Hydraulics & pneumatics: the paper machine's "muscle" and "nerve"
The most easily overlooked yet most fatal lecture in the utility systems: about 80% of hydraulic failures can ultimately be traced to dirty oil.
I. The role of hydraulics and pneumatics on the paper machine
The paper machine itself is "passive" — rolls do not pressurize themselves, doctor blades do not advance or retract themselves, and the fabric-change carriage does not move itself. These motions are driven by two fluid-power systems: hydraulics handles large force, slow speed, and precise force/position control (press line-pressure loading, roll lifting, wire/felt-change cylinders, calender controlled-crown loading); pneumatics handles light, fast, low-force actions (air-tire loading, centering valves, the ejector air supply, valve cylinders, cleaning blow-off). The two complement each other: hydraulics deliver "force", pneumatics deliver "speed".
The empirical rule for selection is clear: use hydraulics when hundreds of kN of force or precise position holding is needed (e.g., calender crown loading); use pneumatics for fast reciprocation, low force, and explosion-proof/dust-proof requirements (e.g., cylinder-operated doors, centering). Using pneumatics at heavy-load positions will "fail to hold", while using hydraulics for fast frequent motions will be "slow to respond and run hot".
In O&M, the common enemy of both systems is contamination and leakage. Statistically about 70%–80% of hydraulic failures can ultimately be traced to oil contamination (solid particles, water, air), while pneumatic failures are mostly due to inadequate air-source treatment (water content, insufficient lubrication, insufficient pressure). So the core of this lecture is not "know how to repair pumps" but "know how to control cleanliness and the air source".
II. Hydraulic power unit composition and oil cleanliness
A paper-machine hydraulic power unit typically contains: reservoir (with level/temperature/breather), main pump (gear or piston pump, constant-pressure variable), relief valve (safety pressure limit), accumulator (absorb pulsation, hold pressure), directional/proportional/servo valves, actuator cylinders, suction/return/high-pressure three-stage filtration, and a cooler. The key point is not the part names but the cleanliness class.
Oil cleanliness is expressed by NAS 1638 or ISO 4406; the smaller the number the cleaner. Servo/proportional valve systems usually require NAS class 7–8 (ISO 16/14/11 level), ordinary systems NAS 9–10. One extra dust-particle class per liter can shorten valve-spool wear life severalfold. The filter differential-pressure alarm is the cheapest warning; running on after it lights up lets particles grind the spool.
Besides particles, water and air in the oil are also invisible killers: water emulsifies the oil, disables additives, and rusts bearings; air causes pump cavitation, system noise, and soft actuation. The reservoir breather should carry a desiccant, and cooler internal leakage (water into oil) must be caught at first sign. Recommended oil operating temperature is 40–55 °C; above 60 °C oxidation accelerates and seals age.
| System type | NAS class (recommended) | O&M implication |
|---|---|---|
| Servo/proportional valve system | NAS 7–8 | Stop pump and inspect on filter differential alarm; replace filter without delay |
| Ordinary hydraulic system | NAS 9–10 | Sample every 6–12 months for particle count + moisture + viscosity |
| Make-up / new oil | No better than NAS 8 | New oil must also be filtered before adding; drummed oil often fails the online requirement |
Oil cleanliness is a "system state", not a "new-oil state". Drain the old oil, add new, but if the filter and reservoir are not cleaned, residual particles immediately contaminate the new oil. Standard practice: drain → clean reservoir-bottom sludge → replace all filters → fill new oil via filter cart → re-test cleanliness after 24 h running. Oil samples must be taken at both return and working points; reading only one leads to misjudgment.
III. Press loading circuit and line-pressure control
Press line pressure is one of the most important process settings on a paper machine: line pressure determines the outgoing dryness, which in turn determines dryer-section steam consumption. Modern presses use a hydraulic proportional valve + pressure-sensor closed loop to control cylinder loading force, then divide by nip width to get line pressure (kN/m). The control loop has three layers: command (set line pressure from DCS) → proportional-valve current → cylinder pressure → pressure-transmitter feedback.
A common fault is "the line-pressure setpoint is right but actual pressure won't rise": first check the accumulator (failed pressure holding drops pressure), then the proportional-valve zero and gain, and finally cylinder internal leakage. Another type is "pressure bouncing": mostly air ingestion in the oil (pump cavitation, return-line leak) or a clogged filter letting the pump's flow pulsation reach the nip directly.
The proportional valve needs periodic zero-point and gain calibration: with power off the command current should be 0 and the corresponding output pressure 0; during calibration compare with the factory curve, and handle any drift beyond 2%. Many "line-pressure drifts" are actually the valve's zero slowly shifting, masked by the operator compensating with the setpoint until one day the compensation hits its limit and is exposed.
