OR WAIT null SECS
Anurag S. Rathore is a professor in the Department of Chemical Engineering at the Indian Institute of Technology Delhi and a member of BioPharm International's Editorial Advisory Board, Tel. +91.9650770650, email@example.com.
An approach to reduce batch time, increase productivity, and decrease costs.
Process intensification is the new catchphrase in biotech processing. It includes concepts such as reducing the process time of a particular unit operation, decreasing the overall process time by improved alignment of various unit operations, and reducing plant idle time by improved equipment utilization. Biotech manufacturers are realizing that cost savings from such efforts, which result in improved productivity for a facility, far exceed those from the traditional focus area of raw materials. This article presents an approach for performing process scheduling and debottlenecking.
The combination of rising costs for discovery, development, and commercialization of biotech drugs, and ever increasing pressure on product pricing from the government and insurance agencies has put renewed focus on containing manufacturing costs in the biotech industry. Process intensification is emerging as one way to achieve this desired objective and includes concepts such as reducing the process time of a particular unit operation, decreasing the overall process time by improved alignment of various unit operations, and reducing idle plant time by improved equipment utilization. Individual unit operations (or process stages) and their relationships within the overall process are identified, including an assessment of resource and utility requirements from the host facility. After process models have been constructed, we can investigate complex and integrated biochemical processes and unit operations without the need for extensive experimentation or disruption to existing operations. The method can be used at all stages of process development from conceptual design to process scheduling and optimization.1
The above-mentioned activities require process scheduling and debottlenecking using platforms such as Microsoft Excel or commercially available software.2–6 Some of the most widely used commercial software include, Aspen Technology's Aspen Batch Plus, Hyprotech Ltd's Batch Design Kit, and Intelligen's SuperPro and SchedulePro. These tools can optimize a manufacturing facility's capacity by scheduling the process to decrease batch cycle time and reduce the amount of equipment used, which will ultimately decrease pr2oduction costs.
Anurag S. Rathore
Several questions arise when performing process scheduling and debottlenecking. How much would it cost to make a fixed amount of the product? How can I reduce the operational cost of the process? How much product can I make in this plant in a fixed time? What are the bottlenecks in my current process and plant? What is the relationship between capital investment and the resulting capacity expansion? How can I relate the scheduling of a process to changes in priorities or production demand?
In this twenty-first article in the Elements of Biopharmaceutical Production series, we introduce some of the key concepts of process scheduling and debottlenecking, and present an approach for performing these activities. Two case studies—one for a single- product biotech facility and another for a multiproduct biotech facility—illustrate our approach.
Process scheduling for a production facility is performed to minimize production time and costs by making decisions about what to make, when to make, with which staff, and on which equipment. The overall aim is to maximize operational efficiency and reduce costs. Commercially available scheduling tools provide the production scheduler with powerful graphical interfaces that can be used to visually optimize real-time workloads in various stages of production, while pattern recognition enables the software to automatically create scheduling opportunities, which may not be otherwise apparent without this analysis.5–6 Manual scheduling, using available platforms such as Microsoft Excel, also can be effective because the user has the advantage of customizing the tool as required, compared with commercial software that are more rigid. Manually developed approaches, however, may lack the user-friendliness of commercial software.
Process scheduling can be conducted by backward or forward scheduling methods and can help allocate plants, machinery, and human resources, plan production processes, and purchase raw materials. Forward scheduling involves planning tasks from the date resources become available to determine the shipping date or due date. Backward scheduling is planning the tasks from the due date or required-by date to determine the start date or any required changes in capacity. The benefits of scheduling include a decrease in process changeovers, reduction in inventory, increase in production efficiency, leveling of labor load, accurate delivery of milestones, and ability to gather real-time information.
