Group technology and part classification coding

Part families, classification and coding (monocode, polycode, Opitz), production flow analysis and rank order clustering, with a worked cell-formation example.

Drafted with Aria, reviewed by the AiCanCode.org team. Spotted an error? Use Give Feedback at the bottom of the page.

Why it matters

A typical batch shop makes thousands of different part numbers in small quantities, and its machines are laid out by process (all lathes together, all mills together), so parts wander across the floor and wait in queues. Group technology (GT) exploits the fact that many of those parts are similar. Grouping them into families and making each family in a dedicated cell cuts setup time, material handling and lead time, and it is the organising idea behind cellular manufacturing, CAPP and flexible manufacturing systems.

Key ideas

Part families. A part family is a set of parts that are similar either in design attributes (shape, size, tolerances, material) or in manufacturing attributes (the sequence of machines and operations they need). Two parts can look different and still share a family because they follow the same routing — and two similar-looking parts can need different processes.

Three ways to form families.

  • Visual inspection — experienced people sort parts or drawings by appearance. Cheap, but subjective and limited to small part populations.
  • Classification and coding — each part gets a code describing its attributes; parts with similar codes form a family. Useful for design retrieval ("does a similar part already exist?") as well as for manufacturing.
  • Production flow analysis (PFA) — uses the routing sheets: build a machine–part incidence matrix and rearrange it so that groups of machines and the parts they process appear as blocks on the diagonal. Each block suggests a machine cell.

Coding-system structures.

  • Monocode (hierarchical) — the meaning of each digit depends on the digits before it, like a tree. Compact (many classes per digit), but hard to build and hard to search by one attribute.
  • Polycode (chain) — each digit position always means the same attribute regardless of the others. Easy to build and search, but needs more digits.
  • Hybrid (mixed) — the most common in practice, e.g. the Opitz system.

The Opitz system. Developed in Germany (Aachen), it has a 5-digit form code (digit 1 part class; digits 2–5 external shape, internal shape, plane-surface machining, auxiliary holes and gear teeth), a 4-digit supplementary code (dimension, material, original form of raw material, accuracy) and an optional secondary code for the company's own use. Digit 1 separates rotational parts by the length-to-diameter ratio — 0 for L/D ≤ 0.5, 1 for 0.5 < L/D < 3, 2 for L/D ≥ 3 — and uses higher values for rotational parts with deviations and for non-rotational parts. (Use the Opitz tables for the remaining class boundaries.) Other systems: MICLASS (TNO, Netherlands), DCLASS (Brigham Young University), KK-3 (Japan).

Rank order clustering (King's algorithm). A simple PFA method:

  1. Read each row of the machine–part matrix as a binary number (column weights 2ⁿ⁻¹ … 2⁰) and sort the rows in decreasing value.
  2. Read each column as a binary number (row weights 2ᵐ⁻¹ … 2⁰) and sort the columns in decreasing value.
  3. Repeat until neither order changes. Diagonal blocks are the cells; 1s outside the blocks are exceptional elements (intercell moves), handled by duplicating a machine, rerouting or subcontracting.

Benefits and limits. Benefits: fewer and shorter setups (similar fixtures and tools within a family), shorter handling distances, less work-in-process, simpler scheduling, standardised designs and process plans, team ownership in cells. Limits: coding an existing part population is slow and costly; cells can leave machines under-used; and families shift as the product mix changes.

Formulas

Row value = Σ (j = 1 to n) a_ij · 2ⁿ⁻ʲ

  • a_ij = 1 if machine i processes part j, else 0; n number of parts (columns). Used to sort rows in rank order clustering.

Column value = Σ (i = 1 to m) a_ij · 2ᵐ⁻ⁱ

  • m number of machines (rows), in the current row order.

L / D

  • Length-to-diameter ratio (dimensionless) for Opitz digit 1 of a rotational part; L and D in mm.

Worked examples

Example 1 (standard). Assign Opitz digit 1 to three plain rotational parts: a flange (D = 120 mm, L = 40 mm), a bush (D = 50 mm, L = 60 mm) and a shaft (D = 30 mm, L = 150 mm).

  1. Flange: L/D = 40/120 = 0.33 ≤ 0.5 → digit 0.
  2. Bush: L/D = 60/50 = 1.2, between 0.5 and 3 → digit 1.
  3. Shaft: L/D = 150/30 = 5 ≥ 3 → digit 2.

Example 2 (GATE level). Form cells by rank order clustering for 5 machines and 6 parts. Machine M1 processes P1, P3, P4; M2: P2, P3, P6; M3: P1, P4, P5; M4: P2, P6; M5: P3, P6.