Changing roll diameter or nip width, the same total force gives very different line pressure. Adjusting line pressure must consider both outgoing dryness and felt life — the two move in opposite directions: higher line pressure raises dryness and lowers steam use, but felt life drops and the risk of crushing and web breaks rises. Rule of thumb: each 1 percentage-point rise in outgoing dryness cuts steam use 4%–5%, but felt life may drop 10%; whether it pays depends on the ratio of steam price to felt price.
IV. Accumulator: pre-charge pressure is the lifeline
The accumulator (mostly bladder type) does three things in press loading: absorb pump flow pulsation, compensate minor leakage during pressure holding, and keep the nip from collapsing instantly on sudden power/pressure loss. It has only two core parameters: nitrogen pre-charge pressure P₀ and the working oil pressure.
Rule of thumb for pre-charge: about 80%–90% of the system's minimum working pressure, and never exceed a certain ratio of the system's maximum pressure (generally P₀ ≤ 0.9 × system minimum working pressure). Insufficient pre-charge → poor pulsation absorption, slow response; excessive pre-charge or bladder rupture → oil pushes straight to the gas valve, bladder bottoms out and is damaged, even nitrogen leaks into the oil line. Judgment method: after shutdown and depressurization, measure bladder gas pressure (with a dedicated charging tool) and compare with the label value; if pressure drops sharply and nitrogen cannot be charged, the bladder is likely ruptured.
For sizing, the effective accumulator volume is estimated by V = (P₁/P₀)^(1/n) gas compression, where n is the polytropic index (isothermal 1.0, adiabatic 1.4). Rough rule: to compensate a single press-holding leakage ΔV, the lower the pre-charge P₀ the larger the releasable volume, but too low loses holding ability. In practice the volume is back-calculated from "pressure drop < 5% within 30 s of holding".
| Symptom | Criterion & action |
|---|---|
| Pressure drops rapidly with motion, frequent re-pressurization | Insufficient pre-charge or bladder rupture → stop and measure P₀; if below label value by 10%, find leak / replace |
| Loud pump noise, system runs hot | Mostly air ingestion or clogged filter → bleed air, check return-line seals |
V. Pneumatic system: FRL and ejector (vacuum generator)
Pneumatic circuits are simpler than hydraulic, but only on the premise that the air source is well treated. The FRL triplet is the baseline: Filter removes water and particles (the auto-drain must work), Regulator stabilizes pressure (pressure fluctuation directly causes unstable cylinder speed), Lubricator supplies oil (can be omitted for oil-less cylinders, but ordinary cylinders cannot run dry). Many "slow or crawling pneumatic motion" faults root in undrained water in the FRL or a fatigued regulating spring.
Paper machines use many ejectors (Venturi principle): compressed air creates a high-speed jet at the throat, drawing the suction chamber to negative pressure, used to hold the sheet, drive the vacuum transfer roll, and clean. Advantages: no moving parts, fast response; disadvantages: high air consumption, sensitive to supply pressure. If vacuum is insufficient, first check supply pressure, then whether the nozzle is worn or clogged with paper dust.
Ejector selection looks at two parameters: maximum vacuum (generally around -80 kPa, depending on supply pressure) and maximum suction flow (determines how large an area it can hold). Each 0.1 MPa drop in supply pressure can cut vacuum substantially. Compared with vacuum-pump systems, it suits intermittent, low-flow, fast-response applications (transfer, cleaning), not high-flow dewatering main systems.
| Air-source fault | Symptom | Action |
|---|---|---|
| Water content | Slow cylinder action, spool rust, water in exhaust | Check auto-drain, add dryer |
| Insufficient pressure | Insufficient cylinder thrust, vacuum drop | Check regulator, line leakage |
| Insufficient lubrication | Cylinder bore scoring, early seal wear | Restore lubricator; for oil-less cylinders, forbid oil |
It essentially converts the kinetic energy of compressed air into negative pressure in the suction chamber, so air consumption rises with the square of vacuum. In design, calculate "total air consumption vs vacuum demand"; blindly enlarging the nozzle only burdens the compressor. Ejectors are cost-effective at intermittently used transfer points; continuous high-flow dewatering is still cheaper with vacuum pumps.
