Manufacturing Systems and Performance Analysis

Introduction

Manufacturing automation is not simply the replacement of manual labor with machines. It is the systematic design, integration, measurement, and improvement of resources that transform raw materials into products. A manufacturing system includes machines, tools, fixtures, material-handling devices, operators, control systems, information flows, maintenance practices, and operating policies. Its performance depends on how these elements interact over time.

Performance analysis provides the quantitative foundation for decisions such as:

• whether a process has sufficient capacity;
• where a bottleneck is located;
• how much work-in-process is appropriate;
• whether automation will improve productivity;
• how equipment reliability affects delivery performance; and
• how production lines should be balanced.

The central challenge is that manufacturing systems are dynamic. Demand varies, machines fail, setups consume time, batches differ in size, and quality losses reduce effective output. Consequently, nominal machine speed is not equivalent to actual system performance.

Learning Objectives

After studying this chapter, the reader should be able to:

1. Classify manufacturing systems and describe their principal components.
2. Calculate cycle time, throughput, capacity, utilization, work-in-process, and lead time.
3. Apply Little’s Law under appropriate steady-state conditions.
4. Explain availability, reliability, maintainability, and overall equipment effectiveness.
5. Identify bottlenecks and explain their effect on system throughput.
6. Balance an assembly or machining line subject to precedence and workstation constraints.
7. Evaluate automation opportunities using technical, economic, operational, and human factors.
8. Interpret performance data critically and recognize the limitations of aggregate indicators.

1. Manufacturing Systems as Transformation Systems

A manufacturing system transforms inputs into outputs. Inputs include materials, energy, labor, tooling, capital equipment, process instructions, and production information. Outputs include conforming products, scrap, rework, emissions, data, and sometimes by-products.

A useful system boundary distinguishes three interacting flows:

1. Material flow: movement and processing of parts, subassemblies, and products.
2. Information flow: orders, schedules, recipes, inspection results, machine states, and traceability records.
3. Control and feedback: decisions that adjust production in response to demand, disturbances, quality results, or equipment condition.

Typical system types include:

No classification is universally superior. A job shop provides flexibility but often has longer and more variable lead times. A dedicated line can achieve low unit cost and high throughput but may be inflexible when product demand or design changes.

A manufacturing system can also be described by its layout: process layout, product layout, cellular layout, fixed-position layout, or hybrid layout. Layout affects travel distance, queue formation, supervision, material handling, and the ease of introducing automation.

2. Time, Flow, and Capacity Measures

2.1 Cycle time and processing time

Processing time is the time during which a part is actively being transformed. Cycle time is commonly used in two related ways: the interval between completed units leaving a process, or the time required for one unit to pass through a workstation. These meanings should be stated explicitly because they are not always identical.

For a repetitive operation with one machine producing one unit at a time, the ideal cycle time may be approximated by:

where CT is the average cycle time per unit, tₚ is processing time per unit, tₛ is setup time, and B is the batch size. The equation assumes setup time is distributed evenly over the batch and excludes failures, waiting, and material shortages.

For a production line, the required takt time is determined by demand:

where Takt is takt time, Tₐ is net available production time during the planning period, and D is required good output during that period. Both quantities must use the same time unit. Takt is a demand rate, not the actual cycle time. To satisfy demand reliably, the effective line cycle time should generally be no greater than takt time.

2.2 Throughput and capacity

Throughput is the rate at which completed, usually conforming, units leave a system. If N good units are produced during elapsed time T:

where TH is throughput, measured for example in units per hour.

If a resource operates for available time Tₐ and has an effective cycle time CTₑ, its ideal capacity is:

This expression assumes continuous availability and a stable product mix. Effective cycle time should include realistic allowances for setups, minor stoppages, and other planned losses where appropriate.

For serial resources, the ideal line capacity is governed by the slowest effective station:

where CTᵢ is the effective cycle time of station i. The station with the largest effective cycle time is the capacity bottleneck. In practice, the constraint may shift because of downtime, changeovers, labor absence, quality problems, or product mix.

Worked Example 1: Capacity and utilization

A machining center is available for 7.5 hours per shift. Its processing time is 2.4 minutes per part. Setup and scheduled minor-loss allowance average 0.6 minutes per part. During one shift it produces 165 conforming parts.

The effective cycle time is:

Available time is 450 minutes, so estimated practical capacity is:

Because the observed output is 165 parts, the assumptions behind the estimate are inconsistent: either the loss allowance was not actually incurred, the available time was greater, multiple parts were produced per cycle, or the data were recorded incorrectly. This illustrates an important professional principle: performance measures must be reconciled with the physical process before decisions are made.

