Deriving from the most suitable methods aimed at the improvement of quality and increased productivity linked to the manufacturing processes in the automotive industries, the most sort after driving aspect that is greatly implemented is the Statistical Process Control (SPC)  and Process Automation. The two mentioned aspects are continuously engulfing quite a big portion of the automotive industry’s funds for investment. In order for related investments in this field to work out and bring about a flow of profit, process automation’s ability to drive an increased level of quality must be integrated with SPC. Quality is quite a challenging issue considering it is not a static factor. The everyday hanging customer preferences and ideas linked up with their needs will always affect or rather cause a deflection in what was considered of good quality previously . This begs the need to always have good innovations for any organization or business that wants to have a competitive advantage as far as customer needs are concerned. One of the most implemented methods to handle the above is SPS which acts as a ruthless tool when it comes to enabling improvements of quality and elimination of waste. This paper will review the different methods implemented for the integration of SPC and the relative automation.
Keywords: Statistical Process Control, Quality control, Quality improvement, automotive industry.
The aspects of control and quality improvements have become the leading factors when developing business strategies in the current technological world. Fabricators, distributors, financial and transport sectors and the healthcare sectors also consider this a very viable and important aspect in attaining effective flow of the business objectives and goals.
The current trend in the automotive  sector shows an increase in globalization coupled with a rise in number of the automotive production capability which creates an intensified competition between the different automotive firms. Quality in the automotive industry is best expressed through customer satisfaction relative to the products and services offered by a particular firm. There is a rise in the demand for quality delivery from the automotive industry. The way forward sort to handle the aspect of quality and the increasing demand is Statistical Process Control (SPC). SPC is the use of statistical methods implemented towards the control and monitoring of processes in order to maintain their operations at their fullest potential. This will help in ensuring that there is production of conforming products with the least waste possible. SPS is also implemented as a strategy aimed at the improvement of capability by reducing variability of products, deliveries, processes, components, perspective and machinery.
To all matters linked to manufacturing and seek to review quality as a factor, SPC is currently the most widely implemented and used method and has generally proven to improve operational and financial benefits. The correct implementation of SPC has been seen to lead to development of decision based on facts, aid in the expansion of the perception about quality enhancements at all levels, implementation of a systematic technique in resolving different problems and generally lead to effective gathering of productive information and experience causing improvements within a firm.
BACKGROUND AND OBJECTIVES
From previously conducted surveys and studies, there is substantial amount of data and findings that are carried out in relation to SPC. It is however ideal that more is required in exploring the utilization of SPC methods and techniques in the different areas of production, manufacturing and design.
In today’s world, many scholars believe that the modern statistical quality control could have started in the United States of America in the 1920s. In this period, in the Bell Telephone company, Walter Shewhart came up with the very first design of a control chart and had it in the application of monitoring and control. Control charts are an aspect that was introduced in the year 1924. Control charts enable the management to have a view of the processes instead of just the products. From Shewhart’s study and in his first book about quality control, he talks about the concept of statistical control and goes ahead to describe the criteria of the three sigma limits.
From Shewhart’s study, he wanted to place a clear cut line between “in-control” and “out of control.” Using an example of having to record some measurements regularly of a particular process, these measurements may be anything involved in important production processes. A number of lines or graphs can be developed to draw an illustration of the intended data in a statistical manner. If a process is then explained to be stable, then the concept behind it is that it is statistically in control. If the general layout of the process keeps shifting over time, then the process is considered out of control. As a scientist, Shewhart then knew that there was always variation gap that needed to be measured every time to keep track of the effectiveness of the entire process. The variation could be huge or overly minimized or between the mentioned extremes but fact remained that there will always be an occurrence in variation.
The next major improvement followed in the 1950s, crafted by W.E Deming, in the post-war Japan. His improvement was influenced by his predecessor’s earlier work. By the time we got to the 1970s, Japan had then developed into a major economic harbor. A control chart develops a prediction that, when there are no assignable causes, the process will proceed in operation as random system and give out the current level of quality in the future. If the attained level of quality is not satisfactory, then certain amendments to the process should be implemented.
One of the advantages of the control charts is that the process itself is given the capability to determine its control limits. The process is able to display clearly exactly what can and what cannot be expected. Thereafter the control limits can then be calculated automatically from the data that is given out by the process. From this, it is clear that control limits cannot be assigned but rather they are brought about by the general operations of the process itself. Control limits refer to concepts relative to the operation of a process but should not be mistaken as corporate goals or the preferences of consumers.