VI. Fault tree: from symptom to root cause
| Symptom | High-probability causes (check in order) | Verification action |
|---|---|---|
| Pressure won't rise | Accumulator failure → relief valve internal leak → pump wear → cylinder internal leak | Measure P₀, check relief temperature rise, listen for pump noise |
| Slow motion / crawling | Air ingestion → clogged filter → FRL water accumulation → stuck speed-control valve | Bleed air, check filter differential, inspect drain |
| High oil temperature | Cooler clogged / insufficient water → large internal leak → relief valve often open | Feel cooler inlet-outlet temperature difference, check relief valve |
| External oil leak | Loose fitting → aged seal → hose pulse fatigue | First tighten fittings, then replace seals, batch-replace hoses at end of life |
| Insufficient vacuum | Low supply pressure → worn/clogged nozzle → suction chamber leak | Measure supply pressure, disassemble and inspect nozzle |
VII. Sealing and leakage management
In hydraulic systems, "internal leakage" is more hidden than "external leakage": external leakage is visible (floor oil stains, high oil consumption), internal leakage is invisible (oil pumped out returns to tank via relief valve or cylinder internal leak, showing as pressure won't rise, high oil temperature, slow motion). Seal life is affected by temperature, speed, and particles. O-rings fear extrusion and compression set; Glyd rings/Step seals suit high-pressure reciprocation; rotary points use oil seals; high temperature uses FKM rather than NBR.
| Component | Typical life | Replacement criterion |
|---|---|---|
| O-ring / Glyd ring | 1–3 years | Replace on disassembly if extrusion, nicking, or hardening is seen; do not reuse |
| High-pressure hose | 3–5 years | Batch-replace by pulse count or age; wire break is not visible externally |
| Filter element | Replace on differential-pressure alarm | Never use past due; handle within 8 h of alarm |
Pulse fatigue breaks the wire layer, invisible externally. Replace in batches at end of life (generally 3–5 years or by pulse count); don't wait for a burst that stops production — a burst not only halts output but sprays oil into the web and electrical cabinets. A tip: axial fine cracks or local bulging on the hose cover are pre-burst warning signs.
VIII. Maintenance points and criteria
- Oil analysis: every 6–12 months take particle count + moisture + viscosity + acid number; trend matters more than absolute value.
- Pressure test: measure system maximum pressure and accumulator P₀ monthly, record trend.
- Air-source drainage: confirm FRL auto-drain action daily, manually drain once a week.
- Preventive replacement: seals and hoses batch-replaced by age; downtime repairs only tighten fittings.
- New-oil filtration: make-up oil through filter cart; never pour drummed oil directly.
- Proportional-valve calibration: zero/gain calibration once a year; handle any drift >2%; never "compensate with the setpoint".
IX. Common field misoperations and taboos
A large share of hydraulic/pneumatic failures are introduced by the "maintenance action itself". Listing high-frequency misoperations as a taboo list is cheaper than emergency repairs afterward: ① change oil but not the filter — old particles immediately contaminate new oil; ② run on after the filter differential alarm — equals letting the system work in "sandpaper oil"; ③ never check accumulator gas — pre-charge goes wrong unnoticed; ④ run hoses until they burst — wire break is invisible, only a burst exposes it; ⑤ hard-compensate proportional-valve zero drift with the setpoint — hides the hazard until compensation hits the limit; ⑥ leave water in the pneumatic drain un-drained — air source carries water, cylinders and spools rust; ⑦ pour new oil directly without filtration — drummed oil often fails online cleanliness; ⑧ disassemble fittings under pressure — high-pressure spray injures people and contaminates the web and electricals.
The common trait of these eight is "save ten minutes now, pay dozens of hours later". A deeper hazard is "temporary fixes" without closure: bypassing an interlock, increasing a relief setting — if not logged and restored by a deadline, temporary becomes permanent. Every temporary fix must have the four elements "who, when, why, when to restore"; better to be a bit slower than leave a latent defect — many major accidents trace to one unclosed "temporary fix".
| Taboo action | Direct consequence | Correct practice |
|---|---|---|
| Change oil but not filter | New oil contaminated by old particles within hours | Clean tank + replace all filters + re-test after 24 h |
| Never check accumulator P₀ | Pre-charge fails, nip collapses, large pulsation | Measure P₀ at shutdown each quarter, within ±10% of label |
| Run hoses until burst | Oil spray, shutdown, web/electrical contamination | Batch-replace by age / pulse count |
X. Daily defense lines for oil and air source
What keeps failures out the door are three defense lines: source line — new oil filled via filter cart, never pour drummed oil directly; process line — suction/return/high-pressure three-stage filtration + cooler temperature control 40–55 °C + breather desiccant; monitoring line — every 6–12 months oil sample for particle count + moisture + viscosity + acid number, watch trend not single points. Among the three, the monitoring line is cheapest yet most often cut — by the time oil degrades, the cost is already valve/pump replacement. The air source has its own three lines: dryer dewatering, FRL pressure-holding and oiling, end drainage; slacking any one is amplified down the chain.
The pneumatic daily defense line more easily "gets away with laziness but explodes late": confirm FRL auto-drain action daily, manually drain weekly, check regulator outlet pressure stability and lubricator oil level monthly, inspect hoses and fittings quarterly. One undrained pool of water can slow a dozen cylinders in sync, appearing as "all pneumatic actions wrong", hard to locate and easily misjudged as valve failure. So pneumatic-source inspection must not be skipped despite being tedious — it is the "water source" of the whole pneumatic chain's health.
