Different sectors that make up the industry are experiencing difficult times and the need for businesses to squeeze the last ounce of efficiency out of resources has never been greater. In several articles spread over the last three years, StitchWorld has concentrated upon the sewing room discussing operator training, garment make up methodology and production management issues. With managers now having to save money wherever they can, it seems to be good time to go back down the production chain and look at a key activity in many knitwear factories, especially in the circular knitting department.
Paul Collyer, a garment industry expert with over 37 years of international exposure in production management, specialist consultancy and training, discusses various pressures that a manufacturer undergoes to improve productivity.
Circular knitting consists of a major chunk of production activity in any vertically integrated cut-and-sew knitwear factory. Due to higher capital investment this section generally runs in three shifts to maximize return on investment. Although from textile manufacturer’s point of view measurement of efficiency means machine run time as against machine idle time but from an apparel manufacturer’s perspective the issue is comparatively complex. During a recent project the following questions were raised:
- How do you measure labour content of circular knitted products?
- How do you decide optimum number of machines per operator?
- Which is more important? Good machine utilization or good operator utilization?
- How do you determine the appropriate machine downtime even when operator loading is optimum?
The following ideas on the above questions should stimulate debates within companies. However we should initially go back to the management mantra “If you cannot measure it, you cannot control it. If you cannot control it then you cannot manage it”. Accordingly the first point to be considered is the measurement of the process. Without accurate measurement it is not possible to answer any of the queries or manage the process.
There are two major measures to be established while measuring labour content of circular knitted products: firstly, the time a machine needs to produce one unit, and secondly, the SMV or labour content involved in producing one unit. Dependent upon the product being manufactured and the method of work, the unit may be expressed in garments, i.e. T-shirt bodies (for flat knits), metres or kilos produced (for circular knits). The SMV or labour content is not the same as the machine time as operators will usually be responsible for more than one machine (see Time Study Top Sheet).
For measurement purposes, the operations within the knitting production processes can be divided into two categories: predictable and unpredictable.
Predictable processes: The machine produces its product and the operator unloads, performs any other necessary task and restarts the machine. This work is completed every time the machine produces a unit. It is possible to measure the cycle of the machine and the time necessary to bring it back into work at the end of the cycle. The machine cycle is easy to time but during the machine cycle if unpredictable elements occur they are isolated to produce an accurate working time and incorporated in later analysis.
End of operation activities can be measured by time study. Narrative time studies will enable multiple machines to be checked making the measurement process more time-effective. The study will time the operator and not the machine. Elements in a narrative study may include activities such as:
1. Fetch materials
2. Perform end of operation activities, recorded by machine
3. Perform unpredictable (in machine cycle maintenance) operations; recorded by machine
4. Patrol (monitor machines)
Unpredictable processes: The machines need maintenance such as replacing needles, cones and mending broken threads. It is known that these necessary actions happen and that they occur during the machine cycle, causing delays but it is not possible to predict how many machine cycles will be performed per action. It is therefore necessary to discover the time they take as a percentage of the machine cycle. This is done using narrative time studies and the results checked using Activity sampling as previously explained.
It is now possible to work out the load factor for each machine. The load factor of a machine is defined as the proportion of the overall cycle time required by the operator to carry out the necessary work to keep the machine running.
If a machine operates for 15 minutes to produce one unit, 2.0 minutes for end of operation activities and unpredictable elements are identified as 4% of machine cycle, then the total time necessary is 17.60 minutes. Load factor is 2.6/17.60 or 15%. The operator should therefore need 15% of 60 minutes (9) in each hour to tend the machine. (Operator rest allowances have been added).
[bleft]A major factor in deciding the optimum loading of the operator is the planned capacity of the machine. By using a measured data it is possible to establish accurate output capacity for each machine[/bleft]
It should be noted that the above is the labour content necessary for the operator to mind the machine to produce one unit, it is not the time needed for costing purposes. If the operator needs nine minutes to mind the machine but is only allocated three machines then the labour content can be expressed as twenty minutes.
The calculation of a Standard Minute Value (SMV) is therefore potentially more complicated as it depends upon the loading of the operator. The combination of possible operator loadings and differing machines can be enormous therefore the company will need to determine a standard to be applied to all calculations of SMVs to enable a practical system to be established.
