Skip to main content

Hydraulic Efficiency: Variable-Speed Pumps, Load Sensing, and Accumulators

A hydraulic power unit converts electrical power into pressure and flow, and on a heavy handling machine much of that power can end up as heat in valves, relief valves, and a pump that runs whether or not an axis is moving. UTEC Industrial designs, engineers, machines, fabricates, and installs custom material handling systems for aerospace and heavy industry from its Spokane Valley, WA facility, integrating Allen-Bradley PLC and motion control with in-house CNC machining, heat treating, and stress relief. This article sets out where a valve-controlled circuit loses energy, what pressure-compensated, load-sensing, variable-speed, and pump-controlled architectures change, what an accumulator can and cannot recover, and how much of the published evidence applies to stationary industrial machines rather than mobile equipment. The power-unit architecture is chosen at the design and engineering links of the chain, design → engineering → parts machining → fabrication → assembly → weld fatigue → stress relief → drives → controls → tuning → monitoring, and it is proved or disproved at the last two.

Where does the energy go in a valve-controlled hydraulic system?​

The Oak Ridge National Laboratory (ORNL) report on the U.S. fluid power industry describes the basic industrial arrangement as an electric motor driving a pump, lines carrying the fluid to the actuators, and flow control valves controlling the velocity or position of each actuator. Its loss overview names four places where energy goes:

  • The pump. Friction and internal leakage; the report says pump efficiency can approach 90% under ideal conditions but can drop well below 75% under time-varying loads.
  • The lines. For hydraulic systems, line losses that are typically due to head losses associated with hoses and couplings.
  • The control valves. Internal leakage (tare flow) and metering losses from the pressure drop used to control flow rate, which the report compares to using resistors to control electric motors.
  • The supply pressure itself. In a fixed-pressure system, the report states, it takes the same amount of energy to raise a light load as a heavy one, with hydraulic energy equal to pressure times displaced fluid.

Ketelsen and co-authors at Aalborg University and the University of Agder describe the conventional layout the same way: a centralized hydraulic power unit (HPU) feeds all actuators through a common pressure rail, and each actuator is controlled by a throttling valve, "leading to significant losses." They add that in valve-controlled systems energy recovery is generally not possible, due to throttling across the control valve (Love et al. 2012, ORNL/TM-2011/14, pp. 6 to 7; Ketelsen et al. 2019, §1).

How efficient are industrial hydraulic systems, and how strong is that evidence?​

ORNL and the National Fluid Power Association collaborated with 31 industrial partners to set what the report calls a rudimentary order-of-magnitude estimate. Across all industries, it finds fluid power system efficiencies from less than 9% to as high as 60%, depending upon the application, with an average of 22%. Its segment figures are 50% as the typical efficiency for industrial applications, 21% for mobile hydraulics, and 15% for pneumatics, and the report states that "The analysis only includes known data."

Three limits govern how far those numbers carry to a heavy handling machine:

  • The industrial figure has a narrow base. Industrial hydraulic data came from the injection molding and machine tool industries, which the report puts at 17.3% of the industrial hydraulics industry; it then scales that sample to about 1.10 Quads of energy per year for all industrial hydraulics. The report's abstract lists material handling among the tasks of industrial hydraulics, but the industrial energy data came from those two industries.
  • The loss breakdown in the report is simulated and mobile. Its Figure 5, "Energy losses in mobile load sensing (LS) hydraulic application," is labeled an LS simulated energy distribution: actuator work 14%, valve losses 43%, fan and charge 25%, pump losses 11%, and friction and other 6%. It is not a measured industrial breakdown.
  • The improvement projections are an expert estimate. Subject matter experts from eighteen fluid power manufacturers at a 2010 NFPA workshop projected that a 5-year best-practices effort could raise average efficiency from 22% to 27%, which the report computes as a saving of about 0.4 Quads per year.

As engineering reasoning, no published average describes a particular coil upender, log turner, or transfer-car lift, and the efficiency of a given machine is a quantity to measure on its own duty cycle (Love et al. 2012, ORNL/TM-2011/14, Executive Summary, Abstract and pp. 8, 11 to 13).

What does throttling cost besides electricity?​

Throttled energy becomes heat in the oil. ORNL's case study on hydraulic transformers states that throttling losses introduce both energy losses and generation of heat. Tiboni, at the University of Brescia, writes that pressure losses that may occur in the piping and components require a pressure upstream that is higher than the actuators actually need, increasing the power consumed, and, citing earlier work, that in many systems a significant amount of hydraulic energy is dissipated across the pressure relief valve when actuator speeds are adjusted dissipatively.

