View metadata, citation and similar papers at core.ac.uk brought to you by CORE METHODS ARTICLE published: 21provided October by 2014 Frontiers - Publisher Connector ENERGY RESEARCH doi: 10.3389/fenrg.2014.00044 Sharing data for production scheduling using the ISA-95 standard Iiro Harjunkoski* and Reinhard Bauer ABB Corporate Research, Industrial Software and Applications, Ladenburg, Germany Edited by: In the development and deployment of production scheduling solutions, one major chal- Gurkan Sin, Technical University of lenge is to establish efficient information sharing with industrial production management Denmark, Denmark systems. Information comprising production orders to be scheduled, processing plant Reviewed by: Pradeep Suresh, The Dow Chemical structure, product recipes, available equipment, and other resources are necessary for Company, USA producing a realistic short-term production plan. Currently, a widely accepted standard for Franjo Cecelja, University of Surrey, information sharing is missing. This often leads to the implementation of costly custom- UK tailored interfaces, or in the worst case the scheduling solution will be abandoned. Addition- Arun Giridhar, Engineering Research Center for Structured Organic ally, it becomes difficult to easily compare different methods on various problem instances, Particulate Systems, USA which complicates the re-use of existing scheduling solutions. In order to overcome these *Correspondence: hurdles, a platform-independent and holistic approach is needed. Nevertheless, it is difficult Iiro Harjunkoski, ABB Corporate for any new solution to gain wide acceptance within industry as new standards are often Research, Industrial Software and refused by companies already using a different established interface. From an acceptance Applications, Wallstadter Str. 59, Ladenburg 68526, Germany point of view, the ISA-95 standard could act as a neutral data-exchange platform. In this e-mail: [email protected] paper, we assess if this already widespread standard is simple, yet powerful enough to act as the desired holistic data exchange for scheduling solutions. Keywords: scheduling, industrial standards, collaboration, integration, re-usability INTRODUCTION production process, can be easily plugged into the overall solution The scope and complexity of scheduling problems are con- landscape, this will greatly help to test and deploy new solutions tinuously increasing due to the more established and efficient of real problems in the process industries more efficiently. From solution technologies and increased popularity and need to a technical perspective, a platform-independent, holistic integra- improve the profitability of industrial production. Trends such tion approach that requires configuration through parameters as enterprise-wide optimization (EWO) (Grossmann, 2005) foster instead of customization by programing is desired to support more integrated problems across business functions. Most typi- easy integration in the production environment. If also applicable cal examples connect supply-chain management and production to a scheduling solution, this would offer a broad applicability, control aspects with the planning and scheduling (P&S) function. capturing most scheduling problems that occur in practice. This results in larger problem instances, more complex data, and A number of scientific journal papers have discussed the iden- mathematical models making the solution landscape in general tified gap between industry and academia (Henning, 2009; Har- more complex (see for instance Chu and You, 2012; Engell and junkoski, 2012) and the hurdles to deploy theoretical results in Harjunkoski, 2012; Zhuge and Ierapetritou, 2012). Traditionally, practice. The target is to define a standard way to plug-in a sched- the production targets of a plant have been defined by enterprise uling solution into a production management environment in resource planning (ERP) systems, which also in some cases have order to ensure also the practical usability of the best research determined at least a rough production schedule for the plant results. In another research line, Muñoz et al. (2013) presents floor. The feasibility of a schedule is typically ensured by a man- an ontological framework to support a sequential optimization ufacturing execution system (MES) or collaborative production of problems from different decision levels by following a hierar- management (CPM) suite. These systems ensure that the pro- chical approach, where also ANSI/ISA-88 and -95 standards are duction targets can actually be realized, include detailed planning reflected. Muñoz et al. (2014) further uses the approach to gener- with finite resources, and take process disturbances into account. ate mathematical models in the domain of enterprise and chemical Having real-time access to the plant floor, they further coordinate processes. Framinan and Ruiz (2012) addresses successful strate- the detailed resources and host the main activities to realize the gies for the development