One of the keys to statistical process control charts is variation. The expressed diversity of variation in a particular process acts as an indication for whether the process is functioning as it should or the opposite of what is expected. At the point when the variation between certain points is found to be huge enough for the process to be considered out of control, then the variation is explained to be because of non-natural causes. The non-natural causes are what are also referred to as assignable causes at most times. This helps bring out the deeper meaning, as earlier mentioned, that a process is considered in control at the point when the results of that process can be predicted accurately. While the opposite is when there is hardly a single way of predicting whether the results of the process will attain the expected targets. In a real-world view point, a process that is out of control can be likened to driving your vehicle where the brakes could be working or not but you have no actual way of knowing this. Ones a process is considered out of control, the following step will be to look for the causes of the seen output variations. If these causes are arrived at, they are referred to as negative when there are multiple sections in the process that have defects and efforts are immediately put in place to develop the reasons for the defects and have implementation of attempts aimed at eliminating the defects. On the other hand, the “out-of-control” cause is considered positive; meaning there is no multiple defects on the process’s sections. At this point the assignable cause is then sort out, and direct attempts put in place to implement it every time. Successful implementation will see that the process grows in its general potential and fulfillment of the desired output.
Processes that bring about data exhibiting natural or common-cause variation will entirely form a distribution graph that portrays a bell-curve. These kinds of processes, the control charts are capable of providing useful and implementable help. However if the data in unevenly distributed and fails to portray a bell-curve, then the process is considered to be already out of control and is therefore unpredictable. In this scenario, ways should be crafted aimed at bringing the process to control. It is needless to say that every process, by its definition, should be able to bring out some regularity. Therefore there should be reorganization of the different data into co-relational sub-groups and having each group being in control.
Considering the general aim of every organization that seeks to improve quality, the first step would be the identification of the processes that need to be improved. Getting such processes can be done through the use of survey, focus groups or by simply having client or customer feedbacks on their general view of the firm’s products. Ones the problematic areas are selected, there is in-depth statistical analysis from the people involved with the processes and have the targeted issues selected alongside a list of their possible causes.
A process is considered to be in statistical control if the distribution of probability that represents the quality traits is even over a period of time. The occurrence of any particular change over the selected time, then the process is considered out of control. There is implementation of either variable control charts or attribute control charts, which is dependent on the type of data. Variable control charts are implemented with an aim of controlling processes of the different parameters of a product. These parameters are measured on continuous measurement scale such as inches, miles, pounds and much more. For processes in the manufacturing sector, the most commonly implemented control charts are the mean and variance that should be monitored closely to maintain high yield in quality. The characteristics possessed by control charts aid in the classification of processes as either good or bad, accept or rejects and much more.
The analysis of process capability has become a vital and well-defined tool in the implementation of SPC towards the continued improvement of quality and increased productivity. The capability of a particular process to meet the needs is assessed by the calculation of one or more capability indices. One of the most easily understandable of these is the proportion of the items that are produced within the process’s specifications. An important trait of the production process as a whole is the un-cut stability that can be attained by regulation.
In the above context, this paper will aim at studying the different relations in stability of processes and put them in context of the different calculations and probability equations involved. The paper will also seek to study and highlight any capable improvements to the application of Statistical Process control technique and laws in the enhancement of effective production in the automotive industry.
The methodology section is a description of the general outline structure use in the research paper and it lays emphasis on a more practical view-point to bring an in-depth understanding of the involved aspects.
A case study methodology is also used to develop vital contribution to the scientific development considering the fact that the research involved in context is complex and requires substantial theoretical bases, expertise and adequate provision of time as a resource. It is needless to mention that there might be few cases that may go undetected in the study due to its vast nature and the large proportion in which it covers.
Production Equipment Capability
It is very vital that the capability of the production equipment is determined before the initiation of any serial production. The intention of the mentioned to develop a conclusion that the production will undergo its operations within familiar patterns and that the machines are capable enough to handle the production within the required tolerance. The mentioned capability of a machine is represented by the use of indices of capability i.e. Cm and Cmk, where the former is used to indicate variance and the latter to indicate process position. Among the very basic, but vital, needs for establishing equipments production capability is the specification of the stability of measured values. Pristavka (2007) mentions that besides the effective measuring device, the production conditions and the general process in statistical control are the basic parameters put in place to examine the capability of a process. Before any serial production, it is important to have at least a sample of fifty products from which a subgroup of a minimum of n = 5 are developed. In order to have the stability test done, the mean and standard deviation of each of the subgroups are calculated.
