Attribution-NonCommercial-NoDerivs 2.0 KOREA You are free to : Share — copy and redistribute the material in any medium or format Under the follwing terms : Attribution — You must give appropriate credit, provide a link to the license, and indicate if changes were made. You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use. NonCommercial — You may not use the material for commercial purposes. NoDerivatives — If you remix, transform, or build upon the material, you may not distribute the modified material.
You do not have to comply with the license for elements of the material in the public domain or where your use is permitted by an applicable exception or limitation. This is a human-readable summary of (and not a substitute for) the license. Disclaimer Master's Thesis Process Mining-driven Performance Analysis in Manufacturing Process: Cost and Quality Perspective Tu Thi Bich Hong Department of Management Engineering Graduate School of UNIST 2016 i Process Mining-driven Performance Analysis in Manufacturing Process: Cost and Quality Perspective Tu Thi Bich Hong Department of Management Engineering Graduate School of UNIST ii Abstract The dynamics of globalization and high expectation of customers make manufacturing enterprise move towards three primarily competitive factors, namely, time, cost, and quality. The desire for continuous performance improvement of manufacturing processes is as old as manufacturing itself.
However, the latest industrial revolution Industry 4.0 and the digital revolution have opened up new avenues, intertwined information systems and the operation processes. As a result, enterprises face a challenge in extracting value from a massive amount of events recorded by today‟s information systems. Process mining is well recognized as a valuable tool for observing and diagnosing inefficiencies in business processes based on event data. It turns out that process mining is a viable solution to this challenge.
Nevertheless, significantly less attention has been paid on investigating cost and quality perspective in process mining. In these respects, this thesis suggests a framework for performance analysis in manufacturing processes based on process mining. The proposed approach focuse on cost and quality perspective. Specifically, the contributions of this thesis are in four-fold (i) to suggest a method to extend event log of manufacturing process with manufacturing information, i.
cost, quality; (ii) to analyze manufacturing information, i. cost, quality with process model; (iii) to utilize various existing process mining techniques and develop new approaches to analyze and predict manufacturing cost; (iv) and to enable quality report in manufacturing process. Keywords: Process Mining, Manufacturing Process, Performance Analysis, Manufacturing Cost, Quality-related KPIs. Application of process mining in manufacturing.
Cost and quality mining in process mining. Cost-related KPIs. Activity-Based Costing (ABC). Direct material cost.
Direct labor cost. Quality-related KPIs. First pass yield. Performance Analysis in Manufacturing Process.
Extension of event log with manufacturing information. Extension of event log with cost-related KPIs. Cost database model. Event log-extended manufacturing cost.
Extension of event log with quality-related KPIs. Quality-aware Event log. Computation of quality-related KPIs index. Manufacturing information analysis with process model.
Visualization of cost breakdown. Resource utilization cost. Process model-enhanced manufacturing information. Visualization of cost breakdown.
Resource utilization cost. Conclusions & Future Works. Summary of contributions. 50 vii List of Figures Figure 1.
Example of first pass yield caculation. A methodology for performance analysis in manufacturing process. Cost database model for cost model extraction. Event log-extended cost of manufacturing process.
An example of cost analysis with process model. Cumulative cost showing per activity. Remaining cost showing per activity. System architecture of performance analysis.
A system architecture for manufacturing information analysis with process model. A screen shot of cost analysis with process model. A screen shot of cost breakdown analysis. Selection of cost parameter in cost breakdown plug-in.
A screen shot of resource utilization cost plug-in. A screen shot of resource cost distribution per month plug-in. A Screenshot of Cost Prediction. Cost per Case.
47 viii List of Tables Table 1. Fraction of event log……………………………………………………………………… 19 ix List of Algorithms Algorithm 1. Extension of event log with manufacturing cost (LCM)…………………………………24 Algorithm 2. Computation of quality-related KPIs index (LQM)…………………………………….
27 x Chapter 1 Introduction This chapter introduces the motivation of the thesis. The problem statement and objectives are discussed in Section 1. The outline of the thesis is explained in Section 1. Motivation Manufacturing process concerns all efforts of an organization to add values to the inputs (raw materials, semi-finished products, know-how, etc.) and transform them into the outputs (finished goods) that meet an expectation of customers [1].
Nowadays, manufacturing environment is becoming more and more sophisticated [2]. Furthermore, increasing the level of competition in the global market has a significant impact on enterprises to stay competitive to survive. Leading manufacturers have seen the continuous improvement of process quality and the reduction of manufacturing cost as strategic weapons in the market to compete against peers. Quality is being known as one of the essential dimensions of business process performance [3-6].
Efficiency in managing of manufacturing cost is considered as a critical factor in obtaining competitive advantages. Also, making production decision is based on cost [7]. Therefore, getting a clear understanding of manufacturing process regarding cost and quality constitutes the most important prerequisite in this respect. The paradigms of information and sensor technology development have enabled large-scale data collection when monitoring manufacturing processes.
The knowledge learned from those data may gain a potential value for performance improvement of the process. However, the discovery of knowledge hidden in the data without proper tools is a challenge due to a significant amount of data. Commonly used approaches in manufacturing process performance analysis have a vast amount of literatures, for example statistical method [8], data mining techniques, i. association rule, sequence data mining [9, 10], and simulation method [11].
