This action might not be possible to undo. Are you sure you want to continue?
“The 1990s have seen a dramatic change in the way that we do business. Rapid advances in technology and increasing regulatory freedom have changed the rules of competition. Companies are now competing globally and traditional barriers between industries are breaking down. To cope with these changes and achieve superior performance, business leaders are moving towards new business paradigms that allow their companies to work more closely with their traditional and new business partners to adapt to the rapidly changing marketplace. This improved integration is the very essence of supply chain management. Supply chain leaders are reconsidering the linkages, not only between functions within their own company, but with other organisations up and down the supply chain.”
Supply chains are becoming more efficient and more responsive to the needs of increasingly demanding customers, driven by competitive pressures and supported by developments in information technology (IT). IT plays a major role in integrating supply chains and managing them more effectively.
Almost every industrial company is now considering the implementation of an advanced system to manage their supply chain more effectively, improve customer service dramatically, and reduce costs as well. These systems are Advanced Planning and Scheduling systems (APS) with marvellous names such as i2/Rhythm, Red Pepper and Manugistics.
With these systems it is possible to answer customer enquiries within seconds instead of hours or days. Speed is just one of the characteristics of APS. It promises that after implementation of APS, better throughput times, delivery times, inventory levels and utilisation rates result in higher levels of customer service and major reductions in costs.
During the recent years system vendors have put much effort in improving the functionality of APS systems. But what is the true value of these concepts? Are they as revolutionary as they sound? Implementations of these kinds of systems have dramatic consequences for the organisation. Is it worth to implement these new software packages?
The objective of this paper is to map the characteristics of advanced planning and scheduling systems and to find out the (use) fulness of these systems. Therefore the following problem has been formulated:
“Why (and how) should organisations implement an Advanced Planning and Scheduling system?”
To solve this problem several questions will be answered: What is supply chain management? What is supply chain integration? What is Advanced Planning and Scheduling? What is the difference between ASP and traditional planning systems? What are the current functionality’s of APS systems? What are the key success factors for implementation?
The integration of the Supply Chain:
“Like the medieval lords who built moats and walls around their castles many organisations have constructed artificial boundaries between themselves and the outside world. While these boundaries do not consist of water and bricks, they are just as difficult to surmount. More importantly, just as social evolution made castle walls obsolete; the new success factors of speed, flexibility, integration, and innovation are making boundaries between organisations less relevant. In fact, hiding behind such boundaries today can be more dangerous than venturing outside.”
World class companies are now accelerating their efforts to align processes and information flows through their entire value-adding network to meet the rising expectations of a demanding marketplace.
Some of the drivers for change that forces companies to overhaul their logistical structure are:
Increased regional and global competition:
The most potent force driving companies to overhaul their supply chains is increased cross border competition, regional and global. For many companies the competitive arena has become worldwide, rather than national or regional.
The role of the single market in Europe: Europe’s single market has intensified competition by tearing down the last protective barriers. At the same time the single market is an important factor which enables supply chain integration across borders. The dismantling of frontier controls has led to the speedup of road transport, which facilitates the switch from national to multi-country distribution centres.
Shorter product life cycles:
The trend towards shrinking product life cycles force a change in logistic management as it augments the risk of being stuck with obsolete inventory.
Changes in the market place:
National and cross border mergers and acquisitions in recent years have led to greater concentration of purchasing power in most sectors of industry. In the wholesale and retail distribution the growth of powerful chains is squeezing out the independents.
Pressure from smarter customers:
Major retailers and industrial end-users are becoming more sophisticated and more demanding. They are reducing their supplier base and are working more closely with the remaining suppliers.
Service as a differentiator: Products are more and more becoming commodities, forcing suppliers to search for new ways to differentiate themselves. Competitive edge will come from service differentiation.
The ability of an organisation to distinguish itself is coming to lie increasingly in the area of customer service. This places heavy pressure on the logistical chain. Delivering goods to customers in the most economic way while providing first-class service and quality is the logistics strategy. This requires more and more integration of the supply chain, in which all parts of the supply chain are linked to each other.
Suppliers and customers cannot be managed in isolation anymore, with each entity treated as an independent entity. More and more, there is a transformation in which suppliers and customers are inextricably linked throughout the entire sequence of events which brings raw material from its source of supply, through different value-adding activities to the ultimate customer. Success is no longer measured by a single transaction; competition is now evaluated as a network of co-operating companies competing with other firms along the entire supply chain.
Analytically, a supply chain is simply a network of material processing cells with the following characteristics: supply, transformation and demand.
An example of a supply chain is shown in figure below:
Supply Chain Management:
Supply chain management (SCM) is defined as a process for designing, developing, optimising and managing the internal and external components of the supply system, including material supply, the transformation of materials and distribution of finished products or services to customers that is consistent with overall objectives and strategies.
The essence of SCM is to develop a sustainable competitive advantage by reducing investments without sacrificing customer satisfaction. Since each level of the supply chain focuses on a compatible set of objectives, redundant activities and duplicated efforts can be eliminated. In addition, supply chain partners share information that facilitates their ability to jointly meet end-users´ needs. IT is an enabler and a key to the development of an integrated supply chain. However, this information must be shared by the partners. Research seems to suggest that there is a reluctance to share key information among partners. Many of these fears subside if partners share similar values and a common vision. Such information sharing heightens the alignment between partners such that effective supply chains share learning’s among partners rather than worry about knowledge expropriation. The goal is to orchestrate this alignment and to ensure that the supply chain is better than the sum of its parts. Adopting the concepts and tenets of SCM requires a new mindset. SCM requires looking at the complete set of linkages that tie suppliers and customers throughout the supply chain.
Supply Chain Integration:
All companies function as links in chains of entities that produce and distribute products. Many companies have viewed their participation in the supply chain from an independent
perspective, and have focused on the maximisation of its own profitability. This traditional view leads to the following types of boundaries in the supply chain, which reduce competitiveness by reducing speed, flexibility, integration and innovation.