The mean of the subgroup is arrived at by using the formula:
Serial number is i
Measured value serial number- j
Quantity of subgroups- k
Size of the subgroup- n
The value measured in i-th subgroup for i=1, 2,…k and for j= 1, 2,…n – Xij
The standard deviation of the subgroups is then calculated as follows:
Calculation of machine capability index
In order to proceed and calculate the machine capability index, we need to develop the standard deviation first:
The production equipment is considered capable when Cm ≥ 1.66 and Cmk≥ 1.67 as defined by Michalek (2006). Once the production equipment is considered to be capable, the control charts are then developed through measuring first then determining the process capability much later.
THE CASE STUDY
The study will be based on the rubber production of the automotive industry the parts. The company is located in northern India. The company is recognized as one of the largest companies in the manufacturing industry associated with the production of automotive rubber parts in the country. Due to the competitive position of the company and the fact that it possesses quite a large portion of the market, the company is known to offer its customers with the largest variety of sealing products for quite a number of application in the field of automotive and other related fields.
Shocker seals are among the leading components in the industry. The production of shocker seals needed a lot of attention due to the percentage rejection of more than 9.1%. The above rejection was the reason why the application of the SPC technique was required to enable the reduction of the percentage rejection. The following were classified as the main causes of the increased rejection of the shocker seals in the industry:
- Moulding: this is the first process in the production of shocker seals. From the study that was conducted by the industry, it was noticed that the various defects in the moulding were the factors responsible for the rejection. The defects that were noticed were presence of air traps, tearing, foreign matter, excess materials and dirty cavity among other reasons.
The recommendations put in place to have the defects of moulding were as follows:
- Approval of the presence of in-process inspection for every manufacturing process
- Creation of adequate vacuum
- Maintenance of proper environmental temperature
- The replacement of manual loading by the implementation of mechanized loading.
- To have the casting temperatures maintained between 1900o C to 2100o C.
- The performance of well crafted trimming
- Have the nozzle hole cleaned up properly
A sample size of (n) was considered in the study in calculation of the outer diameter of the shocker seals. The samples were taken in random and the observation before the carried out study may not be inclusive due to the nature of the vast topic being handled.
The target on the outer diameter of the shocker seals = 62 mm ± 0.10 mm (marked as the tolerance).
The upper and lower specification limits (USL) and (LSL):
- (USL) = 62.10 mm
- (LSL) = 59.90 mm
The process capability ratio can also be calculated as:
There was also the calculation of the process capability index:
In the current world, quality control has become a major factor and issue especially in the automotive industry. There are a lot of technological advancements in the automotive industry all aimed at meeting the overall demand from the global customers. There are a number of challenges that the investors in the industry face everyday among them being: the unending pressure to have the prices of their products as low as possible, the everyday growth of complex global supply chains and the rise in demand for new innovations not forgetting the reduction in time available to market.  Due to these challenges and the uncontrollable factors the need to effectively incorporate quality management in the industry needs urgent attention coupled with real-time visibility. This will be aimed at the overall reduction of operational costs, enable a platform to increase quality based on the everyday improvement procedures and also ensuring compliance of the certification. SPC is the unbeatable most advantageous technique for determining the quality level of any production process.
From Ishikawa’s study, he pointed out that control charts should be viewed as ways of seeing exactly what changes occurred in data over time including the impact felt from the different factors involved in each process. The use of the charts also kept in mind that the processes were bound to change over time and therefore came up with the best analysis of what to anticipate and the suitable measures expected to put in place due to these changes. The major question in many people’s minds is how to tell if an implemented program changes have had the expected effects on the general operations. The implementation control charts is what helps in determining the overall performance of a process and determine whether it is moving upwards significantly.
Using a case study that was done on an automotive industry whose sampling frequency was placed at five pieces per shift with only a single shift per day, there will be a single point recorded on the control chart every day. There are two forms of control charts that implemented for the existing processes. One is by the variables while the other is by the attributes.
The above is a control chart representing the mean and standard deviation.
After the inserting of the relevant data and having come up with the relevant graph shape, like the one above, it is then possible to derive various indices such as capability so as to have a better understanding of the results that are reflected by the graph. Analysis that is carried out on the normal probability plot of the control graph proves the authenticity of the data collected as it is assumed to be in a normal distribution. The Kolmogorov-Smirnov test is also a part of the work carried out. This is due to the normality that is required from the control charts. From the interpretation of the graph, it shows that the data is on a normal distribution.  On the other hand, other statistical techniques, in the case of variable control chart, example being the normality test, a much deeper understanding of the results can hence be made possible.