and deployment of scheduling solutions production, such as dispatching, tracking, visualization, and data and in Harjunkoski et al. (2014) many deployment aspects and collection and -analysis. lessons learned from the industry are discussed. In this regard, it is very important to efficiently transfer and Here, we do not focus on analyzing the decision components of share the data and the information collected between a pro- a hierarchy in the supply chain but assume that the decisions taken duction management system environment, containing a multi- on the scheduling level are well defined. The main challenge is how tude of functions, and the scheduling solution. If the scheduling to enable the data flow between the scheduling and surrounding solution, responsible for the tactical short-term planning of the components. We believe that a standardized approach which is www.frontiersin.org October 2014 | Volume 2 | Article 44 | 1 Harjunkoski and Bauer Sharing data for production scheduling efficient because it is widely understood has been jointly devel- itself to production scheduling but covers many other business oped and is a neutral component, which increases the likelihood functions. This also means that there is no unique way to use the of a broader acceptance. The ANSI/ISA-95 standard (ANSI/ISA- standard. 95.00.03-2005, 2005) has been created by a neutral standardization Simplistically said, the ISA-95 standard is designed for committee, consisting of company and university members. ISA top-down information flow and it is rather straightforward to stands for International Society of Automation and considers all transmit/dispatch the main information of an already computed levels of process automation from the controller (device) level to schedule to the shop floor using the standard. On the other hand, the long-term planning in ERP systems. ANSI is the short name for because of the top-down structure it is not fully straightforward American National Standards Institute, which launched the devel- how to provide all the information necessary to perform the sched- opment of the ISA-95 standard in the 1990s, which has thereafter uling actions. Some decisions are assumed to have been taken been approved as an international standard (known as IEC 62264). on a higher level and information such as sequence-dependent 1 The official definition is:“ISA-95 is the international standard for change-over times, release, and due dates of production orders are the integration of enterprise and control systems. ISA-95 con- not directly covered by the core data elements (Harjunkoski et al., sists of models and terminology that can be used to determine 2013). Naturally, scheduling needs are often very case-specific and which information has to be exchanged between systems for sales, this fact may hit the boundaries of a standard. Here, we try to finance and logistics, and systems for production, maintenance, adopt the standard in a way that fulfills the requirements for the and quality.” Thus, the standard is very broad and does not limit most common scheduling problems. Perhaps the most compact systems view of the ISA-95 standard is provided by the Purdue Reference Model, shown in Figure 1. 1http://www.isa-95.com/ The dashed line indicates the border between ERP and control nd Order al a teri ma t rs Processing ng ceip e mi re Ord nco rgy n (1.0) Product Cost Product I ene tio uc od ity Pr il Accounting Shipping Admin ab ail Av (8.0) (9.0) Finished Goods Inventory Production t s o Scheduling Pack Out Schedule c d P n F m r o d t i r n o i n n e s d t s i a t (2.0) P u c a o s T l r c h y n o t u e C ia g e d io d e g r u n c s d r c o n n e n e e m S ti C o t o a io v G n e ch n a Pr t ti L a e ir e F p m c o d r a r c o o ship Confirm to u b u ship to Release M ul m i o je d q e P lit d b s e l y rf o r an e r O W Short Term Material P P a iv and Energy Requirements e Material and Production Process Data r Product Energy Control Control In Process Waiver Inventory Control S Request (4.0) Material and Energy (3.0) ta (7.0) nd I a C n Inventory rd M c s E o P R a R a O o e n n n M r m s q s d e e t f r o P u e t i d d i C s r a d r r q r i r u t r s o e g m e n M u c me s u i i a t e a r g n e o ul i y P c Q s m a a n d s u r l t t A s t s s e e O t e r P an i R D r n i q o a n o d a m n r e r a d n e t Re t s a n o o d t d a e e u d d u R l e h A n F t S n n c s r t P u c t t e a Q e e c s e r e a e n c o t c n M d c c n a R n d e n e a b a e s d P d Quality a Procurement n n T s s n r e c P e p o t e a T t k c r o n c e o n Assurance i e (5.0) i h n c c a s a n h e s s S n s e M M i ta M c I i s (6.0) a s n c n P i a ur nte f a K C dar l o c n l n R u R h a r F e d e as n m o q sto s q e ce e w u a ui O a e d ir me n r t n s e e rd i d H t d m o a m r e b o s n en r n t e Maintenance a w s e nts t c s R c e k e u c m d Management q o re o r i u r e P u Marketing s P q (10.0) t e R & Sales Research Development and Engineering FIGURE 1 |The Purdue reference model for ISA-95 (source: ANSI/ISA-95).