Strategies and plans are developed independently: Each separate organisation has its own market targets, production plan, and schedule. The other parts in the supply chain are not consulted, which results in an unsynchronised supply chain.
Information sharing and joint problem solving are limited: Organisations withhold information about cost price, profit margins, and problems from other parties in the supply chain. The tendency is to solve these problems alone, often resulting in suboptimal solutions or delayed product delivery.
Resources are utilised inefficiently: In the different parts of the supply chain a lot of resources, expertise and knowledge are held separate from the other parts of the supply chain. All these separate parts use their own resources only for themselves, without the possibility of any other part to use these resources when they are temporarily superfluous.
Accounting, measurement, and reward systems are separate and
Each part of the supply chain has its own accounting, measurement and reward system. Some parts emphasise on quality and others emphasise on sales volume.
Sales force pushes products on salespeople’s terms: Salespeople focus on pushing products to the customers, while each part of the supply chain aims to maximise its own profitability. These salespeople do not listen to the requirements set by the customer which results in dissatisfied customers.
Successful companies will be those that take a systematic, boundaryless view of their participation in the supply chain. They must acquire an entirely new mindset, abandoning the legalistic view of organisations as independent entities linked only by market forces and learning to see themselves as part of an integrated system. By making specific external boundaries more permeable, organisation can dramatically increase speed, flexibility, integration and innovation. In the traditional view each organisation aims to maximise its own profit, while in the new model each organisation aims to maximise total supply chain success. The company in the new model will loosen its external boundaries and will follow a new model.
Business and operational planning are co-ordinated: In the successful supply chain, all members collaborate in both strategic and operational business planning. The goal is not only better product development and production planning, but also common or co-ordinated administrative and operational procedures such as billing, customer service, purchasing, shipping and inventory.
Information is widely shared and problems are solved jointly: As members of a system, participants in a boundary less supply chain share information more freely than before. A production problem in one part of the chain is everyone’s concern, and the best resources throughout the system are applied.
Resources are shared: A systematic view of the supply chain allows companies to deploy resources and expertise more efficiently throughout the chain.
Accounting, measurement and reward systems are consistent: A key requirement for a boundary less supplier-customer relationship is a common scorekeeping and incentive system so that everyone in the supply chain works off the same numbers, speaks the same language, and aims towards the same set of goals. Successful supply chains have jointly accepted methods to determine costs, margins and investments. Agreed-upon performance goals for each organisation unit are derived from
those methods. A matching reward system motivates employees to achieve the systemwide objectives.
Selling is a consultative process: In the boundary less world, successful companies engineer a significant shift in the role of their salespeople. Instead of pushing products, salespeople increasingly consult the customer, helping customers crystallise supply chain requirements and find optimal ways to meet those requirements and best utilise purchased products. In short, salespeople create a pull for a product. Overview traditional and new model view Traditional view Strategies and plans are developed independently Information sharing and joint problem solving are limited Resources are utilised inefficiently Accounting, measurement, and reward systems are separate and unsynchronised Salesforce pushes product on salespeople’s terms New model view Business and operational planning are coordinated Information is widely shared and problems are solved jointly Resources are shared Accounting, measurement, and reward systems are consistent Selling is a consultative process
Four forms of supply chain integration can be distinguished as follows:
Physical integration can be defined as those activities that focus on the improvement of efficiency of the primary process, by which the logistical costs of this process decrease, between minimal two entities in the supply chain. An example of physical integration is the use of standardised transportation devices.
Information integration: A second form of supply chain integration is activities to attune the flow of information. As with physical integration, the primitive form of the logistical process and the management system do not change. An example of information integration is to forward shipping information from shipper to transporters.
Management control integration: Management information, out of other entities in the supply chain, is used in a systematic way to integrate several parts of the supply chain. The goal is not only to generate cost benefits, but also to realise a better customer service level. By connecting the management information between entities in the supply chain, the total supply chain can respond quicker and more effective to the market requirements. An example of this integration is a supplier who receives information from its customer about the inventory level of a specific product.
Organisational integration: Parts of the management activities come to lie at another entity in the supply chain. This concerns more than just the outsourcing of operational activities. It concerns the assignment of logistical planning tasks. An example of organisational integration is a company which partly takes care of the production planning.
Planning in logistical networks takes place on three hierarchical levels: I. Strategic
II. Tactical and III. Operational.
The planning at tactical level aims mainly at minimising the costs associated with the production and distribution of products under all sorts of constraints like available capacity, stock, personnel and finances, while there is a certain demand of customer service.
In this paragraph all the historic planning systems will be described briefly, starting with statistical inventory control (SIC). After the description of Material Requirements Planning (MRP I), Manufacturing Resources Planning (MRP II), Distribution Resources Planning (DRP) and Enterprise Resources Planning (ERP), this paragraph will end with a short description of APS.
Statistical inventory control:
SIC is static in nature and operates solely on the basis of a predicted forecast. This method of inventory management employs a number of mathematical techniques to control inventories, based on historical turnover data. This method of inventory management is easy to computerise.
Material Requirements Planning:
The computerised data-processing techniques introduced in enterprises from 1950 made it possible to perform complex calculations and to process large amounts of data. In this period MRP I systems were developed. For the first time the factor ´time´ made its entry into inventory management. MRP I systems operate on the basic of the existence of socalled dependent demand that can be calculated from a requirement for a product with an independent, predictable demand and the factor time in controlling inventories.
MRP I comprises a number of information-science techniques to plan material acquisition (the inflow of the necessary raw and auxiliary materials and semi-manufactures) and the
production process on the basis of an established production plan for end products. A production plan is determined on the basis of market and turnover expectations. The composition of each product in terms of components (raw materials, auxiliary materials and semi-manufactures) is known and set out in a bill of material.
Given an established production program for a specific period, the planner uses MRP I to calculate which components are required in what quantities and at what point in time, by examining the throughput time or delivery time of the component (scheduling).
