experience, commitment to success, and cost were considered in the selection process. The vendor selection process took approximately six months. Each vendor was given the opportunity to present its benefits analysis and its technical solution. The TASNEE team paid special attention to the APC solution that provided remedies for chronic operational issues.
The project kickoff meeting was held shortly after the contract was finalized and initial preparations were completed. The initial preparation work involved things like installing the APC server, interfacing it with the distributed control system (DCS) and then making the subsequent DCS configuration changes.
TASNEE assigned a project team made up of operations experts, process engineers, process control gurus, DCS experts and IT support. The project team was overseen by a dedicated project manager who also acted as the technical lead. The project team was involved in all aspects of the implementation. Within the first three months, the majority of the APC solution was in place and commissioned. Improvements in process stability and product quality were noted shortly after commissioning began.
To improve reactor stability, the reactor pressure controller was commissioned, along with the polymer production rate controller. For the next four months, the project team focused on fine tuning the controller and installing and testing of the automated transitions. The implementation was completed seven months after the kickoff meeting. The performance test and audit took an additional three months. The project exceeded all performance criteria outlined in the project objectives.
This polymer APC project had to overcome some difficult issues related to the process while remaining compatible with overall objectives. One issue was that the process makes multiple products with varying catalyst activities (due to fluctuating hydrogen concentration) in the reaction loop. This introduced a high degree of non-linearity in the gains and the dynamics.
Another issue was that some of the controlled variables were open loop unstable. This required that the controller respond in a quick and sufficient manner so as not to introduce further instabilities. The team had to be aware that the process was sensitive to impurities in the feedstock and to those generated internally. The impurities tended to affect the catalyst activity and thus the response of the process. Further, the APC solution had to be programmed in a way to recognize that the steady-state location of process variables for a product will vary from run to run.
Another red flag to consider was that some of the final product properties that are important to the end user are not characterized by the parameters measured by the lab or controlled in the process. Clear lines of responsibility also had to be established, so that appropriate technology transfer to the operations and technical groups for maintenance and support could be undertaken. Fortunately, the analyzer performance issues and lab result variability were flagged early in the project and the project team was able to offer solid technical solutions.
The main project objectives were to increase production rate, improve process stability (reduce off-spec) and improve product quality or Cpk. APC solutions for polymers achieve these objectives by implementing a nonlinear multivariable control solution (Fig. 1). An important aspect of reactor stability is the ability to control critical process parameters such as reactor pressure and reactor temperature. Fig. 2 illustrates the pressure control improvement for one of the polymer lines.
| Fig. 1. Typical process flow diagram with |
APC implementation. Source: CB&I1
| Fig. 2. Reactors pressure control before |
and after APC implementation.
Table 1 and Table 2 show the improvements in process stability relative to the baseline during the three month performance period. Table 3 illustrates the projects performance guarantee objectives and Table 4 provides the actual results for the three month performance period.
In addition to reactor stability, improvement in the product quality was an important objective. The product quality improvement was measured by the Cpk number. This is a statistical measure of the plants capability to operate between product limits and near the designated product target. For this undertaking, Cpk for MFR was a project objective. In the end, APC implementation resulted in all products exceeding the Cpk objective.
Improvements in other product properties, including XS, were also observed. Fig. 3 shows the quality improvement for one of the product grades.
| Fig. 3. Quality control performance before and |
after APC implementation.
The APC solution not only improved process stability, increased production and offered better quality control, it also improved product transitions by automating the changes. The product transitions are now faster and more consistent. Transition times were reduced from 50% to 80% depending on the direction of the transitions and the magnitude of the change. This translates to making less off-spec material due to product transitions. A typical MFR transition is illustrated in Fig. 4.
| Fig. 4. Product transitions before and |
after APC implementation.
From then to now
The project was completed seven months after the formal kickoff meeting. The total benefits achieved shows that the ROI was less than three months. A few important factors that contributed to the projects success included the initial work on the project base line, scope definition and preparation. The essential elements are:
- Close collaboration between the TASNEE team and the APC vendor team
- Ongoing training
- Technology transfer to the TASNEE team
- Effective project management.
The application performance was monitored during the evaluation period (three months after commissioning) and the post-audit report shows that all performance criteria were exceeded.
The installed APC solution has been online for over 24 months, with an average utilization of over 93%. The nonlinear model predictive control solution is maintained by TASNEE personnel so that the expected results will continue to be delivered. Before implementation of this solution, TASNEE was focused on improving its lab quality control performance. Implementing this APC solution has allowed TASNEE to reduce the number of lab samples required. This has reduced the lab workload, resulting in more consistent results and improved product quality control. HP
1 CB&I, Parallel configuration: Homopolymers and random copolymers. Available at: www.cbi.com/images/uploads/Config_Parallel_New_3_788x450.gif.
Hanif Poorkar is a senior process engineer for TASNEE in Jubail, Saudi Arabia. He has been with TASNEE for eight years. Mr. Poorkar has worked with different technologies associated with PP processes throughout his 16 year career. He has a petrochemical engineering degree from Dr. Babasaheb Ambedkar Technological University in Lonere, India, and an MBA degree from Karnataka University in Dharwad, India.
Mohammed Al-Zahrani is the APC lead engineer for TASNEE in Jubail, Saudi Arabia. He has been with TASNEE for three years. Mr. Al-Zahrani has worked with different petrochemical polymer process technologies throughout his nine year career. He has a bachelor degree in chemical engineering. Mr. Al-Zahrani has two main priorities: assisting the process engineering polymer department and leading APC projects in the polyethylene (PE) plants within the TASNEE complex.
Hamad Al-Shbrain is the polymer process engineering manager for TASNEE in Jubail, Saudi Arabia. His responsibilities include activities within TASNEEs low density polyethylene (LDPE), high density polyethylene (HDPE) and PP plants. He is a chemical engineering graduate with more than 16 years of experience.
Kami Sefidrou is director of operations for Rockwell Automation in Houston, Texas. He has over 35 years of experience implementing APC solutions. Prior to joining Rockwell, he managed Aspen Technologys polymer group. His experience also includes process control positions with BAPCO in Bahrain, NOVA Chemicals and Syncrude Canada. He has a masters degree in process control and a bachelors degree in chemical engineering , both from the University of Wales, UK.
Brian Lines is a senior technical consultant at Applied Manufacturing Technologies (AMT) in Houston, Texas. Brian has over 34 years of experience in process operations, process design and APC. Mr. Lines has held positions at Pavilion Technologies, Aspentech, NOVA Chemicals, DuPont Canada and Syncrude Canada. Brian holds a number of US patents in non-linear model predictive control. Brian has a chemical engineering degree from the University of Waterloo in Canada.