Dillygence
Flexible factory: modularity, OEE and controlled lead times
Flexible factory: replace rigid lines with modular cells, reduce WIP, and stabilize throughput time through simulation.
Article (WORD)
TITLE: Flexible factory: modularity, OEE and controlled lead times
Introduction: building a flexible factory to keep pace with ever-shorter product cycles
In many European workshops, 20 to 40% of production time is lost to waiting, transfers, adjustments, and rework, according to field diagnostics and the Factory Physics models of Hopp and Spearman. This waste inflates WIP, stretches lead times, and drives overinvestment. When changeovers take an hour instead of twenty minutes, the company buys capacity "to compensate" — i.e., fixed CAPEX (Capital Expenditure). A flexible factory aims for the opposite: more useful throughput without locking in a product mix assumption for five years.
Here, "flexible" refers to industrial organization: how to design and manage a workshop that changes volume, mix, and routing without losing control of flows. The floor quickly passes judgment: if WIP (Work In Progress) rises and lead times lengthen, flexibility remains a promise.
Key takeaway: a flexible factory replaces dedicated lines with reconfigurable modules and a management approach that contains variability. It is validated on concrete indicators — throughput time, WIP, service rate — and decisions come faster when a digital twin tests scenarios before each euro is committed.
I. Defining a flexible factory without illusions
Operational definition: objectives, scope, measurable promise
A flexible factory is a site capable of adapting its production to variations in volume, mix, and products, without major drift in unit cost, lead time, or quality. The promise is measurable: lower WIP, reduced throughput time, higher OEE (Overall Equipment Effectiveness), and better on-time delivery. A flexible architecture keeps options open — adding a cell, modifying a routing, transferring a product family — without major construction work. It protects both the industrial asset and the cash tied up in WIP.
Keeping the terms straight: flexible production, flexible line, reconfigurable factory, modularity, agility
Flexible production describes the ability to manufacture different references with controlled changeover costs. A reconfigurable factory builds on the RMS (Reconfigurable Manufacturing System) logic theorized by Yoram Koren at MIT (Massachusetts Institute of Technology), with modules that recombine. Modularity describes design by interchangeable blocks with common interfaces. Practical agility is measured by reconfiguration time, service stability, and the cost of variability.
The acronyms that always come up: OEE, SMED, WIP
OEE links availability, performance, and quality: it reveals whether flexibility is paid for in downtime or defects. SMED (Single-Minute Exchange of Die) drastically reduces changeover times, thus the "tax" of mix. WIP measures work-in-progress and serves as an alert when variability goes out of bounds. These three indicators guide simple decisions: smaller batches, more changeovers, or conversely stability and leveling.
II. The "roaming" bottleneck that consumes CAPEX
Why the bottleneck shifts when the product mix evolves
A bottleneck does not stay fixed when product families change, because each routing loads workstations differently. Dedicated lines handle this poorly, as they optimize a "flagship product" scenario. The "roaming" bottleneck ends up dictating successive investments, workstation by workstation, without stabilizing the overall lead time. When a workstation speeds up without synchronization with downstream operations, it feeds an intermediate stock that masks the problem rather than solving it.
Mini-case 1: moving from single-product to multi-variant without WIP explosion
What: a mid-sized company (ETI) was moving from a single product to a range of variants, with frequent changeovers.
How: the team applied SMED to two critical workstations, then defined sequencing rules and capped batch sizes, with limited buffers by zone.
Impact: WIP dropped by approximately 20 to 30% within scope, and throughput time stopped drifting during mix peaks, without adding machines. The gain came from better-contained variability, not from higher raw capacity.
III. Liquid factory and flexible production systems: RMS logic applied to SMEs and mid-sized companies
The "liquid factory" describes a flexible factory structured in modular cells, with flows that redirect according to load and routing. The work of Fraunhofer IFF (Fraunhofer-Institut für Fabrikbetrieb und -automatisierung, institute for factory operation and automation) and Koren's RMS concepts formalize a simple idea: reconfiguration must remain fast, predictable, and measurable, with modules and common interfaces. The laws of Factory Physics recall a hard point: the more variability rises, the more lead time and WIP rise if management remains unchanged. Without discipline, the "liquid" workshop becomes an unstable workshop.
IV. The flexibility matrix: 4 axes, 4 decisions, 4 sets of indicators
A flexible factory wins when it links each axis to a design decision and a numerical target. Magnitudes depend on the sector, so it is better to reason in ranges and hypotheses.