Manufacturing Resources Planning:
MRP II is an extension of MRP I, which assumes unlimited capacity. The extension to MRP II involved the calculation of the required capacity. On the basis of a required production program, MRP II calculates back from the delivery data to determine what capacity is required in what quantity and at what point in time in order to deliver the orders punctually. It is important to know at an early stage which capacity element in the process (machine, people, money, supplier, etc.) will constitute the bottleneck and when.
Distribution Resources Planning:
A distribution network consists for the most part of several consecutive inventory points; for example the factory, a central distribution centre (DC) and national sales warehouses. In a distribution network, co-ordination of the various activities (sales forecast, orders, transport and inventories) is essential. The principles of MRP I/II (dependent demand and scheduling) are also used in inventory management in distribution networks: DRP.
DRP is an information system that supports co-ordination within the distribution network. The purpose of such a system is to record goods flows and it requires that information must be available on where stocks are held, which goods are in transit and what are the changes in inventories. DRP makes it possible to co-ordinate the decisions taken at various point in the distribution network.
Enterprise Resources Planning:
ERP is defined as a software architecture that facilitates the flow of information between all functions within a company such as manufacturing, logistics, finance and human resources. It is an enterprise-wide information system solution. An enterprise-wide database, operating on a common platform, interacts with an integrated set of applications, consolidating all business operations in a single computing environment. Ideally, the goal of an ERP system is to be able to have information entered into the computer system once and only once. For example, a sales representative enters an order into the company’s ERP system. When the factory begins assembling the order, shipping can check on the programs to date and estimate the expected transport date. The warehouse can check to see if the order can be filled from inventory and notify production of the number of products still needed. Once the order gets shipped, the information goes directly into the sales report for upper management.
ERP provides a backbone for the enterprise. It allows a company to standardise its information systems. Depending on the applications, ERP can handle a range of tasks from keeping track of manufacturing levels to balancing the books in accounting. The result is an organisation that has streamlined the data flow between different parts of business. In essence, ERP systems get the right information to the right people at the right time.
As a result of ‘island automation’ of individual parts of a company there are hardly, if any, links between those parts. However staff of one department needs a better understanding of other departments’ processes. ERP systems are helpful in this context. These systems take care of the entire administrative process of the various units within a company. A company can use an ERP package to drive all processes, such a financial management, sales forecasting, purchasing, inventory management, production control, logistics, project management, service and maintenance. Examples of ERP systems are Baan, Oracle, JD Edwards and SAP.
Advanced Planning and Scheduling:
“An APS system is a system that suits like an umbrella over the entire chain, thus enabling it to extract real-time information from that chain, with which to calculate a feasible schedule, resulting in a fast, reliable response to the customer. With the help of APS it is now possible to answer customer enquiries within seconds. This is just one of the possibilities of APS. The suppliers of APS can demonstrate impressive results: after implementation of APS, better throughput times, delivery times, inventory levels and utilisation rates result in improved operating results and a higher level of customer service.”
There are two reasons why the interest and demand in APS systems arises at the moment. The first is the development of memory resident servers. Memory resident means that the entire planning engine, model and database are kept entirely in memory. This means very complex manufacturing and supply chain operation models can be stored in memory totally. This development provides a major advantage, because it eliminates disk access time and that gives serious time reduction in solving the planning problems. It allows very fast processing of large datasets, which makes simultaneous material and capacity problem solving possible.
The second reason is that companies are uniting their supply chains. Companies start to understand how the value chain works. Co-operating companies should manage their supply chains in one process. APS systems make it possible to co-ordinate these different supply chains in one system. System suppliers that successfully evolved to this level of planning and scheduling did so because they broke out of the traditional factory-only or distribution-only focus.
APS is a new revolutionary step in enterprise and inter-enterprise planning. It is revolutionary, due to the technology and because APS utilises planning and scheduling techniques that consider a wide range of constraints to produce an optimised plan: I. Material availability II. Machine and labour capacity III. Customer service level requirements (due dates) IV. Inventory safety stock levels V. Cost VI. Distribution requirements VII. Sequencing for set-up efficiency
Planning systems versus supply chain integration:
In this paragraph the planning systems will be classified in a diagram, which is shown in figure below:
Classification of planning systems in an environment/complexity diagram: Complexity Functional Integrated within Environment Static SIC MRP DRP Dynamic APS APS ERP APS Integrated outside APS
The two axes of the diagram are:
The difference between static and dynamic is the level of predictability of the environment. In a static environment there is no need to reschedule or recalculate the plans that are made, because the environment is highly predictable. The organisation is familiar with the (number of) required products for the next period. Therefore it is enough to do the planning or scheduling at pre-defined times for a pre-set period. Instead, in a dynamic environment this predictability is very low. Due to this low predictability it is necessary to be able to reschedule plans very easily, and on a minute to minute basis.
The complexity is divided in three layers of integration. The first layer is a “functional” organisation. In these kinds of organisations the departments try to optimise their own department, without considering that it may not be optimal to the whole organisation. The second layer is “integrated within” one organisation. In this layer a company is process driven and integrated. No outside information is gathered to optimise the
planning. A separate organisation is an organisation with own profit/loss responsibility. The third layer is “integrated outside” the organisation. When information of a production site with own responsibility for profit/loss is shared with the sales-organisation, these organisation is an “outside integrated” organisation.
In the following subparagraphs the planning systems which are named in the above paragraph will be classified in the diagram with the axes environment and complexity.
This planning system will only function in a static environment in a “functional” organisation, because of the limited possibilities of this planning system. Some of these limitations of SIC are: Future requirements cannot always be predicted on the basis of historical data. The specialist know-how that the planners have acquired is not used in the purely statistical approach to inventories. Due to these limited possibilities it is only possible to use SIC in a static environment. It is also not possible to use it for complex problems.
Another disadvantage of SIC is that it results in the Forrester-effect. This effect is the result of the fact that different parts of the supply chain make independent decisions about inventories on the basis of its own stock calculation methods, which are static. These independent decisions result in higher and unbalanced stocks in the whole chain.