If the actual effective cycle time was 2.5 minutes, capacity would be 450 ÷ 2.5 = 180 parts per shift, and output utilization relative to that capacity would be 165 ÷ 180 = 91.7%.

2.3 Utilization and efficiency

Utilization measures the proportion of available capacity used:

or, when output is measured over the same period:

Utilization is not automatically a measure of excellence. Very high utilization can create queues, reduce responsiveness, and leave no recovery time for variation. In a variable system, operating every resource continuously may increase lead time even when apparent utilization improves.

Efficiency often compares actual performance with a standard. Since organizations define efficiency differently, the numerator and denominator must be documented. Confusing utilization, efficiency, productivity, and OEE is a common source of misleading reports.

3. Work-in-Process, Lead Time, and Little’s Law

Work-in-process (WIP) consists of units that have entered production but are not yet complete. It includes parts waiting in queues, parts being processed, parts in inspection, and parts temporarily held between operations.

Lead time is the elapsed time from release or order entry to completion. It may include queue time, processing time, setup time, transport, inspection, rework, and administrative delay. In many factories, processing time is a small fraction of total lead time; waiting is the dominant component.

For a stable system over a sufficiently long observation period, Little’s Law states:

where WIP is average work-in-process in units, TH is average throughput in units per unit time, and LT is average flow or lead time in the same time unit. Equivalent forms are:

and

The law applies to a stable system in which average WIP does not continually increase or decrease, and the measured input, output, and inventory belong to the same system boundary. It is not a formula for predicting transient behavior immediately after a start-up or before a shutdown.

Worked Example 2: Little’s Law

A fabrication cell averages 240 parts in the system and produces 30 good parts per hour. Its average lead time is:

If a scheduling improvement reduces average WIP to 150 parts while throughput remains 30 parts per hour, the predicted average lead time becomes 150 ÷ 30 = 5 hours. The reduction is achieved without increasing machine speed; it results from reducing waiting and excess release into the system.

Little’s Law is powerful but descriptive. It does not identify the cause of high WIP. Causes may include an overloaded bottleneck, large transfer batches, unreliable equipment, poor scheduling, or quality holds.

4. Bottlenecks and Constraint-Based Analysis

A bottleneck is a resource or policy that limits the throughput of the system. The physical bottleneck is often the station with the smallest capacity, but the effective constraint can be broader. Examples include a skilled inspection operation, a limited furnace, a shared tool, a certification step, or a test laboratory.

Bottleneck analysis should proceed through the following logic:

1. Define the product family, demand period, and system boundary.
2. Measure actual effective capacity at each resource.
3. Locate queues and identify the resource that restricts completed output.
4. Protect the bottleneck from starvation and blockage.
5. Exploit existing capacity before purchasing new equipment.
6. Subordinate upstream release and downstream activity to the constraint.
7. Elevate the constraint through redesign, additional shifts, tooling, parallel equipment, or automation.
8. Reassess because the constraint may move.

A local improvement at a non-bottleneck station may produce no increase in total throughput. For example, doubling the speed of a station that supplies a slower downstream station may merely increase WIP. A valid automation proposal must therefore evaluate the whole flow, not only the targeted machine.

5. Availability, Reliability, and Maintainability

5.1 Availability

Availability is the proportion of scheduled or required time during which equipment is capable of performing its intended function. A basic operational availability measure is:

where uptime is functioning time and downtime includes failure, repair, adjustment, and other defined loss categories. Planned maintenance may be included or excluded depending on the organization’s definition; the definition must be stated.

For a repairable system with mean time between failures (MTBF) and mean time to repair (MTTR), an idealized steady-state availability is:

This assumes failure and repair behavior are sufficiently stable and that MTBF and MTTR are measured consistently. It does not account for planned shutdowns, waiting for spares, or external production interruptions unless they are included in the time definitions.

5.2 Reliability

Reliability is the probability that an item performs its required function for a specified time under specified conditions. For a constant failure rate λ, the exponential reliability model is:

where R(t) is reliability at time t, λ is the constant failure rate, and e is the base of natural logarithms. This model assumes a constant hazard rate and is not appropriate for every life cycle. Early-life failures and wear-out failures may require other models, such as Weibull analysis.

For independent components arranged in series, the system operates only if every component operates:

For independent identical components in parallel, the system operates if at least one operates:

These formulas assume statistical independence and clearly defined system success criteria.

