One of the attractions of JIT is its simplicity. Essentially the goal of JIT is the production and delivery of the required items at the required time in the required quantity. JIT is 'pull' or demand driven because nothing should be produced until just before it is needed. The demand for parts or finished goods thus pulls product through the system. The phrase Sell daily - make daily encapsulates JIT but in terms of moving the product along the production line there is an essential difference between JIT and the conventional line.
In the conventional line the work stations produce goods at a speed dictated by the line irrespective of whether the next operator is ready or not. This leads to a build-up of work in progress (and thus cash) in front of any bottlenecks or slower operations. In the JIT system an item is only passed to the next station when they want it. In the words of an American JIT practitioner 'Don't make nothing, no how for nobody, make them come and get it.'
While regarded as a production management system it is much more involved with methodology rather than with technology. It can be a total philosophy for improvement and radical innovation. All production areas can be driven by the single JIT concepts. The real strengths lie in its simplicity and the ease with which it can be explained to all levels of staff. Computerisation of JIT is at an early stage but even when computerized it only involves low level computing power to replace the manual systems. Because it is so accessible it serves as a method for uniting all employees in the common goal of improved performance.
The process of production under JIT has been compared to a river with large rocks. The river is the movement of materials and the depth of water represents Work In Progress (WIP). As the water level is lowered (by driving down inventories) the rocks are revealed. These must be removed to leave less water in the river while allowing it to continue flowing smoothly and swiftly. Typical rocks encountered are quality, changeover times, vendor conformance and product design.
Figure 1: The rocks in the river for JIT
Despite the good things said above JIT is a risky manufacturing philosophy. Inventories are minimal and planning is essentially short term. If the rocks are not removed swiftly then the whole business can founder.
Conventional production planning systems such as MRP and MRPII are 'push systems'. An important measure in push systems is machine usage and the attempt to balance production lines is given a lot of effort. There is often a build up of WIP throughout the process that is made worse at bottlenecks in the operation.
Push systems are often unresponsive to changes in customer requirements. The concept of Economic Batch qualities is generally very important and the need to schedule and run push systems generally results in complex computer programs. In a pull system, components are only produced when there is an order or requirement for the product. This generally results in simpler control systems with greatly reduced inventory levels and batch sizes.
In a JIT environment machine and labour utilisation are not relevant as machines are only used when they are needed. An operator passes along work only when a signal KANBAN is given that the next operator is ready for the work. Such signals are generally visual and easily understood.
- JIT is not simple to implement and many 'rocks' need to be worked on simultaneously.
- JIT is not a quick fix.
Implementation demands careful planning, a firm strategy, the right environment (awareness, industrial relations etc) and determination.
JIT encourages a move away from functional plants to the use of manufacturing work cells. A work cell will contain a natural group of machines and/or people and skills. These various machines and skills will be such that a cell can manufacture a whole product or component e.g. rather than have a welding area where all welding is carried out, the welders will be distributed around the factory. This will result in reduced transport costs (a non-value activity or NVA, reduced WIP and better control of quality. Work cells also create an environment of ownership and autonomy. The cells also control quality and almost naturally seek to improve the process and product.
Before you dismiss the idea of work cells and their efficiency or effectiveness then consider this question: When you were a small fabricator and had 3 people you probably produced around 50 windows per week, when you expanded to 6 people what did your production output go up to? Logic tells us that it should have been 100 windows per week but if it was then I would be surprised. So, what happened to the missing output?
Small groups of 2-3 people can be run by 'team 'leaders' who also do productive work. By the time you get to 6 people you start to need a manager who does no productive work but simply manages - he becomes a 'wealth dissipater' rather than a 'wealth creator'. For the conventional factory idea there is little linearity between the numbers of people and actual output i.e. doubling the number of people rarely doubles the output. Work cells give us a way out of this common dilemma.
Figure 2: Work flow in current factory
The work cell idea can best be explained by considering the work flow in the current factory (shown in Figure 2). Work moves around the factory along a long production line, which is never truly balanced, so that some areas require more labour than others. An increase or decrease in sales requires staff changes which are never quite whole numbers of people and you always end up with a few more or less people than you need.
Note that the distance parts travel (a crude but accurate measure of efficiency) is quite high, indicating low efficiency. In this model of the factory we have a 'serial' type of production flow i.e. the work passes along one line via a series of individual stations. This gives a well defined flow of materials but can give problems with balancing lines and most importantly with poor operational reliability. If a station in a serially coupled system fails then the whole line stops. If each station is working at slightly less than its full capacity then the total line will be working at greatly less than its full capacity.