MRP I/MRP II and DRP:
These planning systems are now still operational in many organisations. In the functional organisation the planning is done separately for the various links in the chain. The planning is executed sequentially. The systems can only handle environments that are static and therefore also result in the Forrester-effect, because the various types of planning Master Production Schedule (MPS), MRP, and Capacity Resources Planning (CRP) affect each other due to the sequential process. The output of for example the MPS is the input for the MRP I/II run.
An ERP system can function very well in an environment which is still very static. An
ERP system is ideal in companies that want to integrate their information flow within the organisation. In multi-site companies this can be viewed by the procedures. Each site (or profit/loss companies) has its own ERP system. It optimises the information flow for only that single site. An ERP system can be seen as a database which is surrounded by all sorts of applications. The database is the device that makes the integration in that company possible.
An APS system can function in a number of environments and types of complexity. When companies start to integrate within their organisation an APS tool can be helpful, because the MPS-MRP-CRP planning process can take place simultaneously. An APS tool really benefits companies integrating with outside organisations. The customer and suppliers are involved in driving the organisation’s logistical chain. Logistical planning and sales are merging in order to be able to respond rapidly to market requirements. The APS tool can be helpful in dynamic environments, because it has the advantage of being really fast in recalculating the plans whenever necessary. Another benefit of this system is that it facilitates the combination of information of multiple sites and that it calculates an optimal plan for a complete supply chain.
Advanced planning and scheduling:
APS can be viewed as an umbrella technology which uses a number of features which are described in the below paragraph. The scope of APS is not limited to factory planning and scheduling. It includes a full spectrum of solutions, both enterprise and interenterprise planning and scheduling systems. Differences are not only the time horizon, but also the level of the planning horizon, such as strategic, tactical or operational planning is considered. Based on Advanced Manufacturing Research, the following solutions can be distinguished:
Strategic and long-term planning: This solution addresses issues like: Which products should be made? What markets should the company pursue? How should conflicting goals be resolved? How should assets be deployed for the best ROI?
Supply chain network design: This solution optimises the use of resources across the current network of suppliers, customers, manufacturing locations and DCs. What-if analyses can be performed to test the impact of decisions to open new or move existing facilities on profit and customerservice level. It can also be a helpful tool to determine where a new facility should be located to fulfil customer demand in the most optimal way. These supply chain network design tools are mostly applied to find the balance between holding more stock at a specific location or making more transportation costs.
Demand planning and forecasting: Both statistical and time-series mathematics are used in this solution to calculate a forecast based on sales history. A demand forecast is unconstrained because it considers only what customers want and not what can be produced. Based on the information from the forecast, it is possible to create more demand through promotions in periods where the demand is less than maximum production.
Sales and operations planning: This is the process which converts the demand forecast into a feasible operating plan which can be used by both sales and production. This process can include the use of a manufacturing planning and/or a supply chain network optimising solution to determine if the forecast demand can be met.
Inventory planning: This solution determines the optimal levels and locations of finished goods inventory to achieve the desired customer service levels. In essence, this means that it calculates the optimal level of safety stock at each location.
Supply chain planning (SCP): SCP compares the forecast with actual demand to develop a multi-plant constrained master schedule, based on aggregate-level resources and critical materials. The schedule spans multiple manufacturing and distribution sites to synchronise and optimise the use of manufacturing, distribution and transportation resources.
Manufacturing Planning: Develops a constrained master schedule for a single plant based on material availability, plant capacity and other business objectives. The manufacturing planning cycle is often only executed for critical materials, but that does depend on the complexity of the bill of material. Also the desired replanning time is a factor that one must take into account when deciding which level of detail is used. For example, with a simple bill of material a complete MRP I/II explosion can be executed in a few minutes.
Distribution Planning: Based on actual transportation costs and material allocation requirements a feasible plan on the distribution of finished goods inventory to different stocking point or customers, is generated to meet forecast and actual demand. With this solution it is possible to support Vendor Managed Inventory.
A solution which uses current freight rates to minimise shipping costs. Also optimisation of outbound and inbound material flow is used to minimise transportation costs or to maximise the utilisation of the truck fleet. Another possibility is to consolidate shipments into full truckloads and to optimise transportation routes by sequencing the delivery/pickup locations.
Production Scheduling: Based on detailed product attributes, work centre capabilities and material flow, a schedule is determined that optimises the sequence and routings of production orders on the shop floor.
Shipment Scheduling: This solution determines a feasible shipment schedule to meet customer due dates. It determines the optimal method and time to ship the order taking customer due dates into account.
APS solutions related to the time horizon
Differences in planning horizons:
The enumerated solutions can roughly be divided into three levels of planning and scheduling: I. Supply Chain Planning II. Manufacturing Planning III. Production Scheduling
Relationships of major planning functions with typical data flows:
The first two levels can be called planningcentric systems. These systems focus on long term strategic and some
tactical objectives. For a global or a multi-site company, these systems can optimise the best possible location in a network of manufacturing locations where a specific order must be produced. The planner enters the business objectives into the system, after which the planning engine determines which objectives might be violated. When objectives are violated in the long term it is possible to adjust the constraints, which results in gained objectives. Adjustments in the constraints might be possible if there is enough time. When there is not enough capacity, in the long term this constraint can be eliminated, because capacity can be enlarged by acquiring an extra production line.
The third level is more a scheduling-centric system. These systems focus more on operational and some tactical objectives. The task of a production scheduling system is to generate a feasible production schedule given a required production output. The constraints it deals with are quite real; they are often given and allow only limited changes.
Supply Chain Planning:
This SCP group takes a forecast and looks at actual demand, after which a constrained operation plan for both manufacturing and distribution is generated. A multi-plant constrained master schedule is the output of the SCP process for manufacturing. To create this output the material availability’s and plant capacities are accumulated. For some industries, transportation requirements and set-up sequencing are considered as well.
Advanced Manufacturing Research (AMR) describes SCP as follows: “SCP determines what should be made given the available resources to achieve business goals.”
The output from manufacturing planning generally is a constrained master schedule for a single plant or a group of similar plants. This master schedule considers the constraints in a more detailed perspective than in SCP. In manufacturing planning a full MRP I/II explosion can be included in the process.
