Improving Railroad Classification Yard Performance Through Bottleneck Management Methods

249 words Body: 5,258 words + 5 Figures (1,250 words) + 2 Tables (500 words) = 7,008 words ABSTRACT Because railroad classification yards can be considered production systems, insight into theBecause railroad classification yards can be considered production systems, insight into the dynamics of a yard system can be gained by adapting production management tools that have led to significant performance improvement in manufacturing. This work focused on improving yard performance by utilizing the concepts of factory physics, Theory of Constraints (TOC) and tools from Lean Manufacturing. The most important manufacturing process analog to improving yard capacity is the bottleneck. In a production system the bottleneck is the process that limits its throughput. As such, the processing rate of the bottleneck sets the rate for the entire system. Improving the performance of the bottleneck is the best way to improve the performance of the entire terminal process. The train assembly (pull-down) process has been identified as the bottleneck in a majority of classification yards. The potential capacity improvement of several bottleneck management alternatives is discussed. One of the principal findings of this work is that the humping process should be subordinate to the pull-down process because the latter is the principal bottleneck in many yards. The hump should be managed and operated so that it provides the bottleneck exactly what it needs when it needs it. The quality of sorting during hump operation directly affects the performance of the pull-down process. A metric for measuring how well during cars in the classification yard have been sorted has been developed and its relationship to yard volume established. Methods for implementing this metric in a classification yard are also discussed. INTRODUCTION While shipping by railroad is usually less expensive than trucking, the lower level of service reliability can produce higher total logistical costs for shippers and receivers. Higher variability in shipment arrival times results in additional inventory having to be carried in order for a railroad customer to maintain a fixed level of customer service (1). Previous studies have established the need for the railroad industry to improve service reliability in order to meet the increasing logistical demands of shippers (2). These same studies have named the classification yard as a key determinant in the service reliability of general manifest (or carload) freight. The trade off between high cost efficiency and reduced service quality is inherent to carload railroad operations. For railroads to continue to grow their business, they must work to overcome the tradeoff. A majority of total trip cycle time is spent in yards. Two major North American railroads have reported that 59% (3) and 64% (Figure 1) of railcar transit time is spent in yards. “This suggests that the reliability of car movements can be improved by reducing the time spent in those activities or by making them more reliable” (4). The transition to scheduled operations by all of North America’s Class I railroads has heightened the interaction between yard performance and service reliability (5, 6) because “efficient high-throughput classification yards are vital to scheduled railroading” (7). Within a classification yard, connections are made by classifying cars from inbound trains into blocks that will be assembled into outbound trains. The objective is to sort cars and reliably connect them to the earliest possible candidate outbound train, while minimizing cost (Barker unpublished date). Kraft has extensively studied the connection reliability problem as it relates to dynamic car scheduling (8) and has developed a hump sequencing algorithm (9), a priority-based classification system (5) and a dynamic block to track assignment scheme with the goal of ensuring connections (6). Kraft raises the issue of inadequate terminal capacity as a barrier to improved service reliability (9). However, the availability of capital and the physical capability to expand some yards may be constrained. Therefore, in addition to considering infrastructure expansion, railroads must also determine how to harness as much capacity from extant infrastructure as possible. This creates the need for new management and operational methods that will increase the capacity of existing facilities. Manufacturers face a similar need and this presents the opportunity for the use of selected techniques from production management. Yard capacity can be improved an estimated 15-30% (3) by adapting an approach known as “Lean Railroading” (10) with emphasis on the bottleneck management component. The pull-down process is identified as the most common bottleneck in hump yards. The macroscopic evaluation method from Wong et al. (11) is enhanced with two additional equations and used to evaluate several improvement alternatives using Bensenville Yard (CPR) near Chicago as the example. To aid in implementing one of the more promising alternatives, a Quality of Sort metric is developed to better manage and understand the interaction between the pull-down process and its immediate upstream process (the hump). LEAN RAILROADING Because classification yards can be considered production systems (12), their performance can be improved by adapting an integrated approach comprised of three proven production management techniques: Lean, Theory of Constraints (TOC) and Statistical Process Control (SPC or “six sigma”). Known as “Lean Railroading” (10), several railroads and railroad suppliers, including Canadian Pacific (CPR), Union Pacific (UP), BNSF, Norfolk Southern (NS), the Belt Railway of Chicago and GE Yard Solutions, are actively applying all or parts of this approach to improving yard performance. In addition, many of the “precision railroading” principles that CN has used to improve their operating performance can also be considered lean. The first step in any lean program is to define value for the ultimate customer and then work to increase value by eliminating waste in the system. Waste is defined as any step or process in a production system that, from the standpoint of the customer, does not add value to the product (13). Waste can be classified into two types: direct waste and variability (14). Direct waste is most easily described as poor railroading practices such as unnecessary moves, mistakes that require an operation to be repeated, lax track maintenance and unsafe operations to name a few. Focusing on these is important, but the goal of eliminating direct waste is as old as the railroad itself. Variability is a fundamentally different source of waste. Hopp & Spearman (1) state, as a law of manufacturing, that, “Increasing variability always degrades the performance of a production system.” Railroad yards are no different: they are subject to both internal (i.e. outages, rework, sorting, etc.) and external (i.e. arrival times, weather, traffic volume, etc.) sources of variability. Another law of manufacturing from factory physics is “Variability in a production system will be buffered by some combination of inventory, capacity and time” (1). In a classification yard, an inventory buffer is seen in the form of railcars sitting in the arrival, classification or departure yards. A capacity buffer takes the form of a process throughput greater than the process demand. A time buffer is the extra time built into each car’s trip plan in order to ensure that the connection will be made and is seen in the terminal dwell. Spearman (14) states, “In many ways, the ‘waste’ discussed in Lean is the ‘buffer’ of Factory Physics. However, this is not always the case. If external variability creates the need for a buffer, is it waste?” Providing different service levels increases variability, but would the railroad be better off if it were to only offer one service level? “The point is that while not all variability is waste, all variability will lead to a buffer which indicates that logistical (but not necessarily financial) performance has suffered” (14). As long as the increase to railroad revenue is greater than the increase to operating costs, profits will increase. Therefore, it becomes the task of yard management to reduce internal variability and the task of network management to manage the external variability so that the bad sources (like arrival variability) are reduced and the good sources (like service level differentiation) increase profit. Implementing Lean Railroading With the advent of scheduled railroading, railroads have already taken an important first step in creating an environment that Lean Railroading can succeed in by reducing external variability for the yard. The implementation steps are: 0. Eliminate direct waste Take a fresh look at the yard as a system by drawing a Value Stream Map (VSM) and try to eliminate obvious sources of waste. 1. Swap buffers Decrease the time buffer (dwell time) by reducing the idle time between processes. This is synonymous with enabling continuous flow. Increase the capacity buffer by focusing on improving the performance of the bottleneck. 2. Reduce variability – a. Address problems in sorting, rework, car damage, down time and setups (apply SPC/”six sigma”) b. Implement standardized work plans c. Work with network management to increase on-time arrival of inbound trains d. Level the production schedule in the yard and set the network operating plan 3. Continuous improvement – “Once variability is significantly reduced, we can reduce the capacity buffer while continuing to identify and eliminate variability. Only at this point do we begin to make real gains in productivity. If we do not reduce variability, we will not be able to reduce the capacity buffer without hurting customer responsiveness. The result is a system that continues to improve over time” (14). The Theoretical Importance of the Bottleneck In order to decrease the time buffer, without