Dependent upon company practices and procedures it may be necessary to add a further activity to the load factor as previously calculated. The calculation measures the time taken by the operator to keep the machine running; there is no allowance for “patrolling”. It can be argued that modern machines should stop when they need attention but as a security measure a person constantly monitoring machines to avoid quality issues and to pre-empt problems, is preferred by most companies. If this practice is followed then both the labour content and load factor will need to be adjusted. There is however no foolproof method of predicting the effects of “patrolling” in that X% of time on patrol will reduce quality issues by a set percentage. All that can be done is set the operator up to patrol for a percentage of his time and monitor results. Then reduce patrol time and see if there is change in output or faults. However before committing to a “PATROL” allowance management should question if they are just maintaining custom and practice or is there a genuine need. Is there a measurable relationship between patrol time and quality fault production?
[bleft]We should be asking questions to ourselves: Does the machine (type) have excess capacity over planned orders? Is machine interference an issue? Which is more important? Good machine utilization or good operator utilization[/bleft]
Having established the labour loading factor to run the machine, it is possible to decide on the maximum number of machines an operator can manage. In the majority of factories an operator will be working on a number of identical machines producing the same product. It should not however be assumed that identical machines perform in the same manner and during the measurement phase every machine should be assessed. Variances in output from machine to machine should be investigated; is the machine set up incorrectly? Does it need investment to bring it up to standard? Is that investment cost-effective? For operator loading purposes the demands of individual machines should be factored in but it is easier when costing to concur to a standard for all machines.
When load factors for each machine and product are known it is of course possible to calculate and load an operator up to a 100% load factor. If load factor in one machine is 20% and operator is assigned 5 such machines to look after, then operator is working up to 100% load factor. But the problem with such a high load factor is the probability of a situation where the machine is waiting for the man. However, as the number of machines the operator is responsible for is increased, the greater the probability that two or more will simultaneously need attention and be idle. This is referred to as machine interference. Machine interference is the queueing of machines for attention, i.e. when the worker is attending to one machine, others have finished their cycle and are idle.
How do you determine the machine down time caused by loading operators to optimum? How do you decide optimum number of machines per operator?
There is no “magic” formula for the above questions but with all known measures for each machine and product it is possible to allocate machines to operators to a predetermined load factor, for e.g. 90%. Activity sampling should be used to determine the machine interference for a given load factor.
Activity categories could include:
– Machine working
– Machine – end of cycle activities
– Machine – unpredictable activities
– Machine idle waiting for operator – machine interference
– Machine idle waiting for maintenance department.
Machine interference as a percentage of available time can then be established. An interesting point is to compare the results from the same group of machines when other operators are working, i.e. during different shifts. The load can then be increased or decreased until an optimum balance between load factor and machine interference is established. A major factor in deciding the optimum loading of the operator is the planned capacity of the machine. By using a measured data it is possible to establish accurate output capacity for each machine; the machine cycle including both predictable and unpredictable activities plus interference allowance for a budgeted load factor. If the machine needed to run 100% of the time, then the load factor of the operator can be reduced and machine interference checked by activity sampling. Alternatively operators can be instructed to give individual machines priority or a mixture of high and low load machines allocated. By including load factor in incentive schemes operators can be encouraged to work to greater loadings!
All of the above are not the “definitive” method of measuring and planning the circular knitting function. They are however a practical way that a company with a competent work study department can establish a means of planning machines and labour, allocating costs and managing the knitting process. The principles also apply to other areas of the apparel industry that utilizes semi-automatic machinery needing machine minding such as embroidery departments.
We should be asking questions to ourselves: Does the machine (type) have excess capacity over planned orders? Is machine interference an issue? Which is more important? Good machine utilization or good operator utilization?
If the machine is running with excess capacity, then it may not matter if the machine stands idle waiting for the operator. All the decisions should be based on facts! The key consideration should be planned machine loading. With accurate measures capacity can be calculated and machines planned and operators allocated to predetermined load factors. In developed countries where labour cost is high, emphasize should be to maximize operator utilization. While in developing countries where labour is comparatively less expensive, machine utilization should be the goal.