Serial flow has some advantages in that it is easy to change the number of workstations and, given that volumes are sufficiently high for each product variant, then changeover between different types can be easy. The disadvantages are that only one type of product can be produced at one time production cannot really be started until the whole line is changed and ready to operate and that there are constraints on the standard order of assembly. By dividing the factory up into work cells we can approach the parallel type of production flow. In this type of flow similar operations are performed in various independent workstations or cells. The model for this is shown in Figure 3.
Figure 3: Workflow in cell factory
This is the equivalent of having 4 small fabrication shops running side by side and using this method we can get good linearity between the number of people and output. Doubling the number of people really can double the output. Increasing or decreasing sales can be coped with by setting up or closing down individual cells and the distance traveled is quite small and therefore the layout is highly efficient.
Parallel systems based on work cells have an inherently higher operational efficiency and many different products can be produced at the same time. You could have a woodgrain cell, a white cell, a door cell, a T/T cell and a casement cell. Any product variations are easily coped with by the cell concept.
Cells offer a method of increasing output, quality, workforce involvement and skill whilst at the same time providing greater flexibility for the manufacturing process.
In the past most of our effort has been put into trying to reduce the cycle time or speed up the output rate whilst totally ignoring the changeover time from one product to another. This has resulted in small batches appearing uneconomic to run and the Economic Batch quantity (EBQ) concept has developed.
Consider how you change from white profile to brown profile when you have an order that consists of 8 white and 2 brown windows. The answer is generally simple - you make the white ones and store them until you have enough orders for brown to make it worthwhile changing over. What would life be like if you could produce alternate white and brown frames? What would it do to your cash flow, storage needs and ability or desire to sell brown products? Reducing set up times allows the introduction of variety as a competitive edge and today we need all the competitive edge we can get.
Reducing the set-up time (on which we rarely concentrate.) can give the equivalent of huge increases in process speed (on which we almost always concentrate). This is all achieved without detriment to product quality, almost always a by-product of increasing output speed.
The ideal of a set-up time reduction plan is to move towards SMED (Single Minute Exchange of Dies) or OTED (One Touch Exchange of Dies). These remove set-up times entirely and make EBQ concepts redundant. Large batches no longer appear on the shop floor, lead times disappear, work in progress disappears, customer response is improved and variety can be increased. Make daily and sell daily becomes the norm rather than a dream.
It has been estimated that 80% of the benefits can be achieved simply by working smarter rather than faster.
So how do we work smarter? In general the sequence for improvement (without investment) is:
Step 1: Analyse your existing changeover times. Try putting these on a chalkboard and you will see an immediate difference simply because you are showing that it is important to you.
Step 2: Divide these times into internal and external components as follows:
Internal - These are operations that can only be carried out when the machine has stopped. Areas for improving internal factors are:
- Quick change tooling/connections
- Standard base plates
- Combining handed tooling
- Parallel operations
- Set-up sheets
External - These are operations that can be carried out while the machine is running. Areas for improving external factors are:
- Pre-setting of tooling
- Pre-kitting of gauges
- Tool kits/special equipment availability
- Standard base plates
Step 3: Convert internal operations to external operations where possible (this may involve duplicates of existing tooling).
Step 4: Streamline all aspects to give best results.
Figure 4: Set up reduction flow chart
Set-up times are a key area in the drive continuously to improve performance. Standards and targets must be set regularly and visibly displayed. There must also be a regular audit to evaluate rates of improvement.
A factory operating on JIT is especially susceptible to variable quality. When a component is produced just in time it must be correct. There are no buffer stocks to rely upon if defects are produced. The JIT method focuses quality control and defect prevention where it really belongs - at the operator level rather than as a separate function.
An essential in JIT is reducing the variability of all products and SPC (Statistical Process Control) is necessary. Both machine and process capability must be established to ensure reliable results. These are be dealt with in more detail as part of the section on Quality Control.
"The Manufacturing Strategy" series is designed to give production managers and their staff some insights into new manufacturing methods and to prompt the industry into considering the benefits of alternative approaches to manufacturing. The series is:
Part 1: Setting the strategy
Part 2: The systems and MRP II
Part 3: Just in time (1) (This section)
Part 4: Just in time (2)
Part 5: Just in time (3)
Part 6: Optimised Production Technology (OPT)
Part 7: A fundamental quality
Part 8: Quality management techniques & tools
Part 9: ‘There's no accounting for manufacturing strategy’
Part 10: Performance measurement
Part 11: Changing roles and things to do NOW!
Last edited: 11/03/10
© Tangram Technology Ltd. 2001
Our standard disclaimer regarding Internet data applies.