Ford Motor Company - Statistical Process Control - Essay Example

? Statistical Process Control Submitted to, Submitted By, of the Submitted on, [November 14th, Introduction: In early days, after the industrial revolution, the manufacturing arena enjoyed the flexibility of application of asymptotic approach regarding the raw resource consumption. Nowadays, the rising manufacturing costs and high resource constraints demand that the manufacturing decisions must be made after careful evaluation of ground facts and realities with proper data acquisition and analysis. Statistical Process Control (SPC) is a comprehensive conceptual and applicable technique that is devised to check, control and optimize the manufacturing process and the quality of the product. With the advent of software applications the SPC is now simpler to implement. Concept: To understand the concept behind the technique of Statistical Process Control, it is necessary to revise some basics of normal distribution and approximate confidence intervals. SPC specifically addresses the processes that exhibit natural variations. These variations are also called as common cause variations. These variations when plotted on a graph as data points represent a bell-shaped curve distribution which is known as Normal Distribution sometimes referred to as natural distribution. Some of the examples of normally distributed data are heights of people or weights of people. According to the same concept; any data which is bell shaped symmetrical in nature and has a tendency to follow the Normal Distribution usually approximately follows the following set of theorems: 1- That approximately 68% observations shall be covered within one standard deviation around its arithmetic mean. i.e. µ±? 2- That approximately 95% observations shall be covered within two standard deviations around its arithmetic mean. i.e. µ±2? 3- That approximately 99.98% observations shall be covered within three standard deviations around its arithmetic mean. i.e. µ±3? So, if we can illustrate the same concept then we might be having the following situation: Normal Distribution with Six Sigma. (Source: The YouMOZ Blog) Statistical Process Control and Control Charts: To limit the variations related to a process in natural limits is the key to the phenomenon of Quality Control. Statistical Process Controls enables the users to frame the correct and incorrect variations in a process. Based on the above mentioned concepts; we can create a process monitoring system which can assure us to observe the whole process running under two tolerance limits namely upper control limit and lower control limit. These limits are obtained by the same concept of approximate confidence intervals. So after setting the limits (which are also termed as pre-alarm) we can run the process and monitor each observation through a Process Control Chart. Therefore we can set the following definition of SPC: Statistical Process Control is a process that uses statistical concepts based Control Charts to analyze and monitor the behaviors of processes by assessing the variations in them. The main strength of Statistical Process Control lies within its Control Charts which furnish an easy and simple way to conclude whether the variations in a process are in the controlled limits or not. These charts are simple time based graphical representations of data that is related to a process. On an x-y axis graph, the time is generally represented on x-axis. The data points plotted are usually the individual readings, averages of subgroups or values representing ranges. There are additional horizontal lines used to represent the control limits. Multiple types of control charts are used to serve their respective purposes. Control charts are used not only to monitor the performance of a process but they can also help identify the factors that hamper the process’s peak performance. The output variance over time is considered out of control when it crosses the specific control limits that represent the normal levels of variations. The control limits are determined and set by the process itself. These limits are never a result of pre-meditated efforts to fulfill the manufacturer’s or customers’ desires. It is necessary to take this fact into account that the term under control does not actually mean that the process is meeting the specifications and fulfilling the pre-determined needs, it simply means that the process is consistent in its behavior. Inside the control limits there may exist, violation of standard rules that are set and documented in the AT&T Statistical Quality Control Handbook. In cases where a process gets out of control, like multiple bottles filled with very low volume of beverage as compared to the required/expected volume, the causes are addressed to rectify the flaws of production. After adjustments, process is continuously monitored further to see whether it exhibits the natural variations with no violation of rules or not? The control charts helps identify the capabilities of a process also. Through the observation of process