AMR describes Manufacturing Planning as follows:
“Manufacturing Planning determines how and when it should be made based on material and resource constraints to meet customer demand.”
The goal of this group is to translate the output of the supply chain planning to an operational plan and work orders. Here is where the ultimate specification takes place on the basis of which the suppliers will deliver., the production departments produce and distribution receives and ships the products. APS supports the planner by continuously adapt or suggest adaptation of the planning and scheduling based on the recent information. Product scheduling is designed to produce the most efficient production schedule (where the throughput times are minimal, the output maximal and the costs are low).
Planning and scheduling:
An APS system uses the following planning and scheduling approach: A planner module which pays some attention to capacity constraints produces a “scheduleable” plan. This plan then feeds a scheduler module, which produces a detailed list of operations showing how capacity will be used and returns this information to the planning function for use in the next planning period. The data regarding current and planned operations can also be used to provide realistic estimates of the ability to meet a new customer order request. This integration of planning and scheduling is described in the following two paragraphs.
The role of planning in APS is to determine what demands on the production system will be met over a given planning horizon. The input to the planning process includes information on manufacturing capacity and demand data. Demands may be of several types: customer orders, forecast, transfer orders (i.e., orders from other plants), released jobs, or replenishments of safety stock. Manufacturing system data includes bills of material, workcenter availability, part routings through workcenters, and inventory (both on-hand and scheduled for delivery). The output from the planning process is a feasible plan, which provides release and completion times for every demand. Like MRP, APS takes into account the availability of materials. Unlike MRP, it also takes into account the capacity of workcenters to process the material and satisfy demands.
This planning process is order-centric, focusing on the demand for end items and determining how much demand can be met in a given time period. Exactly how that
demand will be met, in terms of specific assignments of jobs to workcenters and their sequencing, is left to the scheduling function. It is in fact often desirable for a plan to be somewhat tentative, since it covers a planning horizon subject to disruptions. Forecast may not be accurate. Deliveries may be delayed. Equipment may fail. Unexpected rush orders may be received. Therefore planning is not expected to be highly detailed. Individual machines may be aggregated into a workcenter with no determination of which will be used by a specific order. Setup times may be averaged since sequencing at this time is premature. Buffer times may be defined, especially prior to processing on bottleneck machines, to allow for possible disruptions. The end result is a “scheduleable” plan.
The role of the scheduler module in APS is to produce a detailed list of operations specifying which orders are to be worked on at which workcenters and at what times. The input to this module includes all demands to be satisfied, including the internal orders added by the planner module when an end item required a component to be manufactured. It includes the current material inventory levels as well as planned deliveries or purchased materials. It also includes the same manufacturing system data as that provided to the planner module but uses a more detailed representation of that data. Detailed information used by the scheduler module that is not pertinent to the planner module includes:
I. Variable run times based on the machine and operator actually assigned. II. Rules for selecting machines and operators based on skill sets and quality requirements. III. Variable setup times based on the previous and next part characteristics such as part type, family, colour, width, etc. IV. Rules for sequencing jobs at workcenters, based on minimising setup and other factors. V. Allowable shift overruns. VI. Rules for selecting from a list of prioritised jobs based on due date, slack, cost and other factors.
The result is an accurate representation of what to expect on the shop floor in the immediate future. While the planner module typically considers demand on the system over a few weeks or months, the scheduler module will typically work with a much shorter time frame such as a shift, a day, or a week. The usefulness of a detailed schedule degenerates quickly as time passes, since disruptions on the shop floor or changes to the order mix may require significant adjustments.
Now, lets take a example of a refining company:
Fundamentals of Supply Chain Management for Refining:
Elements of Refining Supply Chain Management:
The refining "Supply Chain Management" (SCM) tools support the activities for running tan oil refinery, from feedstock selection to optimized planning of operations using feasible schedules, to finished product blend planning and scheduling, and finally, optimizing the distribution of products from the refinery to bulk distribution terminals. The purpose of this paper is to provide an overview of these activities and typical tools used in every work in running a modern refinery.
Crude Feedstock Selection and Evaluation
Generally, most refineries run a mixture of a number of crudes, sometimes blended to improve the feed "diet" and obtain better yields. If there are many crude choices available, there are tools that allow crude blending optimization that calculates an optimal blend recipe based on the crude distillation unit yield vectors (used as specs) and crude blend components assay data...
There are three common methods of crude valuations, from the simplest to the most complex, but profitable : I. Gravity and sulfur-based valuation II. Assay-based cutpoints product valuation III. LP Model product-based valuation
All these methods use crude oil assay information as the input for the various yields and corresponding valuation calculations. There are roughly three kinds of crude assay data
available: The US Bureau of Mines, The Oil &Gas Journal, and proprietary assays by producers of crude. They vary from cursory data to comprehensive data provided by crude oil producers (which includes also yield and property data for reforming, cat cracking, etc).
The gravity and sulfur-based valuation: This is the oldest method, and it considers the differentials based on gravity and sulfur differences, e.g. the market relative crude value is Y= a+b*API+c*Wt%S
The assay-based cutpoints product valuation: It uses the crude oil assay set of boiling-point defined distillation cuts to represent the product yields to which the corresponding product market prices are applied to value the crudes.
LP Model product-based valuation: It uses the crude assays together with a Linear Program (LP) refinery model which together with product prices simulates the refinery operations which produces a yield of products that maximizes the total value of products refinable from given crude.
For a given refinery, the "relative value in use" between crudes A and B is the difference in value between the marketable products that can be made from crude A or B.
The complete evaluation of a crude has to be done in the context of the "normal" operation already running a mixture of crudes as a feed to the crude unit, and running” delta" cases with and without the crude being evaluated. In addition, if one chooses to optimize the feed diet, this requires non-linear blending of crude assay data for each of the crude blend components, in addition to calculating an “optimal" blend diet recipe. This obviously is practical only when using the LP model-based tool.