5.3 Maintainability and maintenance strategy

Maintainability concerns how easily and quickly equipment can be restored. Design choices such as modular components, diagnostic sensors, access for technicians, standardized fasteners, and digital maintenance instructions can reduce MTTR.

Maintenance strategies include:

• Corrective maintenance: repair after failure.
• Preventive maintenance: service at planned intervals.
• Predictive or condition-based maintenance: act on measured condition.
• Proactive maintenance: remove underlying causes of recurring failure.

Automation increases the importance of maintenance competence. A highly automated line may require fewer direct operators but more expertise in controls, sensors, networks, robotics, data interpretation, and functional safety.

6. Overall Equipment Effectiveness

Overall equipment effectiveness (OEE) combines three loss dimensions:

where:

The factors are expressed as decimals or percentages consistently. OEE assumes the ideal cycle time is correctly defined and that counts and time categories are reliable.

Worked Example 3: OEE

A line has 480 minutes of planned production time. It runs for 420 minutes, produces 800 units, and 760 are good. The ideal cycle time is 0.5 minute per unit.

Therefore:

OEE is a diagnostic indicator, not a universal productivity target. A line can achieve high OEE while producing the wrong product, running when demand is absent, or accumulating inventory. Conversely, a low OEE may result from planned product changeovers that are strategically necessary. OEE should therefore be analyzed with safety, quality, delivery, cost, and demand data.

7. Line Balancing

Line balancing assigns tasks to workstations so that precedence requirements are satisfied and station workloads are distributed as evenly as practical.

For a line with required cycle time CT and total task time Σtᵢ, the theoretical minimum number of stations is:

where ceil denotes rounding upward to the next whole station. This is a lower bound; precedence constraints may require more stations.

Line efficiency is:

where N is the actual number of stations. Balance delay is:

These measures assume a common cycle time and comparable station availability. They do not capture every practical issue, such as ergonomic risk, skill variation, stochastic downtime, or parallel machines.

Worked Example 4: Basic line balance

Four tasks have times of 2, 3, 4, and 3 minutes. Their precedence is A before B, B before C, and A before D. Required cycle time is 6 minutes.

Total task time is 12 minutes, so:

A feasible assignment is Station 1: A + B = 5 minutes; Station 2: C + D = 7 minutes. This is infeasible because Station 2 exceeds the 6-minute cycle time. A feasible three-station assignment is Station 1: A + B = 5; Station 2: C = 4; Station 3: D = 3. Efficiency is:

The example demonstrates that the theoretical minimum is not necessarily achievable. Precedence and task indivisibility matter.

Automation can improve balance by combining tasks, reducing task times, or enabling parallel stations. However, automation may also create a new bottleneck if automated cycle time, loading, inspection, or recovery time is underestimated.

8. Evaluating Automation Opportunities

Automation should be evaluated as a system intervention rather than as a machine purchase. A structured evaluation includes:

Technical criteria

• repeatability and process capability;
• required cycle time and takt compliance;
• product variety and changeover complexity;
• material presentation and part orientation;
• integration with existing controls and information systems;
• safety, guarding, and maintenance access;
• availability of skills, spares, and service support.

Economic criteria

Investment should be compared with total cost of ownership, not only purchase price. Relevant costs include integration, tooling, programming, training, energy, maintenance, software licenses, downtime during installation, and eventual disposal.

A simple payback estimate is:

This is a screening measure, not a complete investment appraisal. It ignores the time value of money, taxes, risk, and benefits occurring after the payback period. For larger projects, net present value and sensitivity analysis are preferable.

Organizational and human criteria

Automation changes work content. It may eliminate hazardous or monotonous tasks while increasing requirements for monitoring, diagnosis, programming, and maintenance. Employees should be involved early, and training should precede commissioning. Ethical implementation avoids treating workers as disposable costs and considers redeployment, skill development, accessibility, and safe human–machine collaboration.

Data and control criteria

Connected equipment generates data about states, alarms, energy, quality, and maintenance. Data should have a defined purpose, ownership, access policy, retention period, and cybersecurity protection. Artificial intelligence can support anomaly detection, predictive maintenance, scheduling, and visual inspection, but models require representative data, validation, monitoring for drift, and human oversight. An inaccurate model can create false alarms, missed defects, or unsafe maintenance decisions.

9. Applied Case: Illustrative Automated Machining Cell

Background

A fictional manufacturer produces aluminum housings in a four-stage cell: CNC machining, washing, dimensional inspection, and packaging. Demand is 420 good housings per eight-hour shift. Planned production time is 450 minutes after breaks and scheduled meetings.