behavior in controlled limits it is possible to figure out the capabilities of a process. This knowledge may further be used to adjust the specifications, requirements and goals to match the process capabilities or vice versa. EFFECTIVENESS OF STATISTICAL PROCESS CONTROL The effectiveness of the application of Statistical Process Control is well acclaimed all over the world. The historical facts about the FORD Motor Company case are always coded as an example in this regard. Dr. Walter Shewart originated concept of SPC at Bell Labs. His descendent Dr. W. Edwards Deming extended his work. After the World War II he took this technique to Japan and trained Japanese manufacturers. The Japanese Industry adopted this concept whole heartedly. Resultantly the Japanese products enjoyed worldwide acceptance due to their high quality standards. As an acknowledgement to his efforts Dr. Deming is considered as ‘God of Quality’ throughout Japan. The acknowledgement of his works would be incomplete if his association with Ford is not mentioned in them. The Ford Motor Company used to introduce models of cars that were manufactured in both USA and Japan respectively. It was observed that the Japan made transmission versions of the cars were in much higher of a demand than the ones made in the US even though they had to wait for the Japan made models to arrive. Ford Engineers were unable to understand the reason behind the preference of the consumers even though both the transmissions were made on the same preferences. It was only after taking apart the two transmissions that the difference was realized. The US made transmissions had their parts within the specified control limits. Their measurement tolerance was 1/8 of an inch, while the Japanese made transmission parts were manufactured with 1/16 of inch level of tolerance which eventually was closer to the required measurements. This made the eventual running of the Japanese made cars much smoother than their American counterparts. It was not until the association of the Ford Engineers with Deming that the engineers had understood this. Deming inculcated that how very important it was to incorporate quality in manufacturing. How it could be incorporated had many variations. The use of Statistical Process Control to find out and implement the quality coherent levels was one of the means by which quality could be incorporated in production. This eventually resulted in successful selling of the product as customers showed satisfaction in using products that were based on such principles. SPECIFIC SPC TOOLS AND PROCEDURES A number of tools need to be used for the preparation of the Statistical Process Control. A summary of the essential tools may be narrated as follows: Recognizing the Problem. Prioritizing Tasks by using Pareto Charts.. Analyzing Selected Problem by using the Cause and Effect Diagram. Organizing tasks in line with the flow of Execution with the help of flow charts. Analyzing each sub step by using Scatter Plots. Gathering Data by using Check Sheets, Histograms, and Probability Plots etc. Identification of the problems that need to be upgraded in terms of quality is the first step in SPC. Surveys and personal discussions can be used to highlight these problems. Brainstorming with the stakeholders related to the processing further helps in the identification of the causes as well. PRIORITIZE The selected problems are prioritized. The ones having the greatest effects are given the highest priority. They usually are the smallest in number but the highest trouble causing. Pareto Charts- These charts enable graphical representation of problems thus enabling successful prioritization. The principle under consideration in this methodology is: “20% of the sources cause 80% of the problem. ANALYZING SELECTED PROBLEM In order to analyze the selected problem its causes need to be evaluated. Cause and Effect Diagram- A tool used to evaluate the causes of a problem. Also known as the Fish Bone Diagram. The problem in put at the head of the diagram and the causes then made to branch out from a backbone that has been extended from the problem itself. FLOWCHART The process selected for quality improvement should be broken down into a series of logical steps and should be arranged in a flow. SCATTER PLOTS Signifies the correlation between two varying quantities certifying how one becomes the cause of the other. The degree of correlation between the plotted variables can also be calculated by plotting them in this manner. GATHERING OF DATA The subsequent step before initializing the actual quality modification procedure is the gathering of data. Check Sheets- Simple sheets that contain options and check boxes. The check boxes for the selected options are checked thus enabling data input. Well designed and accurate sheets enable accurate data entry. Histograms- The data collected, if plotted in the form of a histogram, enables detection of the form the data is attaining. This helps in evaluating whether the range of the data would enable production of what is required from it. Probability Plot- For the Control Charts to be effective, it is essential that the data plotted be arranged in a normal form. It is a graph that represents the cumulative relative frequencies of data and signifies what certain data is capable of if it is not distributed normally. There are several TYPES OF CONTROL CHARTS which are generally classified into two major classes. These are Variable Data Charts and Attribute Data Charts. 