Compressed-Air Demand Assessment & Load Profiling

REDDY COMPRESSORS

Measure the Work • Separate the Loads • Understand the Peaks • Size From Evidence

QUALITY STATE: Controlled premium engineering draft. Final capacity, pressure, instrumentation and investment decisions require project-specific verification.

Page Control

The compressor should follow the demand—not define it

A compressed-air plant should not be sized from the sum of every tool catalogue value, the motor rating of the existing compressor, a single pressure-gauge reading or the capacity of the machine that happened to be installed previously. Those figures can be evidence, but none is automatically the plant demand.

The output is not merely a CFM number. It is a time-dependent demand profile tied to pressure, production state, measurement quality and operating consequence.

That profile becomes the foundation for compressor capacity, storage, sequencing, pressure control, distribution, temporary-air planning and energy analysis.

1 — Define the Assessment Boundary

Decide what system is being assessed before collecting numbers. A whole factory, production block, compressor room and isolated process can produce different demand profiles. Draw sources, receivers, treatment, major headers, interconnections and significant consumers. Mark where flow and pressure are measured. A measurement without a defined boundary can be numerically correct and still answer the wrong question.

2 — Build a Consumer Register

Create an inventory of significant air users. Record process function, location, required pressure, air requirement and its evidence source, operating pattern, simultaneity, criticality, local regulation/isolation and future status. Grouping small consumers is acceptable when the basis is transparent; false precision is not.

3 — Connected Load Is Not Simultaneous Demand

The sum of all connected consumers is an installed-load total. It becomes a design demand only if those loads genuinely operate together. Establish overlap from machine sequences, schedules, controls, observation or measurement. Avoid a universal diversity factor. Equally, do not assume diversity merely to justify a smaller compressor.

4 — Separate Base, Variable, Peak and Exceptional Demand

Base demand persists through a defined production state. Variable demand moves normally around it. Peak demand is a short higher-flow event that belongs to normal operation. Exceptional demand is unusual. Idle demand occurs without productive operation. Future demand is not yet operating. Keep magnitude and duration visible; an average can hide pressure-critical peaks while a single maximum can overstate sustained capacity.

5 — Time Resolution Changes What You Can See

Logging interval must suit process dynamics and the decision. Long averaging can hide rapid events; excessive sampling can create data without better insight. Preserve raw data where practical and document averaging, filtering and exclusions. A processed graph should remain traceable to its measurement record.

6 — Measure Flow With Its Basis Intact

A flow number is incomplete without units, reference conditions, method and location. FAD, standard-volume flow and actual line volumetric flow can have different bases. Reconcile definitions before comparing measured demand with compressor performance. Unit conversion alone does not reconcile reference conditions.

7 — Pressure Must Be Logged With Demand

Flow without pressure can hide whether useful service was delivered. Log pressure at meaningful source/header and downstream points while demand is measured. A peak with stable point-of-use pressure has a different implication from the same flow with pressure collapse. State measurement location and pressure basis; a compressor controller display is not proof of remote-user pressure.

8 — Link Demand to Production State

Record shift, line state, batch, product family or another useful production indicator. This distinguishes an air-system change from a production change. Lower energy can simply reflect lower output; higher flow can reflect legitimate production. Normalised indicators can help when their denominator and process limits are clearly defined.

9 — Identify Non-Productive Demand

Non-production flow can include leakage, purges, drains, idle-machine consumption, open blowing, abandoned connections or legitimate continuous services. Do not call it all leakage. Identify intended loads first, then classify the remainder using inspection, isolation or supported methods. Estimated leak savings are not realised savings until repairs are completed and verified.

10 — Artificial Demand and Excess Pressure

Some unregulated or pressure-sensitive uses consume more air at higher supply pressure. Raising the whole system to solve one remote pressure problem can therefore increase demand elsewhere. Establish the critical pressure requirement and distribution path before changing source pressure. Do not apply one universal percentage saving to every pressure reduction.

11 — Cyclic Loads Need Event-Based Analysis

Pneumatic machines often consume air in pulses. Average consumption can be modest while a short event produces high instantaneous demand. Record event flow or volume where supported, duration, repetition rate, allowable pressure variation and local storage/distribution.

12 — Use Mass Balance Carefully

For compatible flow bases, a system balance conceptually considers source flow, change in stored air, exports and consumption. Receivers mean source flow and consumer demand need not match instantaneously. During pressure decay stored air contributes; during recovery source capacity replenishes storage. Synchronise timestamps and flow bases before combining measurements.