Volume flexibility — Symptom: overtime on one side, underloaded machines on the other. Lever: reconfigurable resources and organized multi-skilling. Indicators: OEE at the bottleneck, service rate, unit cost across different volumes.
Mix flexibility — Symptom: lead time exploding as diversity increases. Lever: SMED, robust sequences, standardized tooling interfaces. Indicators: changeover time, batch size, WIP, throughput time. Common target: 30 to 60% reduction in changeover time at bottleneck workstations.
Product flexibility — Symptom: new reference requiring a new workstation or end-of-line rework. Lever: process platforms, multi-purpose workstations, strict cycle time management. Indicators: available capacity at the bottleneck, scrap rate, time to production.
Process flexibility — Symptom: dead zones becoming informal storage areas. Lever: modular cells, priority rules, materials handling sized for peak variability. Indicators: WIP by zone, lead time compliance, transfer time. Common target: 15 to 35% reduction in WIP on a stabilized flow.
V. Technology: accelerating an uncontrolled flow mostly accelerates chaos
Robotics and automation bring stability, but they require controlled flows and reliable data. A flexible factory starts by setting the rules, then mechanizes what remains stable. Modularity lives or dies on interfaces: standard tooling, normalized utilities, common data. Robotization becomes profitable when gesture variability remains bounded, when throughput stabilizes, and when quality costs exceed the cost of the robot; semi-automation remains relevant when the mix changes frequently.
VI. Dynamic simulation: the co-pilot of the industrial master plan
Why simulate before investing
Dynamic simulation tests a layout and workshop rules across multiple scenarios, so it avoids locking in a single assumption. It incorporates mix variations, breakdowns, changeover times, and limited human resources. A digital twin turns a debate into a decision, because it displays assumptions and their consequences, and makes trade-offs between throughput, WIP, and variability visible. It also helps size buffers, materials handling, and transfer zones without drifting toward overcapacity.
Mini-case 2: capacity ramp-up with avoided CAPEX and stabilized lead times
What: a site needed to increase throughput on a product family with an unstable mix, and was hesitating between buying a machine or reorganizing the flow.
How: a dynamic simulation compared several scenarios, including adding multi-skilling at the bottleneck, changing the routing, and reducing batch sizes.
Impact: the site avoided a short-term CAPEX and stabilized throughput time, because WIP remained bounded and the bottleneck remained protected. The gain came from upstream-downstream synchronization, not from local speed.
VII. Transforming an existing factory without a long shutdown: a step-by-step trajectory
Transforming an existing site is managed in phases, with pilot zones and planned switchovers. The bottleneck must remain protected, as it funds the transition through the revenue it maintains. Quick wins often focus on SMED, reducing waiting times, and eliminating unnecessary transfers: they reduce WIP, free up working capital (WCR), and fund the next step.
Mini-case 3: workshop re-layout with space gains and throughput time reduction
What: a workshop was saturating in space and accumulating WIP between zones, with long transfers.
How: a re-layout by cells reduced distances, then bounded buffers and priority rules eliminated informal storage.
Impact: usable floor space increased without extension and throughput time dropped noticeably, because the flow remained continuous. The impact varies depending on the starting point, but the mechanism is robust.
VIII. How to decide: a step-by-step method
Flow and constraint diagnosis: identify bottlenecks, waiting times, and rework loops with reliable data — cycle times, changeover times, major stoppages, WIP by zone.
Segmentation by families: group products by routing and quality constraints to reveal pooling opportunities and avoid oversizing for a few rare cases.
3-5 year scenarios: volumes, mix, new products, supplier risks — a few auditable hypotheses, not dozens of parameters.
Architecture choice: compare dedicated lines, modular cells, and hybrids with explicit reconfiguration rules and quantified trade-offs.
Field validation plan: minimum data, tests, stop criteria, and shared governance between production, maintenance, supply chain, and quality.
IX. Daily management: making flexibility "livable" for teams
Daily operations require execution standards: changeovers with stable sequences and measured times, operator dispatch rules toward the bottleneck, batch sizes compatible with SMED and the service level. The decisive KPIs include OEE, OTD (On-Time Delivery), FPY (First Pass Yield), WIP, throughput time, unit cost, and energy per part. An unreliable KPI ruins the decision, even if the tool remains performing.
X. Pitfalls and countermeasures: 5 errors that ruin flexibility
Confusing flexibility with overcapacity: overcapacity ties up capital and masks flow problems. Countermeasure: reconfigurable capacity, multi-skilling, and tested scenarios.
Automating too early: automating an unstable flow accelerates WIP accumulation. Countermeasure: diagnosis and workshop rules first, targeted automation afterward.
