Capteurs et instruments pour bancs d'essais hydrauliques : comment choisir les canaux

Hydraulic test bench sensors and instruments should be selected from the results the bench must produce. Pressure, flow, and temperature are common starting points, but cylinder tests may also need displacement or force, while pump and motor tests may need rotational speed, torque, and case-drain flow. The required channels depend on the device under test and the approved method.

For each quantity, define its location, operating range, allowable measurement uncertainty, response, and recording requirements. Then select the sensor, installation, signal conditioning, and acquisition system as one chain. A high-resolution display or an impressive sensor accuracy figure cannot compensate for an unsuitable range, an incorrect measurement point, or a slow data path. The aim is evidence fit for the test decision, not the largest possible collection of instruments.

Sommaire

1. Start with the test result, then define channels
2. Choose pressure type, location, and range together
3. Match flow measurement to fluid and operating range
4. Add temperature and mechanical measurements where needed
5. Distinguish resolution, accuracy, and uncertainty
6. Specify response, sampling, and synchronization
7. Design installation and signal interfaces
8. Validate channels and make a purchasing schedule
9. Foire aux questions

Start with the test result, then define channels

Write down what the test must establish before choosing hardware. “Test a valve” is too broad: pressure drop, leakage, switching response, and endurance can require different instruments and conditions. A channel should have an identified purpose and a named physical measurement location.

Test objective	Quantities commonly needed	Question before selection
Cylinder motion or force	Chamber pressure, position, time, sometimes direct force	Is the required result motion, hydraulic force, or actual output force?
Valve flow characteristic	Flow, upstream and downstream pressure, temperature, command	Are pressure drop and command synchronized with flow?
Performances de la pompe	Outlet and inlet conditions, flow, speed, and possibly torque	Which efficiency or performance quantity is being calculated?
Performances du moteur hydraulique	Pressure difference, flow, speed, torque, and drain conditions	Are all inputs and outputs measured at compatible conditions?
Fuite interne	Small flow, pressure, temperature, and time	Can the channel resolve the required low-flow result?

These are planning prompts, not a mandatory list for every method. See test methods for different hydraulic components for distinctions between DUTs.

Identify calculated results as well as direct measurements. Efficiency, pressure loss, speed, and average leakage may combine several channels. Their validity depends on definitions, timing, and measurement quality, not merely on having each sensor present.

Choose pressure type, location, and range together

Specify gauge, absolute, or differential pressure as required by the method. Gauge pressure references the surrounding atmosphere; absolute pressure uses an absolute reference; differential pressure compares two points. These are not interchangeable labels.

A pump inlet measurement may need a different range and reference from a high-pressure outlet measurement. Measuring a small pressure drop by subtracting two large-range readings can produce an unsuitable uncertainty. A dedicated differential-pressure arrangement may be appropriate, but its common-mode pressure and installation limits must also be checked.

Separate measurement range from permissible overload and pressure-containing capability. Surviving a pressure event does not mean the sensor measured it accurately.

For an original example, assume a 0–400 bar channel has an error specification of ±0.5% of span under defined conditions. That term equals ±2 bar. It represents 20% of a 10 bar reading, but 1% of a 200 bar reading. These percentages describe the assumed error term only, not the full measurement uncertainty.

Read the exact specification: a value may refer to nonlinearity, combined error, a reference temperature, or another defined condition. WIKA's S-20 data sheet separately presents accuracy-related and operating specifications, illustrating the need to read more than the headline figure. WIKA S-20 data sheet.

Match flow measurement to fluid and operating range

Define whether the test needs supplied flow, return flow, branch flow, case-drain flow, or leakage. A meter before a branch does not directly measure the flow entering one downstream component. Name the location and intended direction.

Select a technology suitable for the fluid, viscosity range, temperature, pressure, contamination conditions, and minimum as well as maximum flow. Turbine, positive-displacement, differential-pressure, and other methods have different operating characteristics. There is no universally best hydraulic flow meter.