Crude scheduling refers to the logistics of handling the crude deliveries and corresponding tank allocations, crude runs, and crude inventories, with the objective of optimizing a feasible daily crude operations sequence while meeting constraints (e.g. available tankage).
The structure of a crude scheduler consists of a time dependent picture at-a-glance of receipts, inventories, and crude runs, with either daily (OK for rough scheduling) or hourly granularity (allows to see problems with tankage, inventories, etc. and be able to take action, e.g. modify tank service allocation, change blends, run rates, etc.).
A crude scheduler uses the following type of information: I. Crude Receipts (via tanker, pipeline, by crude type) II. Crude unit runs (for multiple CDU's, individually, by crude types) III. Crude inventories (by crude type, calculated as Opening inventory+ReceiptsRuns) IV. Crude rates (by crude type) V. Tank min/max capacities and service allocation VI. Equipment availability flags(tanks or CDU's out of service, etc) VII. Crude assays VIII. Crude prices IX. Product prices
Typically, a coastal refinery receives crude via ship and pipelines. The crudes are fed to one or more Crude Distillation Units (CDU), some of which must operate on blocked runs to meet specialty product demands, like lubes. The crude operations coordinator/scheduler must set crude blocks and rates to contain the crudes in the available tankage as well as meet the planned crude runs for the month. The crude scheduling tools allow the crude coordinator to follow inventories and the effect of the planned runs on a hourly/daily basis, for up to 3o to 45 days.
The technology can vary from spreadsheets to sophisticated mixed integer linear programming (MIP) for event sequence optimization, to expert systems encapsulating refinery scheduler know-how in the form of "knowledge-based" rules. The man-machine interface can be tabular and/or graphical, but this is secondary to its practical usefulness.
Typical benefits realized by using a crude scheduler are: I. Better resolution of potential conflicts with the timing of crude oil receipts; this allows reduction of safety stocks by at least one day.
II. Better management of refinery crude tankage to optimally blend crude oils. This reduces freight costs by allowing receipt of bigger parcels. III. Minimization of feedstock switches to the crude units, thereby maximizing yields and minimizing off spec products and reruns. IV. Maximizing the length of crude runs improves the optimization of refinery performance to meet product demand and minimize reblends. V. Invaluable tool in troubleshooting operating problems experienced with some type of crudes, or evaluating impact of refinery upsets or equipment failures and evaluating alternatives VI. Improved communication between planning and operations
Refinery Planning and Scheduling:
Traditionally, there is a confusing vocabulary when describing planning and scheduling of refinery operations, intimating that they are one and the same, except for the time element. NO, they are not the same, because they have different objectives :
Planning tools develop a refinery operating plan while attempting to maximize profits and minimize costs while making a variety of products and meeting tight environmental requirements. This is done through the use of process unit’s models and using an optimizer to maximize/minimize the desired parameters. The results are the best long term "average"' operation targets for the process units to meet the specified economic objectives.
Scheduling tools are concerned with practical feasibility, not economic optimization, and by its very nature, it is a time dependent problem. Scheduling handles the practical aspects of implementing a feasible refinery operations plan - a doable operations sequence schedule, on a daily basis, taking into account inventories, liftings, availability of process equipment and tanks.
In terms of commercially available software, there are both separate programs for planning and separate packages for scheduling, or hybrids, using planning software with a multiple time period "staging" functionality to simulate a scheduler's day-by-day activities.
Refinery Operations Planning:
Typically, an operating plan is made for the next year (long range planning), 6 months (medium range), monthly (short range plan) and a weekly plan (the one that is passed to the scheduler for feasible scheduling purposes.
Plans are always averages over the time horizon, developing a good set of operating goals for a future period. As a consequence, the results are the best "average"' operation targets for the process units to meet the specified economic objectives.
The tools use either commercially available process simulators coupled with an optimizer, or Linear Programming (LP) with non-linear optimizers with built-in process models as part of the structure of the LP equations. The difference between the two is in the accuracy of the process unit models embedded in the tools. The models in the process simulators are quite accurate, because process simulators have other uses, such as process design and troubleshooting; for operations planning, it might be an overkill, given the time horizons involved and the uncertainty of data used in the models. The Linear Programming (LP) with non-linear optimizers use less accurate process models to find quickly a single feasible best solution given a possible range of conditions, and generally are the major planning tool of choice in with refinery planning and economics departments.
Both techniques use a refinery model built of individual process units. One can use a full scale, highly accurate model of each process, or limited mini-models incorporating only the critical processing limitations. A variation of the mini-model is the Extreme Point Modeling LP formulation (EPLP) of the refinery, which is an extremely condensed version of a full scale refinery model which, when given the same input data (crude price and availability, product volumes and values, and unit capacities), will give the same answer as the detailed refinery model. The "condensing" is done by representing only the major handles /modes that are available to the planner to alter the operations.
EPLP planning models have two major parts:
I. The Base case, which is an optimized refinery plan operation for some period, say, one month; it contains feed purchases, product sales, intermediate volumes, capacity utilization, and key product qualities based on results from a detailed refinery model. II. Delta feed and process shift logic, which defines the ultimate yield effects on a refinery-wide basis, and delta capacity utilizations for changing feeds or varying process alternatives, again, the data being based on results from a detailed refinery model.
The planning tool outputs usually are detailed reports with the operating targets, yield, and balance and stream data for the process units, and a summary of refinery economics. The reports are usually tabular, with some limited graphical summaries of key results. As one expert planner with many years of experience put it, "the relevant information of an LP output is often submerged in a sea of detail. The LP should provide a clear message on business drivers, constraints, and flexibility…on one "A4" page" .
Refinery Operations Scheduling:
The scheduling objective is to come up with the BEST practical (feasible) daily work plan on how to sequence crudes, provide instructions on running the process units, what tanks to use, what product blends to do in what sequence (so one does not run out of components), and finally, the product shipment sequence.