13 — Estimate Only When Measurement Is Not Yet Available

Preliminary work may rely on manufacturer consumption, cylinder calculations, tool schedules, existing-machine history or analogous production. Keep every assumption visible and label its evidence status. An estimate does not become measured demand because it appears in a spreadsheet.

14 — Future Capacity Requires a Confidence Register

Classify future loads as committed, probable, possible or unknown. Build scenarios rather than one inflated demand number. A base case, committed-growth case and defined expansion case make capacity decisions understandable and allow infrastructure provisions to be separated from immediate compressor purchase.

15 — Representative Measurement Periods

The logging period should capture operating states that materially affect the decision: shifts, batches, changeovers, idle periods, seasonal conditions or maintenance events as applicable. There is no universal number of days that proves representativeness. Keep an event diary so shutdowns, service, temporary equipment and unusual operations can be interpreted correctly.

16 — Demand Duration Curve

A duration curve sorts demand from high to low and shows how much observed time is spent at or above each level. It helps distinguish sustained load from infrequent peaks, but it removes chronology. Use it alongside time-series data and state the dataset and exclusions.

17 — Convert the Profile Into Design Cases

Do not reduce the study to one maximum. Define normal sustained production, normal peak, low-demand production, idle/non-production, critical process, committed expansion and exceptional-event cases as relevant. For each case state flow, pressure, duration/pattern, quality condition, consumers and evidence status.

18 — Capacity Margin Must Have a Reason

A provision can address measurement uncertainty, defined future demand, expected deterioration or another identified need. Record what each margin addresses and avoid overlapping margins at consumer, branch, plant and compressor levels. The goal is not zero margin; it is traceable margin with an engineering reason.

19 — Capacity and Pressure Are Coupled

A plant demand expressed only as flow is incomplete. Candidate compressor performance must be compared at the pressure required after the downstream system is assessed. Do not combine maximum flow and maximum pressure from different operating points. Pressure losses should be addressed as system issues rather than hidden inside an arbitrary compressor-pressure increase.

20 — Validate the Profile Before Procurement

Review the demand profile with production, maintenance, engineering and the people who understand unusual operating events. Reconcile obvious conflicts between calculated loads and measurements. Identify missing evidence and open assumptions. Freeze a dated design-basis version before equipment quotations are compared; later changes should be revisions, not silent replacements.

21 — Core Demand Register

22 — Review Gate

23 — What Must Never Claim

Installed tool CFM automatically equals plant demand.

Average demand alone is sufficient for compressor sizing.

The single highest logger point is automatically continuous design duty.

All non-production airflow is leakage.

One diversity factor applies to every factory.

One logging duration is universally adequate.

CFM values can be compared without checking their reference basis.

Compressor discharge pressure proves point-of-use pressure.

A standard percentage capacity margin is always correct.

Future growth should be hidden inside an unexplained oversized compressor.

A larger compressor automatically corrects a pressure-distribution problem.

Estimated savings are realised savings before verification.

24 — Frequently Asked Questions

Should I add the CFM of every machine?

Only for a case in which those machines genuinely operate together. Build simultaneity from the actual production sequence or measurements.

How long should flow logging continue?

Long enough to represent the operating states relevant to the decision. The correct period depends on shifts, batches, seasons, changeovers and unusual events.

Can compressor motor kW tell me plant air demand?

No. Motor rating is equipment information, not a direct measurement of plant airflow.

Use controlled preliminary estimates from equipment data, cycles and operating schedules, clearly label assumptions, and plan measurement where the decision warrants it.

Should peak demand determine compressor size?

Not automatically. Retain peak magnitude and duration, then assess source response, storage, distribution and allowable pressure variation.

Can night or weekend flow be treated as leakage?

Only after legitimate continuous loads, purges, drains and other intended consumption are identified.

How should future expansion be included?

Separate committed, probable and possible loads and compare scenarios instead of adding an unexplained blanket percentage.

Why log pressure with flow?

Because the same airflow can represent acceptable or unacceptable service depending on pressure at the consumer and the behaviour of the distribution system.

What is the most useful final output?

A frozen, revision-controlled set of design cases containing flow, pressure, duration/pattern, quality requirement, evidence status and future assumptions.

The next engineering pages should reconcile airflow definitions and compressor capacity against the demand profile before moving deeper into pressure, treatment, storage and distribution.

26 — Final CTA

DO NOT BUY CAPACITY BEFORE YOU UNDERSTAND DEMAND.

