DATA MINING , ANALYSIS AND IMPROVEMENT SOLUTIONS OF THE EFFICIENCY PRODUCTION.

Tools and idea considered to write this article are:

-       Excel.

-       OEE, stand for Overall equipment effectiveness

-       Pareto for analysis.

-       Kaizen philosophy.

-       Lean six sigma.

1)   GOAL: Improving up to 10% efficiency of a product line. Specifically, we have chosen 4 machines:

Factory divisions 

We have worked in printing division. This division is composed of 4 machines. We analyzed them. The printing division showed the Worst OEE (Bottleneck).

1 -     OEE Valuation

OEE is the indicator that measures the production efficiency of the line or machines in an established time.

OEE = availability of the machine X performance of machine X quality.

OEE is a percentage

-       Solar time = 356 days

1)   PLANT OPERATING TIME = Solar time – Unscheduled time = the shifts of the work for example if we have two shifts per day we can calculate it by 16 hours/day.

2)  Part of the plant operating time is used for PLANNED SHUTDOWN, during which there are maintenance activities planned, training of workers, etc.

3)  PLANT OPERATING TIME – PLANNED SHUTDOWN = PLANNED PRODUCTION TIME. This is the time available for producing.

4)   Planned production time doesn’t correspond to the effective production because of breaks down of machine, set up, waiting for lack of workers, waiting for product missed. The duration of each waste of time is called DOWNTIME LOSSES.

5)  PLANNED PRODUCTION TIME – DOWNTIME LOSSES = OPERATING TIME. The operating time is the AVAILABILITY  

6)     SPEED LOSSES = lack efficiency, slowdowns, micro stops. It depends on the machine you are analyzing and could Be:

-       Machines slowdowns.

-       Check by the operators.

-       Missing attachment of the sucker on the carton.

The effect is having an output different from what planned. Speed losses made the PERFORMANCE.

7)     QUALITY LOSSES = the percentage of the production pieces on total production. QUALITY

2 -      INDEX

3 -     EVALUATION OF THE KIND OF ANALYSIS

Important is the evaluation of the analysis to individuate the right machines to study.

Critics’ machines mean the ones who determine the major losses.

The next evaluation is brought by “Abdul Samat et al. (2012)” says that the identification phases of the critic machines are divided into the following point:

-       Losses identification = losses linked to a specific machine, all kind of losses.

-       Frequency Analysis of occurrence of losses = Using Pareto tool.

-       Machine selection = Selecting the machine in base on the Pareto Analysis.

4.1 PARETO ANALYSIS

The Pareto Analysis says that 20 % of the causes determine 80% of the effects. Solving this minority we can get great improvements.

How to draw Pareto:

-       Draw a table.

-       Insert the Losses you want to monitor.

-       Minutes sorting by descending, in base on the analysis you want.

-       Calculate the total of minutes, the percentage of the total of minutes and the cumulate percentage.

Pareto on Excel

Results of Pareto analysis. It shows the most important losses correlated with the time of Data Mining.

From the picture above we can see which issues need more attention in order to improve the

line efficiency.

After five months of data mining, we solved problems of this three issues increasing OEE of
11%.
Important is DON'T GIVE UP.

Emanuele