Process debottlenecking generally includes identifying and removing bottlenecks in equipment and resources. Equipment-based bottlenecks can be eased or eliminated by adding or removing equipment. Resource-based bottlenecks, however, may be unavoidable and may limit the net productivity of the plant. Feed throughput for a production facility is directly proportional to the batch size and inversely proportional to the cycle time; thus, for constant batch sizes we can reduce the cycle time by extensive scheduling, which will subsequently increase throughput. To increase a batch size, we will need to increase the process efficiency or perform scale-up of the process.3
Economic evaluation is necessary when deciding between in-house production and outsourcing. Building a new production facility to manufacture a biotech product is not only a major capital expenditure, but also a lengthy process. To make this decision, information regarding the required capital investment and time to complete the facility is necessary. Even if the decision is made to outsource production, the above-mentioned cost analysis will still be useful as a basis for negotiating with contract manufacturers.4 Irrespective of the decision on where to manufacture, further scheduling and optimization is always useful for maximizing profitability. In a recent publication, the authors illustrated how process scheduling and optimization for a monoclonal antibody product made in a multiproduct biotech facility was used for performing material and energy balances, sizing equipment and utilities, estimating capital and operating costs, and analyzing cycle times.1
This section proposes a simple approach that could be used for process scheduling and optimization. The approach is manual and requires common platforms, such as Microsoft Excel. As mentioned earlier, however, commercially available software also can perform this task. The key inputs for this analysis are the process flow diagram; information on time, labor, buffer, and media (if any) required for each step, and equipment specifications (for existing process).
Step 1 focuses on evaluating process economics. First, we calculate the net fix, expenditure for each step for a single batch, including recurring cost and non-recurring cost, (eq. 1),
in which NE(i) is net expenditure for each step per batch; t(i) is batch time (h), W is average labor wage (per day); Nl is number of laborers required for operation; T is shift time (h); ReC(i) is other resource costs (media, WFI, etc.); C(i) is equipment cost; Nb is number of batches per year; and Te (i) is equipment life (in years).
Next, we calculate the production profit per batch (PP) (eq. 2),
in which SP is the selling price of the product.
Assuming that the amount of raw material used is proportional to the production capacity, a new parameter E can be defined for cost analysis (eq. 3),
in which, E is constant and does not depend on the number of batches produced per year (Nb).
The cost index (n) can then be defined in (eq. 4),
in which, Nb (i) is the changed number of batches per year because of the addition of new equipment (i = 2,3,4...); and Nb (1) is the number of batches with each single piece of equipment.
The cost index (n) is greater than 1 if the change is profitable. Therefore, scheduling should be continued while n is increasing. This strategy ensures that any change made improves overall profitability.4
Step 2 involves creating an equipment- occupancy (EO) chart for the existing equipment. The chart can be in the form of a Microsoft Excel worksheet, that lists equipment on the Y-axis and time scale on the X-axis. Once the chart has been created for the first batch, the EO chart for the second batch is created by shifting data on the horizontal time scale toward the right until the bars do not overlap. This step can be repeated for subsequent batches. Commercial software also can generate similar EO charts.
Step 3 entails batch-time analysis. Cycle time is the current bottleneck in the production scheme, but it can be reduced by adding equipment, which will increase the number of batches per year. The EO chart can be used for this purpose, and separate charts can be created for the upstream and downstream processes if they are performed independently. This activity is continued until the cost index increases, but attention should be paid to ensure that cycle time remains the same after the final scheduling so that storage capacity is not exceeded. If this step is performed for a multiproduct biotech facility, debottlenecking is first performed for the different unit operations for one process and then for the different processes that are run in the same facility. A given process segment (e.g., harvest) must be complete for one process before the equipment is used in the second process. For a typical multiproduct facility, this step can be tedious because of the large amount of equipment, coupled with various process constraints.
Step 4 involves optimizing tanks and other equipment required for buffer and media preparation and storage. The EO charts can be used to calculate the volume and type of buffer and media required during each shift (8 h), and a preparation scheme is then prepared considering the stability of the buffer/media. The number of times a buffer/media preparation tank can be reused in a shift is calculated by dividing the shift time by the average time required for preparation of any buffer. The number of tanks needed can be minimized by maximizing the reuse of tanks.
This approach is illustrated in the two case studies presented below. Both of these case studies are based on real biotech therapeutic manufacturing facilities.