Build the boundary. Identify the consumers. Measure flow and pressure. Record production state. Separate sustained load from peaks. Classify future demand. Convert the evidence into design cases. Then compare equipment.

Secondary CTAs: Submit Flow & Pressure Data • Review Compressor Capacity • Plan a Plant Air Audit • Review Expansion Demand • Request Temporary Air Assessment

Checkpoint Record

Future revisions must be separately versioned; no silent shortening, replacement or regeneration.

V2 adds deeper engineering substance, calculations, worked examples, audit logic and publishing controls without deleting V1.

V2 STATUS: STAGING READY — TECHNICAL / EDITORIAL VERIFICATION GATED

V2.1 — Demand Assessment Is a Measurement Problem Before a Sizing Problem

A defensible compressor selection begins with a defensible demand model. Define what is measured, where, under which production state, and whether the dataset represents the plant's variability.

The objective is not one maximum number. It is a time-dependent description supporting capacity, storage, controls, distribution, treatment and reliability. Preserve the distinction between measured demand, calculated demand, manufacturer data, estimates and future assumptions.

V2.2 — Controlled System Boundary

Map every source feeding the study boundary, every export, treatment item, receiver, major header and material consumer. A meter downstream of storage can differ instantaneously from compressor delivery because stored air is charging or discharging.

Cross-connected compressor rooms, purge dryers and temporary sources can invalidate a simple source-versus-demand comparison. Every important measurement needs a physical boundary.

V2.3 — Consumer Register

Record asset/process, location, point-of-use pressure, air requirement and source, operating pattern, cycle information, simultaneity, criticality, regulation and future status.

Preserve whether manufacturer data is average, maximum, cycle volume or another definition. Calculated cylinder demand should retain geometry, stroke, cycle rate, pressure basis and assumptions. Estimates remain estimates until verified.

V2.4 — Connected Load, Diversity and Simultaneity

Connected load is the arithmetic total of installed users. Simultaneous demand is what operates together in the relevant case. Do not use a universal diversity factor.

Build an operating-state matrix from machine sequence, PLC timing, production schedules, observation and measurement. This also prevents under-sizing when multiple machines can purge, index or actuate together.

V2.5 — Demand Classes

Separate base, variable, peak, exceptional, idle/non-production and future demand. Base informs sustained capacity; variability informs trim/sequencing; peaks inform storage and pressure stability; idle demand supports leakage analysis; future demand supports expansion.

Mixing them into one maximum destroys engineering information.

V2.6 — Time Resolution

Sampling interval must suit process dynamics. Slow logging can hide a five-second event; excessive frequency can add complexity without decision value.

Preserve raw data and document aggregation. If one-second data becomes one-minute averages, do not present the averaged series as though it captured instantaneous peaks.

V2.7 — Flow Measurement Quality

Record meter technology/model, range, pipe/location, installation requirements, orientation, units, reference conditions and calibration/verification status.

Confirm whether the instrument displays actual or compensated/reference flow. A precise meter configured to the wrong basis can generate repeatable but misleading data.

V2.8 — Synchronise Pressure With Flow

Log source/header and representative downstream pressure on aligned timestamps. A 30 m³/min event with stable point-of-use pressure is different from the same flow with pressure collapse.

The second condition may involve source response, storage, distribution restriction or multiple effects. Flow alone cannot show whether useful service was maintained.

V2.9 — Production Context

Attach shift, line state, batch, product, machine count or output to the dataset. Higher airflow may simply reflect higher production; lower energy may reflect downtime.

Where normalisation is useful, define the denominator and ensure unlike products/states are not forced into one misleading KPI.

V2.10 — Idle Demand and Leakage

Night/weekend flow is not automatically leakage. Identify legitimate continuous loads, purges, drains and standby services first.

Use isolation or area testing where practical. A leak register should preserve location, quantification method, repair status and post-repair verification. Estimated savings are opportunities, not realised savings.

V2.11 — Artificial Demand

Unregulated or pressure-sensitive uses can consume more air at higher pressure. Do not apply one universal pressure-to-flow saving percentage.

Identify unregulated users, map critical pressure, correct restrictions and test controlled pressure changes while monitoring production performance.

V2.12 — Cyclic Demand

For cyclic consumers, volume per event, event duration and repetition rate may be more useful than one average. Conceptually, Q_avg = V_event × N cycles per unit time, on a consistent reference basis.

Instantaneous event demand can be much higher than Q_avg. That distinction drives compressor response, local storage and distribution decisions.