That heat bears on the pump's operating window. Bosch Rexroth's data sheet for one axial piston variable pump series gives an optimal operating viscosity and efficiency range of 36 to 16 mm²/s, a continuous-operation range of 400 to 10 mm²/s (which, for a VG 46 oil, it equates to about +4 °C to +85 °C), and a short-term minimum of 10 to 7 mm²/s, limited to 3 minutes at no more than 0.3 times nominal pressure. It requires a cleanliness level of at least 20/18/15 to ISO 4406 (the data sheet names no edition), tightened to 19/17/14 at the drain port when viscosity falls below 10 mm²/s, and it gives HLP 46 oil a viscosity of 10 mm²/s at 85 °C. As engineering reasoning, every kilowatt removed from throttling is a kilowatt the cooler no longer has to reject, and a circuit that runs hotter moves its oil away from the pump maker's optimum viscosity band (Love et al. 2012, ORNL/TM-2011/14, p. 9; Tiboni 2023, §1; Bosch Rexroth RE 92711, 2021, pp. 4 to 5).

How does a pressure-compensated variable-displacement pump reduce losses?​

A swashplate axial piston pump varies its output without changing speed. Rexroth's data sheet states that flow is proportional to drive speed and displacement and can be infinitely varied by adjusting the swashplate. Its theoretical relations, without efficiency, are flow q = Vg × n / 1000 (L/min, with Vg in cm³ and n in rpm) and power P = q × Δp / 600 (kW, with Δp in bar); one size in the series, at 45 cm³ per revolution, is rated 68 L/min at 1,500 rpm and 32 kW at 280 bar, as theoretical, rounded values.

With the basic pressure controller, the data sheet says the pump "only supplies as much hydraulic fluid as is required by the consumers": when working pressure exceeds the setting, the pump regulates to a smaller displacement. The controller's setting range is 50 to 280 bar, with a pilot fluid consumption of approximately 3 L/min maximum. Two cautions from the same document apply to any efficiency design built on it:

  • "Pressure controllers are not safeguards against pressure overload." The data sheet instructs the designer to add a pressure relief valve to the hydraulic system.
  • Not all versions of the product are approved for use in a safety function according to ISO 13849 (no part or edition given), and the data sheet directs anyone who needs reliability parameters, such as MTTFd, for functional safety to Bosch Rexroth.

As engineering reasoning, a pressure-compensated pump removes the flow that a fixed pump would otherwise push across the relief valve, but it still holds the full set pressure, and the metering valve still drops the difference between that pressure and the load pressure (Bosch Rexroth RE 92711, 2021, pp. 1, 8, 12 and 47).

What does load sensing add, and how much can it save?​

Rexroth's data sheet calls its load-sensing version a pressure flow controller: in addition to the pressure controller function, an adjustable orifice, for example a directional valve, is used to adjust the differential pressure upstream and downstream of the orifice, this is used to control the pump flow, and "the pump flow is equal to the actual hydraulic fluid quantity required by the consumer." The standard differential pressure setting is 14 bar (range 14 to 22 bar), and relieving the control port to the reservoir gives a zero-stroke ("standby") pressure about 1 to 2 bar higher than that differential, "however, system influences are not taken into account." Pilot fluid consumption is a maximum of approximately 3 to 4.5 L/min, and the variant with no unloading between the control port and the reservoir carries the notice that "The LS must thus be unloaded in the system." As engineering reasoning, holding a set difference across the metering orifice puts the pump pressure at the load pressure plus that difference.

The arithmetic below shows the scale. It uses assumed values and the data sheet's power relation without efficiency, and it ignores pump, line and pilot losses:

  • Assumed duty. One cylinder needs 100 L/min at a load pressure of 120 bar; useful hydraulic power is 120 × 100 / 600 = 20.0 kW.
  • Constant-pressure supply at 280 bar. Pump output is 280 × 100 / 600 = 46.7 kW, and the valve drops 160 bar, or 26.7 kW.
  • Load-sensing supply at a 14 bar margin. Pump pressure is 134 bar, output is 22.3 kW, and the metering orifice drops 2.3 kW.

As engineering reasoning, the advantage shrinks when one pump serves several actuators at different pressures, because the pump follows the highest load and the other branches throttle the difference; in the example, a second branch at 40 bar and 50 L/min would drop 94 bar, or 7.8 kW.

The only quantified load-sensing saving in the federal report is from a crane and is not ORNL's own measurement: ORNL reports that Liang and Virvalo show an efficiency increase in a hydraulic crane from 10.6% to 27.4% using LS pumps. The research for this article found no measured load-sensing saving on a stationary industrial machine in a neutral primary source, and Rexroth's data sheet makes no efficiency claim for its controller (Bosch Rexroth RE 92711, 2021, pp. 8, 14 to 15; Love et al. 2012, ORNL/TM-2011/14, p. 9).