Check whether performance is stated as a percentage of reading, full scale, or a combination, and over which part of the range it applies. Webtec's DHM information provides an example where the stated flow accuracy depends on the specified measurement range and operating conditions. That product-specific wording should not be generalized to all meters. Webtec DHM measurement information.

A meter sized for the bench's largest flow may be unsuitable for a small leakage test. Consider separate ranges or a dedicated channel where the method requires it. Confirm how changing ranges affects plumbing, calibration, and software identification.

Review pressure loss and installation requirements using the selected meter's manual. Do not assume one straight-pipe rule, orientation, or reverse-flow capability applies to every technology. The instrument must not alter the test condition beyond what the approved method allows.

Add temperature and mechanical measurements where needed

Measure temperature where it helps define the test condition. Reservoir temperature may differ from fluid entering the DUT or passing through a meter. Identify the location, immersion or contact arrangement, response, and uncertainty relevant to the method.

Temperature can affect fluid viscosity and component behavior. A test report should therefore connect the reported temperature to the actual measurement period and location. A slowly responding surface sensor and an immersed sensor can answer different questions.

Use direct force measurement when the required result is actual mechanical output and the method calls for it. Calculating force from pressure and area may omit friction, opposing chamber pressure, alignment effects, and external forces. The calculated quantity must be named correctly.

Position sensing can support stroke, displacement, speed, and stopping measurements, but derived speed depends on time resolution and processing. A simple end switch cannot provide a continuous displacement trace.

For rotary equipment, define speed and torque measurements at the relevant shaft, including direction and sign conventions. Mechanical installation, coupling, alignment, and protection need engineering assessment. Adding a torque sensor changes the mechanical assembly; it is not simply another electrical channel.

Distinguish resolution, accuracy, and uncertainty

A display with more digits may show smaller changes without making the result closer to the quantity being measured. Similarly, a repeatable sensor can consistently produce a biased reading.

Terme	Practical meaning	Selection implication
Gamme	Interval over which the channel is intended to measure	Cover the required points without sacrificing needed low-end performance
Résolution	Smallest indicated or distinguishable increment	Ensure useful changes are visible, without equating digits with correctness
Répétabilité	Agreement under repeated comparable conditions	Evaluate variation separately from systematic effects
Stated accuracy or error limit	Manufacturer-defined performance under specified conditions	Read the definition, included terms, and exclusions
Incertitude de mesure	Quantified uncertainty associated with the result	Assess the complete method and chain for the acceptance decision

Include relevant calibration, drift, environmental, installation, signal-processing, and repeatability contributions in the project's uncertainty evaluation. Avoid adding unrelated percentage specifications without a valid measurement model.

NIST's traceability guidance connects measurement results to a documented calibration chain and uncertainty. Calibration evidence is therefore part of the result's support, not a replacement for evaluating the installed application. NIST metrological traceability.

Agree how uncertainty affects conformity decisions through the Liste de contrôle pour la réception du banc d'essai. The required uncertainty should follow the test tolerance and decision rule, rather than being chosen only because a particular sensor is inexpensive.

Specify response, sampling, and synchronization

The sensor, pressure connection, signal conditioner, filter, acquisition hardware, and software all affect dynamic measurement. A fast sampling setting cannot restore a transient removed by a slow sensor or heavily filtered signal.

Distinguish the acquisition rate from the screen refresh rate and saved-data rate. A system may acquire rapidly but store only one averaged value per second. That could be adequate for a steady condition and unsuitable for a short switching event.

As an illustrative timing example, sampling at 2 Hz means one sample every 0.5 seconds. A 0.1-second event can occur entirely between samples. At 50 Hz, the interval is 0.02 seconds, but capturing more points still does not guarantee an accurate peak or waveform: sensor bandwidth, filtering, triggering, and timing remain relevant.

NI explains the relationship between bandwidth, sampling, and aliasing, including the need to limit unwanted frequency content before conversion. Use that principle to design the measurement chain; do not turn a basic sampling theorem into a universal hydraulic test rate. NI analog-signal acquisition guide.