The objective of this case study was to perform process scheduling and debottlenecking to enable a five-fold process scale-up with minimal capital expenditure on equipment. We also proposed an optimized plan for buffer and media preparation and storage and calculated the number of storage and preparation units required. Data on the existing equipment were provided.
Figure 1. Equipment occupancy chart for the upstream process. The X axis denotes process time and the Y axis shows key process equipment.
The upstream facility produced a 30-L harvest batch every fourth day. The major equipment that existed in the facility included a disposable 40-mL flask, three 500-mL spinner flasks, two 5-L bioreactor, three 30-L bioreactor, a microfiltration unit, and a 30-L harvest storage tank. An EO chart for the upstream process was generated using SuperPro software and is shown in Figure 1. Analysis of the EO chart identified bottlenecking equipment, timescale of equipment usage, reuse of equipment, and excess equipment. In this case, the 30-L bioreactor was the bottleneck and it was evident that the number of spinner flasks was excessive; instead of three, only two are required.
Figure 2. Optimized equipment occupancy chart for the upstream process. The X axis denotes process time and the Y axis shows key process equipment.
To increase capacity, a future plan entailed purchasing two 150 L bioreactors and reusing the rest of the equipment from the current facility as much as possible. Figure 2 shows the minimum amount of equipment required and the optimized scheduling for producing a batch in the shortest time period. As we planned to use some of the equipment from the previous facility, the final list of equipment that must be purchased included a disposable 40-mL flask, a 500-mL spinner flasks, a 5-L bioreactor, a 25-L bioreactor, and a 150-L harvest storage tank. It was also decided to reuse the microfiltration unit, which was capable of handling a five-fold increase in throughput. With this scheduling, a batch could be manufactured every six days and 4 h, and the bottlenecking was now the 150-L bioreactor.
Figure 3. Equipment occupancy chart for debottlenecking for the five-fold scale-up. The X axis denotes process time and the Y axis shows key process equipment.
Figure 3 shows that adding another 150-L bioreactor shifts the bottleneck to the 5-L and 25-L bioreactors and reduces batch production time to five days. Because the downstream process (discussed in the next section) has a batch time of seven days, further debottlenecking of the upstream process was not performed.
Figure 4. Equipment occupancy chart for the downstream process. The X axis denotes process time and the Y axis shows steps for each unit operation.
The downstream process underwent process scheduling and optimization following the procedure outlined above for the upstream process. The resulting EO chart is shown in Figure 4. To enable a five-fold scale-up, it was decided to use the same equipment for chromatography 1( C-1), ultrafiltration (UF), and C-2 unit operations. Scaling of C-3, C-4, and C-5 unit operations was performed in accordance with a scaling factor of 2.5X, and C-6, C-7, and filtration unit operations in accordance with a scaling factor of 5X. The scaling factors were chosen as per the following guideline. The cost of the chromatography columns increases significantly with column size. Also, the column size is smaller for the chromatography steps that are later in the process. Hence, it is economically optimal to split the batch for the first half of the process (i.e., C3 to C5) and then later combine the two process segments in the second half (C6 and thereafter). This strategy minimizes the batch time, capital cost, and labor cost.
Figure 5. Buffer preparation scheme for a single-product biotech facility. Each column denotes a unit operation of the process and the rows denote the time in days.
The buffer preparation plan was scheduled and optimized in view of buffer stability and requirements. Many options were evaluated. The optimal solution with the minimal amount of buffer storage and preparation tanks is shown in Figure 5. An EO chart was made to minimize the amount of equipment by reusing equipment again if permissible. Figure 6 shows one such EO chart. Similar charts were created for all buffer preparation and hold-up equipment to identify the minimum equipment required. For this case study, the requirements are shown in Table 1.
Figure 6. Optimized equipment-occupancy chart for buffer preparation and hold-up tanks. The rows denote tanks used in the process, and the columns denote the duration of usage in days.
The above case study demonstrates how process scheduling and debottlenecking enable an economically optimal scale-up of the process and optimize the number of buffer and media preparation units. Batch size was increased five fold while buying only two 150-L bioreactors, five chromatography columns (i.e., C-3, C-4, C-5, C-6, C-7), and one 125-L harvest storage tank. In addition, the optimal number of buffer preparation and storage units was also estimated for preparation, and storage of 3,290 L of 30 different buffers per week.