Refinery operations scheduling is one of the most challenging activities, where the technology and tools have not been successful across the board. The reasons for this is COMPLEXITY, covering the crude system logistics, process units, blending, offsite, and shipping, each with vastly different time constants, economic models, procedures, and constraints. In addition, the scheduler has to react to daily process and business variability, human interaction unpredictability, and time consuming integration of data from various disparate information sources. As a consequence, traditional scheduling activities were divided into:
I. Crude scheduling. II. Process unit scheduling. III. Blending and offsites dispatching scheduling.
The crude scheduling is concerned with crude tank allocation, optimal sequence of crude batches in pipelines, and optimal crude blending operations to feed the best crude diets for crude distillation. The tools and techniques were reviewed previously.
Process unit scheduling is concerned with determining unit operating conditions, unit by unit, time period by period, product status, characteristics of each stream and each tank period by period - this is what we are reviewing here, in this section. Blending and offsite shipments scheduling is concerned with the sequence of product blends and optimal recipes so as to optimize component and product inventories, status of receipts and shipments period by period, oil movements schedule, and the status each tank or group of tanks quantity and quality, period by period. The tools and techniques will be reviewed in the next section.
Many years ago, the thinking was to use the refinery planning tools, by running it sequentially for multiple time periods, that is, a normal planning model is dropped into an LP system multiple times. The models are tied together with variables that represent transfer of streams (for example, inventories) from one time period to the next. The disadvantage of this approach is that the information is inadequate for scheduling, (e.g. pooled qualities vs. individual tank qualities), and the scheduler gets swamped with information that is not relevant to scheduling activities, so he has to spend more time "fishing out" what he needs - not exactly very efficient when you run on low manpower levels, and you are pressed for time to exploit "opportunistic" commercial purchases and sales.
The next evolution in thinking was to divide the scheduling tools into three distinct categories for crude scheduling, process unit scheduling, and blending/offsite scheduling, addressing the specific requirements and idiosyncrasies of these operations. There are individual schedulers/coordinators to handle each one of these areas. THIS IS CURRENTLY THE MOST USED APPROACH.
Finally, modern scheduling systems are optimized for efficient and rapid generation of feasible and optimal operations schedules. The approach is to use a generic refinery model encompassing the crude system, process units, the piping network, and the offsite facilities, together with a knowledge base to introduce the scheduler experience and constraints, an event-driven optimizer to handle equipment status/mode of operation, and an object-oriented graphical data base for easy human manipulation and customization.
The scheduling "time horizons" are generally a week to ten days (to cover a week with a long weekend holiday) to a month to a month and a half; obviously, the quality and credibility of the data decrease with a longer time horizon. The typical granularity of schedules is daily, hourly, (although some vendors provide minute resolution, this is overkill unless you are doing the scheduling for a fast reactor….)
The result is a "PRODUCTION SCENARIO", including crude blends, process unit instructions (cut points, severity, feed rates, etc), tank levels and quality profiles. The scheduler can modify the production scenario interactively. Although graphical user interfaces are in vogue, a tabular display provides equal or more meaningful information for quick assimilation or modification by the scheduler.
Product Blending Planning and Scheduling
The product blending operations are generally batch blends of multiple components, using either sequential pumping to a blend tank, or inline blending to a tank or pipeline or ship. The blend components come from a mixture of closed tanks, running tanks and direct process unit rundowns.
Blending is also special because of the intimate interactivity between blend planning and scheduling: scheduling of blends is dependent on planned recipes which themselves are dependent on component availability and tank containment problems, which, of course, are caused by recipe determined component drawdowns…that is why, generally, blending uses INTEGRATED planning and scheduling tools.
To complicate matters, the new environmental rules and regulations, products have to meet 20 to 30 property specs simultaneously, compared with 2 to 3 specs 20 years ago. With the new mergers, acquisitions and consolidations, numerous merchant refineries blend products for others marketers, so one has to produce numerous slightly different grades of products (e.g. it is not unusual to see 60 different grades of gasolines made by one independent refiner).
In addition, there is economic pressure to minimize working inventory and maximize the facilities (tanks, pipelines, pumps, berths) utilization factors to reduce capital and operating costs.
The purpose of product blending planning and scheduling is to address these operating objectives: I. Meet quality specifications. II. Meet shipment schedules. III. Minimize quality giveaway. IV. Minimize blend cost. V. Maintain inventory targets within minimum/maximum levels.
The time horizon for blending planning and scheduling is determined by two factors: I. Accuracy of information regarding product and components supply and demand schedules II. The need to maintain an average inventory to satisfy a less accurate supply and demand schedule
In practice, that means a two time period horizon, the first period covering one week to 10 days, when the demand is quite accurately known, followed by a second "average" operation period, of say 3 to 4 weeks, where the information might be "soft". This insures that the actual blends done in the first week are not going to use all the "good" components, making it impossible to make decent blends in the subsequent weeks.
The output from the blend planning and scheduling tools are: I. A detailed product blend schedule
II. Optimal blend recipes III. Component and product tanks to use IV. Predicted properties of the blend recipes V. Blend component and product inventories as a function of time
This information is usually included in a "pumping order" sent either electronically or via paper to the offsites operations people for execution. Modern control systems use an online optimizer (copy of the planning optimizer) together with an on-line multi-property analyzer to obtain the benefits of an optimal blend recipe.
What are the benefits of a blending planning and scheduling system? I. Meet quality specifications without reblends II. Meet shipment schedules and avoiding demurrage III. Minimize quality giveaway even under delivery time pressures IV. Minimize blend cost by using the cheapest components available V. Minimize inventory levels without running out of components or overflowing tanks VI. Minimize lab work load VII. Lower component costs by adjusting component cuts/severity to match product grade demand VIII. Buying/selling components to minimize costs IX. Reacting quickly to opportunistic commercial demands
Typical benefits for 125,000 bpd gasoline blender exceeds $9 millions/year .