When combining pressure, flow, valve command, and position, establish whether channels share a clock or have known timing offsets. Otherwise an apparent cause-and-effect relationship may partly reflect acquisition delay. Keep timestamps and processing settings in the record.

Design installation and signal interfaces

Choose accessible, rated measurement points that represent the quantity required by the test. Adapters, dead volumes, long pressure connections, and restrictions can influence dynamic behavior. The mechanical installation must also suit pressure, temperature, vibration, and fluid compatibility.

Do not add a sensor by improvising a connection to a live system. The approved installation and pressure-testing procedures govern modifications and commissioning.

For electrical interfaces, specify supply requirements, signal type, input compatibility, connector pinout, cable routing, shielding or grounding requirements, and isolation where needed. Qualified personnel should design these details for the actual equipment and environment.

Consider channel failure and loss of signal. The acquisition system should distinguish invalid data from a valid zero where the instrumentation permits it. Reports must not silently substitute zeros or previous values without recording that handling.

Measurement channels and protective functions may share information, but their requirements must be defined separately. A general-purpose data logger is not automatically a validated safety controller.

Relate instrumentation choices to the bench's exigences en matière de pression, de débit et de puissance, so measurement equipment remains suitable across the actual operating envelope.

Validate channels and make a purchasing schedule

For each channel, retain its unique identifier, sensor serial number, calibration evidence, installation location, scaling, units, and configuration version. Confirm the displayed and exported values against the approved verification method.

Where a sensor is replaced or its range changes, check the entire chain again. A new calibration certificate cannot prevent an old software scale from being applied to the replacement. Record changes to filters and averaging as well as hardware changes.

Schedule field	Information to specify
Channel and purpose	Quantity, DUT function, direct or calculated result
Location and reference	Named port, shaft, surface, or position datum; gauge or absolute where relevant
Operating range	Minimum, normal, maximum, direction, and transient conditions
Performance	Required uncertainty, response, and relevant error definitions
Hydraulic or mechanical interface	Connection, pressure capability, fluid, temperature, mounting
Electrical interface	Supply, signal, connector, acquisition input, and isolation requirements
Timing and data	Sample rate, bandwidth, filtering, synchronization, storage, and units
Vérification	Calibration conditions, installed-channel checks, and acceptance evidence

Ask suppliers to state exceptions row by row. Two proposed pressure channels are not equivalent if one meets its specification only at a reference temperature while the other is evaluated over the full test environment.

Also specify which processing steps are reversible. If the system stores only smoothed data, later reviewers cannot recover a discarded transient. When raw data are required, preserve them with the calibration and configuration identifiers needed to interpret them. Name calculated columns clearly, especially when a reported value is an average, a filtered peak, or a difference between channels rather than a direct sensor reading.

Before acceptance, run representative steady and dynamic cases using an approved reference method. Confirm that range switching, data export, and calculated results behave as specified. The checks should demonstrate the intended measurement capability rather than merely prove that all channels display numbers.

For recurring use, assign responsibility for calibration review, intermediate checks, and configuration control. Define actions after overload, repair, suspicious results, or sensor replacement. The interval and method should follow the instrument's use and risk, not an invented universal calendar rule.

Foire aux questions

Are pressure, flow, and temperature enough for every bench?
No. They are common channels, but the test objective may require force, displacement, speed, torque, leakage, or other measurements.

Does higher sampling rate always improve accuracy?
No. It can help represent changing signals when the rest of the chain supports them, but it cannot correct unsuitable calibration, range, installation, or bandwidth.

Can one large flow meter measure every leakage rate?
Only if its characterized range and uncertainty support the required result. Many applications need a separate low-flow measurement arrangement.

Should I buy the sensor with the smallest accuracy percentage?
Compare the complete definitions and operating conditions first. The best choice is a channel that meets the test requirement after installation, acquisition, and uncertainty are considered.