Problem

The company plans to automate loading at the CNC machine because operators report fatigue. Management expects the investment to increase output, but the inspection station has a cycle time of 1.25 minutes per housing, while the CNC machine has an effective cycle time of 0.90 minutes and the washer has 1.00 minute.

Analysis

The demand takt is:

Inspection is the slowest station at 1.25 minutes. Its ideal capacity is 450 ÷ 1.25 = 360 units per shift, below demand. Therefore, automating CNC loading alone cannot achieve the required output. It could increase CNC capacity and create a larger queue before inspection.

The team measures inspection losses and finds that 12% of planned inspection time is lost to changeover, rechecking, and waiting for calibration. A revised inspection method, fixture standardization, and a second inspection position are considered. Two parallel inspection positions, each with 1.25-minute cycle time, provide a combined ideal capacity of:

This does not mean 720 good units will be achieved; staffing, quality, upstream capacity, and availability must still be verified. The CNC loading automation is then evaluated separately for ergonomic risk reduction, stable feeding, and possible labor redeployment rather than being justified by output alone.

Approach and outcome

The company first improves inspection capacity and standardizes data collection. It then installs semi-automated CNC loading with a guarded fixture, error-proof part orientation, and manual recovery mode. The project is accepted because it reduces ergonomic exposure, supports consistent loading, and does not overload the downstream constraint. Operators are trained in fault recovery and inspection-system verification.

Learning points

1. The bottleneck must be identified before automation is specified.
2. A local speed increase can increase WIP without increasing throughput.
3. Safety and ergonomic benefits may justify automation even when capacity does not increase.
4. Capacity calculations must include quality, downtime, changeover, and parallel-resource assumptions.
5. Automation projects require process ownership, training, maintenance planning, and data governance.

10. Measurement Practices, Limitations, and Best Practices

Performance measures are only as reliable as their definitions and data collection. A robust measurement system should:

• define the system boundary and observation period;
• distinguish planned, unplanned, and external downtime;
• separate good units, scrap, and rework;
• record product mix and batch size;
• use synchronized time stamps;
• validate automated counts against physical counts;
• report averages together with variation and exceptional events;
• prevent metric manipulation through transparent rules; and
• connect indicators to decisions and improvement actions.

Averages can conceal instability. Two lines may have the same average cycle time but very different variation, queue behavior, and delivery reliability. Engineers should examine distributions, downtime Pareto charts, changeover duration, first-pass yield, queue length, and time-series behavior.

Safety must take priority over throughput. Automated systems require risk assessment, safeguarding, emergency-stop design, lockout/tagout procedures, safe control logic, and compliance with applicable occupational safety and machinery regulations. Cybersecurity is also a production concern because unauthorized access to programmable controllers or manufacturing networks can affect safety, quality, and availability.

Chapter Summary

Manufacturing systems are integrated networks of material, information, equipment, people, and control. Their performance is described using cycle time, takt time, throughput, capacity, utilization, WIP, and lead time. Little’s Law links average WIP, throughput, and lead time in a stable system, but it does not explain the cause of delay.

The bottleneck determines practical system throughput, so automation must be evaluated across the complete process. Availability measures operating readiness; reliability describes the probability of successful operation over time; maintainability concerns restoration speed. OEE combines availability, performance, and quality, but it must be interpreted with demand, safety, delivery, and cost information.

Line balancing distributes tasks subject to cycle-time and precedence constraints. Automation opportunities should be assessed technically, economically, organizationally, ethically, and from a safety and cybersecurity perspective. The best improvement is not necessarily the fastest machine; it is the intervention that improves the system’s constrained performance while preserving quality, safety, flexibility, and sustainable human work.

Key Terms

Manufacturing system: Integrated arrangement of resources that transforms inputs into products.

Cycle time: Time interval associated with producing successive units or completing a workstation operation.

Takt time: Available production time divided by required demand.

Throughput: Rate at which completed units leave a system.

Capacity: Maximum achievable output under specified conditions.

Utilization: Proportion of available capacity that is used.

Work-in-process: Units released into production but not yet complete.

Lead time: Elapsed time from release or order entry to completion.

Little’s Law: Relationship WIP = throughput × lead time for a stable system.

Bottleneck: Resource or policy that constrains system throughput.

Availability: Proportion of defined time during which equipment is operationally capable.

Reliability: Probability of performing a required function for a specified time and condition.

Maintainability: Ease and speed with which equipment can be restored.

MTBF: Mean time between failures for a repairable item.

MTTR: Mean time to repair or restore equipment.