Measurements used for applications such as planning, scheduling, and optimization are subject to errors which affect the credibility and usefulness of the results of the applications that use them, such as planning predictions. The errors can be random or systematic, and are due to causes such as: I. Noise and vibration II. Drift III. Bias IV. Deterioration of components V. Wear of parts VI. Corrosion of sensors VII. Fouling
VIII. Improper calibration The Data Reconciliation (DR) technique works only with errors that are normally distributed random variables. Thus, the first step in the technique of data reconciliation is to identify and eliminate gross (systematic) errors using statistical quality control procedures, and then develops mass, volume, and energy balances around a process unit or a plant that must balance EXACTLY. The imbalances from forcing an exact balance are then used to correct the raw measurements using a mathematical "least square" minimization technique. DR uses various linear and nonlinear optimizers to minimize the DR objective function, the sum of the squares of the ratios of the measurements adjustments to the measurements standard deviation, subject to the balance constraints.
The raw measurements used for DR must be from the steady-state operation of the process unit. There are techniques to detect “steady-state" operation, but most frequently on can use the readily available process variable hourly or daily averages in a DCS historical data base as the "raw" values.
What are the DR benefits? I. More credible operating plans II. Better monitoring of process performance and higher accuracy of process yield measurements III. Aid in detecting faulty instrumentation IV. More accurate operating data for troubleshooting analysis and process improvement V. More accurate accounting and loss control
Aggregately, these DR benefits are in the range of 0.5 to 1.5 M$ in complex refineries.
Supply Chain Management Problems and Promises:
Refining SCM has the potential to add 60 to 80M$/year in hard benefits accruing from increased yields and thruputs, lower inventories, higher equipment utilization factors, and manpower efficiency. What are some of the barriers to achieving this?
Integration Today one might have a planning department responsible for planning and economics, plus 3 to 4 coordinators (crude, mogas, distillates, fuel oils). Each one of these use different sets of tools, and have to get their data from widely scattered sources of information, some electronically, some by FAX, some by phone, pieces of paper, etc. Then, this data needs to be checked, and then integrated in the respective planning/scheduling tools, and then run test cases, and finally, the case of interest today. This is very time consuming, and people in these jobs are under enormous stress: the time factor, with ships coming, process unit upsets, missed pipeline shipments, arguments with the Lab, problems with the blender controls or the on-line analyzers, psychological impact of mergers, on and on.
All of the above activities and tools have been integrated before, done electronically and automatically by some innovative refineries, but they are the exception. Planning and scheduling systems create work orders/instructions sent automatically for execution by process control with operator/supervisor approval .
There is no excuse for this state of affairs, given the cheap computing power available and refinery-wide networks. Also, there is no excuse for lack of integration between commercial vendors own software, that is, their planning software not "talking" with the scheduling software, etc, by this, meaning automatic data transfers, without human intervention. The management consulting companies and systems integration companies do this type of integration work, but they charge "an arm and a leg", and you are still stuck with "one-of-a-kind", hard to maintain system.
Databases A modern relational data base is perfectly capable to hold years of process, performance, quality and economics data. It can run on inexpensive PC-based servers connected to the refinery-wide network, and can be accessed by the planning and scheduling tools running on PC's in the refinery using industry standard interfaces. One central planning and scheduling data repository eliminates the multiple "private" data bases used by each planner/coordinator, together with discrepancies, redundant and inconsistent data or even wrong data.
Quality of Data and Models One thing that has not changed from the beginning of the computer age is the "garbage in - garbage out" dictum.
Data has to be checked for accuracy, validity, consistency, and completeness. This does not have to be exaggerated, but has to be done, preferably with "automatic" data filters. The accuracy has to be put in the context of how data is to be used; there is no point in having data accurate to 1 part per million if 1% is the most the application can handle.
Models, similar to data, have to be suitable for the purpose. There is no point having a high fidelity rigorous model with 100,000 equations when you are doing long, medium or short range planning; it's a totally different story if you want to use the model for closed loop, real-time optimization. Expert Systems and Artificial Intelligence (ES/AI)
Cheap computing power made it possible to use (ES/AI) for: I. Establishing knowledge-based rules to handle difficult problems, like data filtering, or scheduling conflict resolution II. Using neural networks for non-linear model development from historical data, and pattern recognition for early problem detection III. Genetic algorithms for non-linear mixed integer optimization problems, thus avoiding a host of mathematical problem handling issues IV. Stochastic planning models using statistical methods to handle uncertainty (Monte Carlo simulations together with Genetic Algorithm nonlinear optimizers)
What is needed are knowledgeable people working with vendors to widely incorporate these advances in commercial tools.
The Web and the Internet
This new computer communication revolution holds great promise in increasing the cooperation between various users of computer applications, and applications results, at a very low cost, independent of the location of the users.
It relies on the use of Internet browsers with built-in "applets" to access BOTH applications and data from "web-enabled" servers around the world.
"E-commerce" portals are Internet-based electronic exchanges and auction houses, making instantly available to buyers and sellers information on what's available for sale, assay and pricing information, etc. This opens up a completely new real-time trading opportunities for buying and selling feedstocks and products online. Of course, this realtime trading will make planner and scheduler's life more hectic and stressful by introducing the additional time pressures associated with opportunistic trades.
The same browser technology allows planners and coordinators to access, no matter where they are or in what time zone, the most up to date refinery information, and do their work remotely. The results stored in a central data base will be available to all the other authorized users. There are in fact, commercial application service providers (ASP), where one can "rent" a planning application package; the ASP is responsible for supporting the application and applying the latest and greatest updates and improvements, while the user benefits from all the improvements at the lowest possible cost. The large bandwidth makes the use of application the same as from a local PC desktop. There are other ASP's that specialize in providing data base servers and process data historians for refineries to store many years of data. All these capabilities are supported by large communication bandwidths, security firewalls, and multiple redundancy.
And these are just a couple of recent examples…the beginning of new way of efficient working, "distance engineering".
The Bottom Line:
The SCM proved in practice that benefits of $0.50 to $1.00 /bbl potential refinery margin increase are attainable by refineries applying an integrated approach. However, experienced people are the key to success: software and computers, no matter how
powerful and glitzy, are not substitutes for understanding the refining business and processes.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue listening from where you left off, or restart the preview.