Table 1. Requirements for the first case study
For this case study, process scheduling and debottlenecking was conducted for a biotech facility manufacturing four therapeutic products to minimize the required number of buffer preparation tanks. The harvest for the second product is generated as a byproduct from the first process. For the third product, harvest was generated from the first or second process. Similarly, the fourth product was generated from the first, second, or third process.
Figure 7. Equipment-occupancy chart for a multiproduct biotech facility. Rows denote various steps of each unit operation. The columns denote the process time in days.
Figure 7 shows the overall scheduling for the processes that require buffers. Using the available data on the volume and composition of buffer required for each step and the scheduling of each step, the requirement of buffer per shift and the total buffer requirement were calculated. Next, the tank requirement scheme was developed for each shift based on the volume of buffer to be prepared. All the buffers are prepared only one day before to avoid storage and stability issues. The scheme is shown in Figure 8. Next, tank usage was scheduled by maximizing the use of each tank by vertically shifting rows (e.g., we can use a 600-L tank for the preparation of 400-L buffer, if it is available). If buffer stability is not a problem, horizontal shifting can be performed to minimize the number of tanks and to level the daily workload. The optimized tank usage scheme is shown in Figure 9. Data are also shown in Table 2.
Figure 8. Tentative number of tanks required. Rows denote various tank sizes. The columns denote the process time in days.
The approach illustrated in this second case study is simple and effective; however, we did not consider the actual buffer preparation time in our analysis and instead approximated it using the average time for buffer preparation, which varied from 2 to 3 h. Tank wash and media transfer time was accounted for in the average time only. Although, this made the analysis quite simple and accurate for application here, alterations should be made for cases where this approximation may not be valid.
Figure 9. Optimized tank requirements. Rows denote various tank sizes. The columns denote the process time in days.
This case study shows the application of the suggested method for identifying the optimized number of buffer and media preparation tanks for a multiproduct biotech process. The above numbers of tanks are proposed for preparing a total of 48,235 L of 25 different buffers per week using only eight tanks with a total volume of 5,750 L.
Table 2. Requirements for the second case study
The methods in this article illustrate how process scheduling and debottlenecking can result in process intensification, thus reducing batch time, increasing plant productivity and, ultimately, decreasing the cost of manufacturing. The first case study focused on the procedure for scheduling a biotech process, debottlenecking, economical and optimal scale-up, developing a buffer preparation model, and optimizing the number of buffer and media preparation and hold-up tanks. The second case study demonstrated the development of a model for buffer preparation and the optimization of the total number of tanks required for a multiproduct biotech facility.
Anurag S. Rathore is a biotech CMC consultant and a faculty member, and Tarun Sengar is a student in the department of chemical engineering, both at the Indian Institute of Technology, Delhi, India, +91 9650770650, firstname.lastname@example.org. Rathore is also a member of the BioPharm International editorial advisory board.
1. Gosling I. Process simulation and modeling strategies for the biotechnology industry. Optimizing productivity in multiproduct batch facilities. Gen Eng News. Sep 2003. Available from: www.chemsim.com.
2. Toumi A, et al. Design and optimization of a large scale biopharmaceutical facility using optimization and scheduling tools. Pharm Eng. 2010;30(2):1–9.
3. Petridis D, Koulouris A, Siletti C. Throughput analysis and debottlenecking of biomanufacturing facilities. Biopharm Int. 2002 Aug suppl;28–64.
4. Yuan JQ, et al. Profit optimization for mycleia fed-batch cultivation. J Biotech. 1997;54:175–93.
5. Papavasileiou V, Koulouris A, Siletti C, Petrides D, Trans I. Optimize manufacturing of Pharmaceutical products with process simulation and production scheduling tools. ChemE Part A. 2007;85(A7):1086–97.
6. Petrides D, Siletti C. The role of process simulation and scheduling tools in the development and product commercialization. Pharm Eng. 2002;22(1).