Extrusion Intelligence, OEE (Overall Equipment Effectiveness) for extrusion plants Second part “Performance” This is the second article of a series of three concerning the application of OEE method in extrusion, wherein the production performances will be deepened. The goal is to go beyond the basic analysis of a generic KPI “Performance”, to create and manage more specific KPIs that allow an immediate identification of any lack of performances and the related causes and possible solutions: define more specific KPIs is needed as a generic one is just a simple average, although a weighted one. Design and analyze more specific KPIs allows to know and improve situations that a generic KPI can highlight as good ones but that actually have vast room for improvement. To reach this goal, as specified in the previous article, it must be defined an analysis path that, starting from aggregated data related to the productivity, allows to analyze the single specific detail. How to measure production Performances? The pic. A shows the generic KPI Performance = D/C (Real net Kg/h / Theoretical net Kg/h; highlighted with the red rectangle) which measures the lack of performances due to extrusion speed decreases, missing billet optimizations, technical scraps and a bad planning. This document will analyze the performances increase function of technical scraps decrease and define the basic KPIs needed to check them. Check and analyze the KPIs related to the production is mandatory to apply the continuous improvement philosophy. The performances analysis does not consider the downtimes occurred during the production process as they were already considered in the first article (“Availability KPI”). Pic. A

The following chart shows the logical operation of “Availability” and “Performance”. The base of the KPI “Performance” is the real operating time and the KPI “Net KGs/h”, without considering the downtimes, to separate the KPIs related to downtime from the ones related to the puller speed and the billet optimization (technical scraps). Such analysis is made basing on the data of a company using a billet optimization system, whose length is calculated basing on the order and characteristics of the alloy. Because of this the technical scrap percentage is calculated all the time basing on the optimal length of the billet. The KPI “Availability” is calculated with the following formula: B/A= Availability (Operating Time / Planned production time) KPI “Availability”

The KPI “Performance” is calculated with the following formula: D/C= Performance (Real net Kg/h / Theoretical net kg/h) C= Theoretical net kg/h = theoretical net KGs deliverable on Planned Production Time D= Real net kg/h = real net KGs produced on Operating Time

KPI “Performance”

The following table shows the KPIs recommended to be kept under control to have a good management of the technical scraps; red-highlighted the ones analyzed in this document. Description Production (Gross KGs) Production (Net KGs) Total scrap (KGs)

U.M. Kg Kg Kg

Theoretical net KGs/h (operating time) Real net KGs/h (without considering the first billet) Real net KGs/h (considering the first billet) Real net KGs/h (considering a longer first billet) Scrap (KGs/h)

Kg/h Kg/h Kg/h Kg/h Kg/h

Total technical scrap (%) Technical scrap (%; without considering the first billet) Technical scrap (%;considering the first billet)

% % %

Billet length weighted average Handling table length weighted average Lot size weighted average

mm mm kg

Technical scraps Technical scraps means head & tail scraps, joints & butt-ends, the billets required to form the profile and the scraps generated at the shear, that is any scrap which cannot be eliminated but that can let you make a better production run if optimized. What’s the value of a one percent (1%) decrease of the technical scraps? The pic. B shows the different results obtained considering different production runs on European plants, with 6060 alloy, billet between 7” and 10” and assuming a one percent average decrease. Billet diameter

inches 7 8 9 10 Pic. B

mm 178 203 228 254

Current Production

% scraps decrease

t/year 6000 8200 10200 14500

% 1% 1% 1% 1%

Production increase t/year t 60 82 102 145

The analysis will be based on data collected on a three-presses facility and a theoretical technical scrap KPI equal to 10%. The pic. C shows the daily production progress, highlighting gross KGs, net KGs and scrap percentage. The weighted average of the real scrap, for the three presses, is equal to 10,9%; the graph shows a growing trend but basically a positive situation with a 0,9% deviation from the theoretical KPI. Starting from this data the user can check if there is room for improvement, so the causes of this scrap must be understood and, above all, the economic advantages due to its decrease must be valued.

drill down

Pic. C Besides a “what-if” analysis that, assuming to decrease to the theoretical KPI value (10%) any die exceeding it, can be done: such analysis calculates the theoretical advantages due to this decrease.

The pic. D shows the user interface where the user defines the scrap percentage which will be the base of the what-if analysis (10% in this situation) while the following image shows the results obtained. The simulation calculates all the production runs made during the time range set (two and a half months on three production lines), with interesting results: to make the same net production of 3.631t the raw material used is decreased by 2,15% (about 87t for the time range analyzed; 400t an annual basis). Assuming to keep constant the raw material used (4.077t) the net production is increased by 78t (87t – 10% technical scrap), equal to 2,17% (about 370t on annual basis).