1. Variable Data Charts: These charts are based on continuous natured data that varies and can be measured up to the required precision. The examples of data are length (of any object or stay), weight or volume etc. Individual charts, Average charts, Range charts and Moving Range Charts are Variable Data Charts. Individual Charts or I-Charts are used to plot individual process observations. These charts are especially used to identify the trends in process performance according to the average. Any exceptional trend of consistent below or above average performance may initiate the need of investigation. Average Charts or X-Bar Charts are used to plot the average values of subgroups formed using individual observations. Subgroups are selected data samples from the possible data. The subgroups are preferred when the original/individual data set is large. These subgroups can be acquired using multiple criteria like the processes furnished by different operators or through time slicing etc. The X-Bar Charts are used to measures variations in subgroups. These are one of the majorly used control charts. Figure 1: Average volume in subgroups of beverages. In Figure 1, the cyan horizontal line shows the proposed average of the product which has to be monitored, and two control limits are respectively drawn on the upper and lower sides of the theoretical (average) line. Each observation then projected and plotted on the chart area which indicates whether the observation (product unit) is closer or far from the average. This process should be continued till any observation even touches the pre-alarm or the realization (line plot) start making some similar trends. Range Charts or R-Charts are used to plot the variation in the ranges of variations exhibited by a process. The differences between highs and lows of each subgroup are considered as ranges. Moving Range Charts or mR Charts are also one of the majorly used charts. Variability among measurements that is based on the differences between successive data points is represented via these charts. Figure 2: Moving Range chart for volume of beverages 2. Attribute Data Charts: These charts are based on data that can only me measured in integers or whole numbers. These charts focus on the attribute or qualitative values produced by a process like number of defective and non-defective items etc. Here the subgroups comprise of the groups of units to identify the required and unwanted results. Defects and Rejects charts are examples of Attribute Data Charts. Defects Charts are used to plot the number of defects. They can be illustrated as c-Charts and u-Charts. The c-charts or control charts use constant subgroups sizes. These are used to indicate the consistency and predictability of defects. The u-charts or unit charts represent varying subgroups sizes. These are used when there is variability in the number of items tested in a subgroup. The main purpose of these charts is to indicate the number of defects per unit. Rejects Charts represent the count of rejected items in a subgroup. The two types and p and np charts that respectively represent the percentage of rejection and number of rejects. The p-charts are used for varying subgroups sizes. The examples are incomplete forms and incorrect dosages. REFERENCES: BenMore. (2010, July 14). The YouMOZ Blog: Statistics 101: Deviations. Retrieved from http://www.seomoz.org/ugc/statistics-101-deviations. Six Sigma SPC (9 June 2006) Statistical Process Control Overview. Six Sigma SPC. Retrieved from http://www.sixsigmaspc.com/spc/statistical_process_control.html Six Sigma SPC (13-Apr-2008) Six Sigma SPC's ZeroRejects Features. Six Sigma SPC. Retrieved from http://www.sixsigmaspc.com/products/zero-rejects.html Martinez, R. (n.d.) A Leadership Guide to Quality Improvement for Emergency Medical Services (EMS)Systems. NHTSA.gov. Retrieved from http://www.nhtsa.gov/people/injury/ems/leaderguide/index.html (n.d.) Statistical Process Control. NSTA. Retrieved from http://cache.micron.com/$assets$/beb62e98-5e25-47bf-b765-6804a3086726/spc.pdf ACS. (2011)Introduction to Statistical Process Control. Static.com. Retrieved from http://www.statit.com/statitcustomqc/StatitCustomQC_Overview.pdf StatSoft, Inc. (2011). Electronic Statistics Textbook. Tulsa, OK: StatSoft. WEB: http://www.statsoft.com/textbook/reliability-and-item-analysis/Retrieved Lean Sigma UK. (n.d.) Key people featured. Retrieved from http://www.leansigmauk.com/key-people.php Buttrey E. S. (June 2009). An Excel Add-In for Statistical Process Control Charts. Journal of Statistical Software. 30 (13)Retrieved from http://www.jstatsoft.org/v30/i13/paper Read More