XI. Commissioning and acceptance of hydraulic systems
A newly installed or major-overhauled hydraulic system must do two things before use: "flushing + acceptance", but these are often skipped and it goes straight to pressurized operation, so fresh particles immediately jam the proportional valve. Flushing uses a dedicated flush pump to circulate; the target is online cleanliness within NAS 8 and particle count no longer dropping for two consecutive hours, only then are weld slag, iron chips, and sealant cleared from the piping. Many "valve jams as soon as the new system goes live" faults root in no flushing or sub-standard flushing — not a valve quality problem, but an installation debt.
Acceptance looks at four numbers: system maximum pressure (relief valve setpoint matches label), accumulator P₀ (within ±10% of label), each actuator's action time (deviation from design <5%), and pressure drop over 30 s holding (<5%). Record all four in the acceptance report so that later "valve slowly slows", "pressure slowly won't rise" have a baseline to compare. Without a baseline, all "seems worse than before" cannot be quantified and relies on the veteran's feel — but veterans retire, feel cannot be taken with them, baseline can.
The acceptance report should also attach a "typical operating-condition parameter snapshot": pressure, temperature, flow, and action time under three conditions — full load, 50% load, no load. Any future anomaly can be compared with the snapshot to immediately judge "systematic degradation" versus "normal load-switching behavior", avoiding misjudging normal load fluctuation as a fault. The snapshot's collection cost is near zero, yet its value spans the equipment's whole life cycle — it turns vague complaints like "this machine seems less powerful than before" into comparable data.
A often-overlooked detail is the timing of oil sampling: it should be taken when the system is running stably with normal oil temperature, sampling separately from the return line and the working point, avoiding the transition period just after make-up oil or filter change. Single-point sampling is easily misjudged — freshly added oil temporarily lowers the particle count, falsely suggesting the system is clean. Trend analysis is more reliable: only when three to five consecutive samples show flat or falling particle counts is cleanliness truly controlled; moisture rising slowly from 0.03% to 0.1% warns far better of breather failure or cooler internal leak than a single absolute value.
XII. Mnemonics for this lecture
Hydraulics and pneumatics are not "visible and tangible" like mechanical parts; they rely on discipline not feel. Condensing this lecture into a few mnemonics, checked before and after shifts, beats flipping through the manual in a pinch:
- Dirty oil causes seven-tenths of failures; stop the moment the filter differential alarms — don't run the system in "sandpaper oil".
- Measure accumulator P₀ every quarter; label ±10% is the red line; wrong pre-charge makes the whole curve wrong.
- Line pressure is total force divided by nip width; if it won't rise, check the accumulator and filter first, don't rush to the valve.
- Pneumatic baseline is FRL; water in the air source slows the whole machine, rusts spools, scores cylinders.
- Replace hoses in batches by age; wire break is invisible, only a burst exposes it — that is a time bomb.
- Oil change must include filter change, temporary fixes must be closed out — save ten minutes, pay dozens of hours.
One sentence summary: 80% of hydraulic/pneumatic failures are in "cleanliness and air source", the remaining 20% in seals and accumulators. Control these two ends and the system is mostly stable. Link it with the CBM system of M01 — a continuously rising filter differential, slowly climbing oil temperature, gradually falling P₀ are all lead-time signals of "knowing it will fail before it does"; feed them into condition monitoring and it is no longer "repair the pump after it breaks".
Key takeaways
- Hydraulics deliver "force", pneumatics deliver "speed"; use hydraulics for heavy-load positioning, pneumatics for fast reciprocation. Common enemy is contamination and leakage; about 80% of hydraulic failures root in dirty oil.
- Cleanliness is a system state not a new-oil state: servo/proportional valve systems NAS 7–8, ordinary NAS 9–10; oil change must clean tank + replace all filters + re-test. Control oil at 40–55 °C, prevent water and air ingress.
- Press line pressure = total force ÷ nip width, three-layer loop (command → proportional valve → cylinder → transmitter feedback); if it won't rise, check accumulator and filter first; calibrate proportional valve yearly.
- Accumulator P₀ ≈ 80%–90% of system minimum working pressure; measure P₀ at shutdown, check bladder if below label by 10%. Back-calculate volume from "pressure drop <5% over 30 s holding".
- Pneumatic baseline is FRL: dewater, stabilize pressure, oil; the ejector "exchanges air for vacuum", sensitive to supply pressure, suits intermittent low flow.
- Fault-tree order: pressure won't rise → accumulator/relief/pump/cylinder internal leak; slow motion → air ingestion/filter/FRL; insufficient vacuum → air source/nozzle.
- Internal leak is more hidden than external; replace hoses in batches by age, wire break not visible externally; select seals by material (FKM for high temperature).