Forgetting interfaces and maintainability: without tooling and utility standards, each cell move becomes a costly mini-project. Countermeasure: integrate maintainability into the design.
Neglecting training and multi-skilling: improvised multi-skilling creates errors and tension. Countermeasure: training plan, clear authorizations, dispatch rules.
Managing without rules or reliable indicators: the workshop slips into "fire-fighting" mode. Countermeasure: a few KPIs maintained daily and problem-solving loops with clear responsibilities.
XI. Performance and carbon: when the modular factory also reduces waste
When WIP falls, the workshop reduces handling, transfers, and rework, so it consumes less energy per part. When scrap falls, the company also reduces material and the associated unnecessary operations. ROI is often built on three levers: lower WIP that frees up working capital, reduced changeover times that free up capacity, and stabilized service rate that avoids purchases. A flexible factory links performance and resource efficiency through measurable mechanisms, not slogans.
FAQ — Flexible factory
What is a flexible factory?
A flexible factory is a site capable of adapting volumes, mix, and variants without major drift in lead time, quality, or unit cost. It relies on reconfigurable modules, standardized interfaces, and management rules that bound WIP. It is assessed on indicators such as throughput time, WIP, OEE, and service rate. It is designed through scenarios rather than a single assumption.
What is flexible production?
Flexible production describes the ability to manufacture multiple references or families with controlled changeover costs. It relies on reducing changeover times, stabilizing quality, and planning that is consistent with variability. It aims for stable customer service despite a changing product mix.
What are the key characteristics of a flexible factory?
The main characteristics include modular cells, common tooling and utility interfaces, and a structured multi-skilling organization. Management bounds WIP and protects the bottleneck. Investment decisions rely on auditable scenarios and hypotheses.
What are the technological levers of a flexible factory?
Levers include targeted robotics, traceability, tooled scheduling, and interface standardization. RMS systems strengthen reconfigurability when modules remain interchangeable. The prerequisite remains reliable data and maintenance standards.
How do you transform an existing factory into a flexible factory without stopping production?
The transformation is managed in phases, with a pilot zone and then successive switchovers. The bottleneck is protected first in order to maintain throughput and fund the transition. Initial gains often come from SMED, reducing waiting times, and simple priority rules. Dynamic simulation helps prepare each switchover and limit operational risks.
How does a flexible factory reduce bottlenecks and stabilize flows?
It avoids isolated local optimization and synchronizes upstream and downstream. It bounds WIP by zone, thus preventing silent congestion. It protects the constrained workstation through scheduling rules, adapted buffers, and bottleneck-oriented multi-skilling.
How does a flexible factory use a digital twin to make faster decisions?
The digital twin compares architectures and workshop rules across multiple mix and volume scenarios. It makes assumptions explicit, thus reducing opinion-based debate. It quantifies impacts on WIP, lead time, OEE, and service rate, with parameter traceability.
How does a flexible factory secure investments and capacity ramp-up?
It favors reconfigurable modules over monolithic CAPEX. It validates the master plan through simulation before purchase, thus avoiding oversizing. It funds capacity ramp-up through gains on changeovers and WIP.
What ROI can be expected from a flexible factory?
ROI often comes from freed-up capacity, lower WIP, and reduced lead time. A 15 to 35% drop in WIP on a stabilized flow can quickly free up working capital. A 30 to 60% reduction in changeover times at constrained workstations can avoid a purchase and improve service rate. ROI becomes credible when hypotheses and scenarios remain auditable.
What methodology should be deployed to standardize a flexible factory across multiple sites?
Multi-site standardization relies on common modules, identical interfaces, and a shared data repository. Each site retains local latitude, but applies the same measurement and management rules. Simulation helps validate that a module remains performing in multiple contexts.
Which KPIs should be monitored to balance flexibility, productivity, and costs in a flexible factory?
Priority KPIs include OEE, OTD, FPY, WIP, throughput time, and unit cost. Energy per part completes the trade-off analysis when variability increases handling and rework. Monitoring must distinguish the bottleneck workstation from the rest of the flow, as trade-offs are often made there.
What daily management standards keep a flexible factory alive?
Standards cover changeovers, priority rules, buffer management, and operator dispatch. They include problem-solving routines with clear responsibilities and data discipline on cycle times and stoppages. A flexible factory endures when these standards reduce urgency and stabilize quality.
Dillygence combines industrial expertise and digital twin to test flexibility scenarios before investment and accelerate robust decisions, from the master plan to flow management.
Incremys