The above work helped in the deeper understanding of SPC in aspects such as equipment capability and how the various calculations can aid in the final plotting of a control chart. There is also knowledge acquired on the overall interpretation and testing of the data provided in the control chart so as to ensure evenness in the normality. By the use of control charts in the study above, it can allow for the identification of the different problems or short-comings in the different automotive industries. Through the use of the control chart model, in connection to the various calculations of probability and standard deviation, it is possible for an individual to link the theoretical part and implement it so as to increase the knowledge and improve learning process through a step by step application of the vast statistical process control. The study also touches, on a shallow basis, the automotive industry so as to prove the applicability of the SPC technique to the different companies and industries despite them having different processes, operational routines and different adaptation specifics.
From the analysis of results of the attribute control chart, it shows that attaining a production level of all the pieces required is quite impossible especially within the specified requirements. However through such information, the firm can then implement strategies in the production of the pieces in a better understood way that is correctly analyzed and implemented through correct actions in order to have a general improvement in its production.
The Statistical Process Control analysis can easily help in the general improvement of the effectiveness of manufacturing process which will also lead to a decrease in the quantity of defective products. A reduction in the number of defective products will help in saving quite a lot in terms of budget as there will be reduced amount spent on re-doing the spoilt or defective products. Generally it will also help in proper utilization of time as a resource. For each of the specified products, there will be a substantial saving in the amount of time and money that is spent.
An example of the implementation of SPC in the manufacture of shocker seals saw the improvement of the process capability to a level better that what was anticipated. The improvement came despite the rejection level of the process being as high as 9.1%. The high rejection level of the shocker seals was a major concern for the company. However, with the implementation of required suggestions and recommendations from the statistical control process technique, as already mentioned, the company experienced quite an improvement it its process capability. It is therefore seen that the determination of a process’s capability in connection to developing the control charts is a powerful technique is SPC that aids in bringing out the most in a process, in terms of productivity, while still maintaining the required quality or even enhancing the quality of the process’s products.
|||M., Yusuff, R. M., Hong, T. S. and Ismail, M. Y. B Abdolshah, “Loss-based process capability indices: a review, International Journal Of Productivity and Quality Management,” vol. Volume 7, no. 1, pp. 1-21, 2011.|
|||Dr. D. R. Prajapati, “International Journal of Emerging Technology and Advanced Engineering,” Implementation of SPC Techniques in Automotive Industry: A Case Study, vol. Volume 2, no. Issue 3, March 2012.|
|||R. Vrabič, P. Butala e A. Sluga G. Škulj, “Statistical Process Control as a Service,” An Industrial Case Study, vol. 7, pp. 400-410, 2013.|
|||LobonĠ, C. V., Kifor,C., Oprean,O., Suciu L., “TEHNOMUS – New Technologies and Products in Machine Manufacturing Technologies,” IMPLEMENTING A SPC INTEGRATED SYSTEM TO IMPROVE MANUFACTURING PROCESSES IN AUTOMOTIVE INDUSTRY, 2011.|
|||American Society for Quality. (2015) Statistical Process Control. [Online]. http://asq.org/learn-about-quality/statistical-process-control/overview/overview.html|
|||Roger Edgell. Statistical Process Control – SPC. [Online]. http://www.managers-net.com/statistical_process_control.html|
|||P.-H. Lee, ““Joint statistical design of and s charts with combined double sampling and variable sampling interval,”,” European Journal of Operational Research, vol. 225, no. 2, pp. 285-297, 2013.|
|||H. Luo e X. Zhang Z. Wu, ““Optimal np control chart with curtailment,” ” European Journal of Operational Research, vol. 174, no. 3, pp. 1723-1741, 2006.|
|||Ph.D. & Robert F. Hart, Ph.D. Marilyn K. Hart, “Introduction to STATISTICAL PROCESS CONTROL TECHNIQUES,” Statit Software, Inc., Corvallis, Oregon, 2007.|
|||C. Perry, ““Processes of a case study methodology for postgraduate research in marketing,” ” European Journal of Marketing, vol. 32, no. 9-10, pp. 785-802, 1998.|
|||G. Barbato, E. M. Barini, G. Genta e R. Levi F. Aggogeri, ““Measurement uncertainty assessment of Coordinate Measuring Machines by simulation and planned experimentation,” ,” CIRP Journal of Manufacturing Science and Technology, vol. 4, no. 1, pp. 51-56, 2011.|
|||T. C. and Gan, F. F. Chang, “Shewhart Chart for Monitoring the Variance Components,” Journal of Quality Technology, vol. 36, no. 3, pp. 293 – 308, 2004.|
|||C. S. Nelson, “The Shewhart Control Charts,” Test for special causes, Journal of Quality Technology, vol. 16, pp. 237-239, 1984.|