An enterprise has applied various quality control techniques to improve the process quality by reducing its unevenness. These include Six Sigma [12], Statistical Process Control (SPC) tools, Total Quality Management (TQM) [13], Business Process Re- engineering (BPR) [14, 15], and Business Process Improvement (BPI) [16]. However, related methods have not provided an adequate solution to this issue yet. For instance, statistical and data mining results are sometimes too complicated to understand [17].
Furthermore, simulation has the limitation that it takes too much time and cost to build acomplex manufacturing process model [17]. Process mining is well recognized as a valuable tool for analyzing an operational process and 1 tracking down its problems or inefficiencies using event logs [18]. These event logs can be obtained from Process Awareness Information Systems (PAISs) such as Manufacturing Execution System (MES), Enterprise Resource Planning (ERP), Customer Relationship Management (CRM), and so forth [19, 20]. Furthermore, process mining framework also allows us to discover models, analyze bottlenecks, conduct conformance checking and analyze process performance [21, 22].
Process mining has been applied in various fields such as finance, healthcare and port. In recent years, process mining has begun to be implemented in manufacturing areas. For instance, Son et al. [23] proposed a method to analyze an overall production process based process mining.
Rozinat et al. [24] introduced a way to improve the test processes in ASML using process mining. Park et al. [25] suggested an approach to measure the performance of the ship block manufacturing processes.
Hence, process mining is the most promising approach for performance analysis of manufacturing processes. Problem Statement In the manufacturing industry, it is essential that enterprises reduce their manufacturing cost and improve process quality continuously. Unfortunately, they encounter a challenge to analyze manufacturing cost due to the complexity of manufacturing processes. For instance, cost for a particular task or an individual resource is hard to be calculated.
The questions regarding consuming cost can be only answered at the end of the process. The quality of a process refers to the ability of that process to „produce and deliver quality products‟ [26], and to „conform to manufacturing specifications‟ [27]. Interest in improving the quality of a process has grown throughout the manufacturing community. Unfortunately, one of the difficulties in improving the quality of a process is figuring out where to start and deciding whether the implemented changes are beneficial.
Therefore, quality aspects of a manufacturing process are often neglected or deferred due to the difficulty in measurement [28]. Considering these issues, there is a need of proper method allowing access to the cost and quality status of each stage or each activity in the process. Although process mining literature has discussed a wide range of its application to solve these issues, not all of them have adequately supported these perspectives. Firstly, although significantly amounts of research have been conducted in process mining, they have focused on time and resource perspective rather than the perspective of cost and quality.
Next, there exist few studies on cost mining in process mining [29]. However, these efforts have not considered manufacturing processes [30-32]. Furthermore, they have not adequately assisted cost mining of manufacturing process where cost relies on not only time but also production volume. Production volume means the total units of products 2 coming into a process.
Assume that there is no loss or omission of goods during the manufacturing process. Meanwhile total output equivalent to total input. The output may include good products and defect products. Lastly, the approach is taken by [23] enables a way to control the quality of manufacturing process through yield analysis by combining the input and output of each activity with event logs.
It has become an emerging topic in the field of quality engineering using process mining. Nevertheless, this thesis has not entirely proposed a systematic and holistic method towards the quality perspective of manufacturing process yet. These drawbacks necessitate the development of a proper method to handle the entire cost and process quality mining in the manufacturing sector. Objectives To overcome such limitations, this thesis suggests a framework for performance analysis in manufacturing processes based on process mining and focuses on cost and quality perspective.
The primary objectives of the study are as follows: 1) To extend an event log of manufacturing process with manufacturing information, i. cost, quality 2) To analyze manufacturing information, i. cost, quality with a process model 3) To utilize various process mining techniques and develop new techniques to analyze and predict manufacturingcost 4) To develop a method to generate quality report in manufacturing process 1. Outline The rest of this study is organized as follows.
Chapter 2 explains related works which are process mining, its application in the manufacturing sector, cost and quality mining based on process mining. An overview of performance analysis with cost-related KPIs and quality-related KPIs is presented in Chapter 3. Chapter 4 introduces a performance analysis in manufacturing process. An implementation is illustrated in Chapter 5.
Finally, this thesis ends with conclusions and possible future works in Chapter6. Firstly, in order to extend an event log with manufacturing cost, this study first formalizes and extracts manufacturing cost model. Furthermore, a cost database structure is presented to support the extraction cost model from various information sources. Later, this study formalizes event logs of manufacturing process.
With the respect to enhancing the event log with quality and quality model 3 with cost information is generated based on quality-related KPIs. Then, the event log is updated with quality components, and a quality index is directly computed in event log using the quality model. Secondly, the manufacturing information would be easier to be interpreted and more significant and precise for decision makers if they are associated with the corresponding elements of the process model. Therefore, a process model extension with manufacturing information such as cost and quality are presented.
Thirdly, various visualization methods are used to present analysis results, such as cost analysis with a process model, visualization of a cost breakdown, and structure resource utilization cost.