V2.13 — Cylinder Demand Estimation

Preliminary cylinder demand can be derived from swept volumes, stroke, cycles and pressure/reference assumptions. Double-acting cylinders require cap-end and rod-end volumes.

Dead volume, valve behaviour and leakage mean simplified calculations remain engineering estimates unless validated. Record every geometry and operating assumption.

V2.14 — Storage and Mass Balance

Receivers decouple short-term source delivery from consumer demand. During pressure decay stored mass supports consumers; during recovery source capacity replenishes storage.

Conceptually, source inflow minus exports minus accumulation equals consumption for a defined boundary. The accumulation term is not zero while pressure changes.

V2.15 — Demand Duration Curve

Sorting flow from high to low shows how much observed time occurs at or above each level. It distinguishes sustained demand from infrequent peaks.

Sorting removes chronology, so retain the time series. Similar duration curves can conceal very different event sequences and storage/control requirements.

V2.16 — Percentiles Are Descriptors, Not Automatic Design Points

A 95th or 99th percentile can describe a dataset but does not automatically define compressor size. A legitimate critical event may sit outside it; an abnormal event may distort the maximum.

Map excluded events back to production reality and document why each material event is included or excluded.

V2.17 — Worked Demand Example

Illustrative only: sustained production varies around 18–22 m³/min, a recurring 27 m³/min event lasts 15 seconds, low-demand operation is 9–12 m³/min, and idle flow averages 4 m³/min.

Do not immediately select 27 m³/min continuous capacity. Confirm flow basis, simultaneity, storage contribution and point-of-use pressure during the event. The sustained case may follow the upper sustained band while the peak becomes a transient case for storage/control/distribution.

Investigate idle flow separately. If legitimate continuous load is 1 m³/min and verified repairs later reduce avoidable flow, issue a revised post-improvement demand case rather than assuming savings before they occur.

V2.18 — Future Loads as Scenarios

Classify future loads as committed, probable, possible or unknown. Committed equipment with defined demand belongs in a future design case; possible projects remain scenario analysis.

This allows infrastructure to be prepared for expansion without forcing speculative compressor capacity into today's purchase.

V2.19 — Measurement Campaign Design

State the engineering question before choosing instruments. Capacity study, pressure-loss diagnosis, leakage study and energy optimisation can require different sensor placement and sampling.

Create a measurement register with instrument ID, range, accuracy, location, sampling interval, synchronisation, verification and owner. Keep an operating diary for shutdowns, maintenance, temporary compressors and abnormal events. Inspect data for gaps, flat-lined sensors, impossible values and clock drift before removing instruments.

V2.20 — Uncertainty and Sensitivity

Assign evidence quality to material inputs. A calibrated representative measurement is not equivalent to a catalogue estimate.

Where uncertainty can change selection, test sensitivity—for example 20, 22 and 24 m³/min sustained demand—rather than hiding uncertainty inside a blanket 25% margin. This shows which missing evidence deserves further measurement.

V2.21 — Frozen Design-Case Register

Later changes require a revision, not silent downstream edits.

V2.22 — Common Failure Modes in Demand Studies

Failures include logging only a convenient shift, using a meter outside range, ignoring cross-connections, losing flow reference conditions, calling all idle flow leakage, averaging away peaks, using compressor nameplate as demand, double-counting future load and failing to synchronise pressure.

A pre-selection review should explicitly check these failure modes.

V2.23 — Rental and Temporary-Air Demand

Temporary air requires its own duty. Shutdown, breakdown and commissioning states can differ materially from normal production.

Define flow, pressure, quality, duration, connection boundary and operating pattern. Do not simply rent the nominal CFM of the failed compressor; existing treatment, storage and distribution may or may not remain available.

V2.24 — Multi-Compressor Demand Assessment

Overlay flow and pressure trends with machine states: loaded, unloaded, stopped, trim, base and standby. This reveals whether installed capacity is being used effectively.

A low average plant demand can coexist with high energy use if multiple compressors remain unloaded or poorly sequenced. The overlay becomes evidence for later controls engineering.

V2.25 — Acceptance Criteria for the Demand Study

Confirm that measurement represents relevant production states, material consumers are accounted for, flow basis is known, pressure is synchronised, abnormal events are classified, future loads are separated and uncertainties are documented.

Another competent reviewer should be able to reproduce the final design cases from the source evidence without relying on undocumented judgment.

V2.26 — Premium Demand-Assessment Worksheet