Pic. D

Pic. E, a graph generated by the simulation, shows the difference between real values and simulated ones, for each press: it’s evident that there is a vast room for improvement, in particular on presses 1 and 3.

Pic. E

Press 3 analysis With a “drill-down” on the daily production progress graph (pic. C) a new graph, including the alloys used during the time range analyzed, is shown (pic. F).

drill down

6060

6063

6082

6005

6005A

6061

Pic. F A new drill down let you analyze in detail the data about 6082 alloy, with a 17,9% average technical scrap. The analysis consists in the research of all the dies with a technical scrap higher than 10% and understand if it’s possible to decrease them to 10%. The drill down shows the dies with the highest incidence and shows clearly that there is vast room for improvement (pic. G; shows the 10 dies with the highest incidence). Obviously the dies can be shown as sorted lists in addition to the graphic visualization, according to the user choice.

Pic. G This analysis example shows that, even starting from an average value generally good (10,9% average scrap compared with the theoretical 10%), a proper analysis of the data can let you identify dies with technical scraps values which justifies an intervention to improve them.

Technical scraps and first billet A technical scrap usually is valued with a percentage related just to head, center and tail scraps, butt-end thickness and first billet required for profile formation. What is not considered many times is the size of the production lot, while it’s clear that the incidence of the first billet varies according to the lot size; therefore, defining the technical scraps KPIs, the head, center and tail scraps and butt-end thickness are separated from the scraps caused by the lot size. The pic. H shows this situation (technical scraps according to lot size and first billet length). The calculations have been done basing on the following data: Press 1650t, double handling table, profile weight 1600gr, bar length 6500mm, total scraps 2550mm, buttend 20mm, speed 20 m/min. The lines on the graph represent: - Brown line, scrapped first billet (400mm) - Light brown line, scrapped first billet (250mm) - Orange line, without first billet - Purple line, longer first billet Both in the situation with a first billet length equal to 250mm and in the one with a longer one (400mm) it has been done an optimization to reach the puller coupling: in this second situation the optimization set an elongation of the first billet, so the scrap will be just a single butt-end. It’s clear that production runs of very small lots are often the cause of high technical scraps while a good planning which groups many orders to a single, bigger, production one is the best possible solution to reduce these scraps.

Technical scrap % function of the quantity extruded, and the first billet 20 18 16 14

% Scrap

12 10 8 6 4 2 0

Pic. H

Production Order 3000 kg

Production Order 2000 kg

Production Order 1500 kg

Production Order 800 kg

Production Order 500 kg

Production Order 300 kg

Scrap 1° Billet 400mm

12,5

12,9

13,2

14,3

15,4

17,5

Scrap 1° Billet 250mm

12,3

12,5

12,7

13,4

14

15,4

Scrap no 1° Billet

11,9

11,9

11,9

11,9

11,9

11,9

Scrap Long 1°Billet + 230mm

12,2

12,4

12,6

13,3

14,1

15,3

The pic. I shows the productivity (KGs/h) basing on production lot size. Using a longer first billet, when possible, allows to increase considerably the productivity as almost all the first billet cycle time is eliminated.

Productivity, function of the quantity extruded, and the first billet 1600 1400 1200

Net Kg/h

1000 800 600 400 200 0

Pic. I

Production Order 300 kg

Production Order 500 kg

Production Order 800 kg

Production Order 1500 kg

Production Order 2000 kg

Production Order 3000 kg

Kg/h 1° Billet 400mm

1026

1154

1240

1325

1351

1379

Kg/h 1° Billet 250mm

1060

1180

1260

1337

1361

1386

Kg/ no 1° Billet

1440

1440

1440

1440

1440

1440

Kg/ 1° long billet

1403

1413

1421

1429

1431

1434

An example of an optimal check is shown in the pic. L, a user interface including the following indicators, monitoring three events kind: - net KGs/h - % scraps - Setting speed The KPIs included in the pic. L (user interface of press performances control) are represented with colored vertical bars: - blue (theoretical values) - orange (real values without considering the first billet) - brown (real values considering the first billet) Theoretical values are real-time calculated, billet by billet, basing on the referred parameters stored in a technological database; these calculations are done without considering the first billet as the incidence of its cycle time and the scraps may vary considerably according to the lot size (real data are collected automatically from the press). The KPIs included are: - theoretical net KGs/h (blue bar) - real net KGs/h without considering the first billet (orange bar) - real net KGs/h considering the first billet (brown bar) - theoretical scrap % (blue bar) - real scrap % without considering the first billet (orange bar) - real scrap % considering the first billet (brown bar) - theoretical speed set (blue bar) - real speed set (orange bar) The left area includes the data grouped by shift while the right one includes the data of the die selected. The horizontal bars are indicators representing the difference between theoretical values and real ones: - KGs/h ones – Highlights a positive situation if >=1; a negative one if