OEE: Product of availability, performance, and quality.

Line balancing: Assignment of tasks to stations while respecting cycle time and precedence.

Precedence constraint: Required order in which tasks must be performed.

Automation: Use of control systems and technology to perform tasks with reduced direct human intervention.

Total cost of ownership: All significant acquisition, operation, maintenance, integration, and disposal costs.

Review Questions

Short Answer Questions

1. Distinguish processing time, cycle time, and takt time.
2. State Little’s Law and identify its principal applicability condition.
3. Why does the slowest station in a serial line commonly determine capacity?
4. Define availability, reliability, and maintainability.
5. Name the three components of OEE.

Long Answer Questions

1. Explain how material flow, information flow, and feedback interact in an automated manufacturing system.
2. Discuss why high utilization can increase lead time and WIP in a variable production system.
3. Explain the assumptions and limitations of OEE as a performance measure.
4. Describe a systematic procedure for identifying and improving a manufacturing bottleneck.
5. Discuss the technical, economic, human, ethical, and safety factors that should be considered before automating a process.

Analytical / Application Questions

1. A cell contains 180 units of average WIP and produces 24 good units per hour. Calculate average lead time and explain two operational causes that could make the lead time worse even if average throughput remains unchanged.
2. A line has task times of 5, 4, 3, 2, and 4 minutes, with a required cycle time of 6 minutes. Calculate the theoretical minimum number of stations, explain why precedence constraints may require more stations, and propose a method for evaluating a feasible assignment.
3. An automated station has 480 minutes of planned production time, 420 minutes of operating time, 900 total units, an ideal cycle time of 0.4 minute, and 855 good units. Calculate availability, performance, quality, and OEE. Interpret whether improving speed alone is necessarily the best intervention.

Multiple Choice Questions with Answers

Question 1

A production line has station cycle times of 0.8, 1.1, and 0.9 minutes per unit. Which station is the capacity bottleneck under stable conditions?

A. The 0.8-minute station

B. The 0.9-minute station

C. The 1.1-minute station

D. All stations have equal capacity

Correct Answer: C. The 1.1-minute station

Question 2

A system has an average WIP of 120 units and throughput of 20 units per hour. According to Little’s Law, what is the average lead time?

A. 0.167 hour

B. 6 hours

C. 20 hours

D. 2,400 hours

Correct Answer: B. 6 hours

Question 3

A line has 420 minutes of available time and demand for 350 good units. What is the takt time?

A. 0.83 minute per unit

B. 1.20 minutes per unit

C. 1.67 minutes per unit

D. 770 minutes per unit

Correct Answer: B. 1.20 minutes per unit

Question 4

Which statement best describes reliability?

A. The percentage of planned time a machine is scheduled

B. The probability that an item performs its required function for a specified time

C. The ratio of good units to total units

D. The time required to complete a setup

Correct Answer: B. The probability that an item performs its required function for a specified time

Question 5

A machine has MTBF of 90 hours and MTTR of 10 hours. Under the stated idealized assumptions, what is its availability?

A. 10%

B. 50%

C. 90%

D. 100%

Correct Answer: C. 90%

Question 6

Which three factors are multiplied to calculate OEE?

A. Cost, delivery, and flexibility

B. Availability, performance, and quality

C. Throughput, WIP, and lead time

D. Reliability, safety, and utilization

Correct Answer: B. Availability, performance, and quality

Question 7

A line has total task time of 18 minutes and a required cycle time of 5 minutes. What is the theoretical minimum number of stations?

A. 3

B. 3.6

C. 4

D. 5

Correct Answer: C. 4

Question 8

Why can improving a non-bottleneck machine fail to increase total system throughput?

A. It always reduces machine reliability

B. It may only create additional WIP before the bottleneck

C. It necessarily increases takt time

D. It eliminates all inspection requirements

Correct Answer: B. It may only create additional WIP before the bottleneck

Question 9

Which condition is required for the simple availability formula MTBF ÷ (MTBF + MTTR) to be a reasonable idealized estimate?

A. Failure and repair behavior are measured consistently and are sufficiently stable

B. The equipment never experiences downtime

C. Demand is exactly zero

D. Every product has a different cycle time

Correct Answer: A. Failure and repair behavior are measured consistently and are sufficiently stable

Question 10

A line operates for 400 minutes during 500 minutes of planned production time. It produces 800 units, with an ideal cycle time of 0.4 minute, and 760 are good. What is its approximate OEE?

A. 60.8%

B. 68.4%

C. 76.0%

D. 80.0%

Correct Answer: B. 68.4%