  1. Column weights for P1…P6: 32, 16, 8, 4, 2, 1. Row values: M1 = 32 + 8 + 4 = 44; M2 = 16 + 8 + 1 = 25; M3 = 32 + 4 + 2 = 38; M4 = 16 + 1 = 17; M5 = 8 + 1 = 9. Row order: M1, M3, M2, M4, M5.
  2. Row weights in that order: 16, 8, 4, 2, 1. Column values: P1 = 16 + 8 = 24; P2 = 4 + 2 = 6; P3 = 16 + 4 + 1 = 21; P4 = 24; P5 = 8; P6 = 4 + 2 + 1 = 7. Column order: P1, P4, P3, P5, P6, P2.
  3. Recompute row values with weights 32…1 for P1, P4, P3, P5, P6, P2: M1 = 32 + 16 + 8 = 56; M3 = 32 + 16 + 4 = 52; M2 = 8 + 2 + 1 = 11; M5 = 8 + 2 = 10; M4 = 2 + 1 = 3. New order: M1, M3, M2, M5, M4.
  4. Column values with weights 16, 8, 4, 2, 1: P1 = 24, P4 = 24, P3 = 16 + 4 + 2 = 22, P5 = 8, P6 = 4 + 2 + 1 = 7, P2 = 4 + 1 = 5 — order unchanged, so stop.
  5. Cell 1: machines M1, M3 with parts P1, P4, P5. Cell 2: machines M2, M5, M4 with parts P3, P6, P2. One exceptional element: P3 needs M1, so either duplicate M1 in cell 2 or accept one intercell move for P3.

Common mistakes

  • Grouping by appearance alone; manufacturing families are defined by routings.
  • Mixing up monocode and polycode — in a monocode, the meaning of a digit depends on the previous digits.
  • Forgetting to re-weight rows after the columns are reordered (and vice versa) in rank order clustering.
  • Stopping after one pass of ROC instead of iterating until neither order changes.
  • Treating every exceptional element as a reason to add a machine; compare its cost with an occasional intercell move.

For GATE PI

Expect MCQs on monocode, polycode and hybrid codes, the structure of the Opitz code, and GT benefits. Numericals typically give a small machine–part matrix and ask for the rank-ordered matrix, the number of cells or the exceptional elements, or ask for an Opitz digit from given dimensions. Practise binary weighting quickly and carefully.

Quick check

  1. In which coding structure does each digit have a fixed meaning independent of other digits?
  2. How many digits does the Opitz form code have?
  3. A rotational part has D = 80 mm and L = 30 mm. What is its Opitz digit 1?
  4. A row reads 1 0 1 1 for parts P1–P4. What is its ROC row value?
  5. What is an exceptional element?

Answers: 1. Polycode (chain). 2. Five. 3. L/D = 0.375 → 0. 4. 8 + 2 + 1 = 11. 5. A 1 outside the diagonal blocks — a part needing a machine in another cell.

Try answering each one aloud before you open it.

  1. 1.What is Group Technology in the context of manufacturing?Concept

    Group Technology is a manufacturing philosophy that seeks to improve productivity by grouping similar parts or products into families. These families are processed together, which helps in reducing setup times, improving workflow, and enhancing overall efficiency. The idea is to capitalize on similarities in design and production processes.

  2. 2.Explain the concept of part classification and coding in Group Technology.Concept

    Classification sorts parts into families by similar design or manufacturing attributes; coding records those attributes as a string of digits so similar parts get similar codes. Codes can be monocode (hierarchical, each digit's meaning depends on the previous ones), polycode (each digit has a fixed meaning) or hybrid, such as the Opitz system with a 5-digit form code and 4-digit supplementary code. Designers use the codes to retrieve existing similar parts, and planners use them to form families for cells and CAPP.

  3. 3.Why is Group Technology important in Computer Integrated Manufacturing?Application

    Group Technology is important in Computer Integrated Manufacturing because it helps in organizing production processes more efficiently. By grouping similar parts, it reduces setup times and inventory levels, enhances machine utilization, and improves production scheduling. This leads to cost savings and increased productivity, which are crucial in a competitive manufacturing environment.

  4. 4.How does part classification coding contribute to reducing production lead time?Application

    Part classification coding contributes to reducing production lead time by enabling the identification of part families that can be processed together. This reduces the need for frequent machine setups and changeovers, leading to faster production cycles. Additionally, it helps in better planning and scheduling, which minimizes delays and bottlenecks in the production process.

  5. 5.What are the potential challenges in implementing a part classification and coding system?Application

    Implementing a part classification and coding system can be challenging due to the complexity of accurately categorizing a wide variety of parts. It requires a thorough understanding of part attributes and manufacturing processes. Additionally, the initial setup can be time-consuming and may require significant changes to existing workflows. Ensuring consistency and accuracy in coding is also a critical challenge.

  6. 6.Explain how Group Technology can lead to improved quality control in manufacturing.Application

    Group Technology can lead to improved quality control by standardizing processes for similar parts, which reduces variability and defects. By processing part families together, it becomes easier to monitor and control quality parameters consistently. This approach also facilitates the identification of common defects and their root causes, enabling more effective corrective actions.

  7. 7.What happens if a part is incorrectly classified in a Group Technology system?Application

    If a part is incorrectly classified in a Group Technology system, it may lead to inefficiencies such as increased setup times, improper machine utilization, and potential quality issues. The part may not be processed with its correct family, leading to delays and increased production costs. It can also disrupt the workflow and scheduling, affecting overall productivity.

  8. 8.Discuss the role of computer software in implementing part classification and coding systems.Application

    Computer software plays a crucial role in implementing part classification and coding systems by automating the categorization process. It helps in managing large datasets of part attributes, ensuring accuracy and consistency in coding. Software tools can also facilitate the integration of classification systems with other manufacturing processes, such as inventory management and production scheduling, enhancing overall efficiency.

Finished this topic? Mark it so your progress, study plan and readiness keep up.

Stuck on something here?