When does a variable-speed pump beat a variable-displacement pump?​

The one industrial comparison read for this article answers that question for a single machine only. Tiboni compared five power-drive architectures on a variable speed and force hydraulic blanking press for semi-finished brass products: a standard fixed vane pump with relief valve, a regenerative valve circuit, a high–low two-pump circuit, a variable-displacement axial piston pump, and a fixed-displacement pump on an induction motor whose speed an AC frequency converter adjusts continuously. The last three all keep the regenerative valve. Each was sized for the same load cycle, a 1.4 s cycle that includes a 0.3 s table rotation with the cylinder waiting. The study found:

  • Regeneration first. Introducing regeneration in the directly controlled valve cut average active power by 50.5% and cut the pump flow needed for rapid descent from 230.8 L/min to 117 L/min.
  • Ranking for this press. The greatest savings were, in order, high–low, then the variable-speed pump, then the variable-displacement pump, and the high–low and variable-speed results differed by only 1.8%. Across all architectures, active power fell by 50.5% to 64.3% against the standard circuit.
  • Why the displacement-controlled pump trailed. Tiboni attributes it to the additional control oil flow that must be continuously ensured to maintain the axial piston pump's flow control.
  • Why high–low did well. The press alternates high flow at low pressure with low flow at high pressure, which Tiboni says is typically the case with punching.
  • Cost differs from power. On Italian energy-market rules, the variable-speed architecture had the lowest energy cost, a saving of 15% against high–low, which Tiboni attributes to the capacitor bank in the drive limiting the reactive power drawn.

Only the variable-speed plant was measured: 14.85 kW average active power against a modeled 16.06 kW, an overestimate of about 8%. The other results come from sizing models. Neglecting motor inertia underestimated the variable-speed architecture's power by 1.63 kW, a difference Tiboni attributes to strong dynamic effects: motor speed fluctuations on the order of 1,000 RPM within fractions of a second, against about 10 RPM for the other architectures. These are one press's results and do not show that a variable-speed pump wins on every duty cycle (Tiboni 2023, §§2.1.1, 2.3 to 2.6, 3.1 to 3.2, 4 and 5).

Do variable-speed drives save as much on hydraulic pumps as on centrifugal pumps?​

Not by the same mechanism. The U.S. Department of Energy's motor and drive sourcebook classes positive displacement pumps with constant-torque loads, for which the relationship between flow and power is linear, and states that energy savings from reduced-flow operation are reduced compared with variable-torque pumps and fans; adjustable speed drives can also save energy on such systems, but significant savings occur only in certain applications, such as rotary-screw compressors with variable loads. Elsewhere it states that the load characteristics of machinery such as positive displacement pumps "often do not favor the use of VFDs," and that in these applications the linear relationship between output, power and equipment speed tends to favor other control technologies.

Among VFD misapplications, the same sourcebook warns that where torque increases at low speeds, such as certain mixing processes, motor power does not drop significantly at lower speeds, and the integral motor fan may not provide sufficient cooling at lower speeds. Two pump-side limits apply as well. Ketelsen and co-authors, citing an earlier study, note that most pump types are not able to run continuously at low rotational speeds (below 300 RPM), and that running at low speed may reduce reliability and durability. Rexroth's data sheet for one pump series notes that a hydraulic system is an oscillating system, that operation at constant rotational speed over a long period can excite its natural frequency, and that this pump's excitation frequency is 9 times the rotational speed frequency.

As engineering reasoning, the case for a variable-speed hydraulic pump rests on matching delivered flow to the cycle, slowing or stopping the pump between moves, rather than on the affinity-law savings quoted for centrifugal pumps. The data sheet's torque relation, M = Vg × Δp / (20π × ηmh), contains no speed term, and as engineering reasoning a fixed-displacement pump holding pressure at low speed still needs the torque that pressure requires, the condition to check against the motor's cooling (U.S. DOE AMO 2014, DOE/GO-102014-4356, pp. 12 and 39; Ketelsen et al. 2019, §5; Bosch Rexroth RE 92711, 2021, pp. 8 and 47).

What do pump-controlled and self-contained actuator drives change?​

A pump-controlled drive removes the control valve from the power path. Ketelsen and co-authors define it as a differential cylinder, a hydraulic supply, an oil reservoir, and auxiliary components, with the cylinder's motion controlled by varying the pump's flow and at least one pump per actuator. Self-contained versions replace the vented tank with a sealed accumulator. The review states that pump-controlled systems can generally recover kinetic or potential energy, for example during load lowering. Its savings figures are per application and per cited study:

  • Tote dumper. A 21% reduction in energy consumption against a valve-controlled system for a given load cycle, in a circuit whose counterbalance valve for load holding was found to preclude energy recovery during lowering.
  • Overall. "Some references report energy savings up to 75% for specific applications," and "The real industrial breakthrough is however yet to come."

Tiboni's literature review reports that Koitto and co-authors studied a closed-circuit, servomotor-driven fixed-displacement pump on a stationary industrial material-handling application that sequentially lifts and lowers a fixed mass, with energy savings of 53 to 87% against a conventional valve-controlled approach; the same summary notes that the actuator tended to vibrate on reaching the desired position, that system pressures varied greatly as the cylinder moved, and that the system's dynamics did not quite meet the stated requirements. The vendor saving claimed for self-contained actuators, and the heat and controllability limits the review names for them, are set out in the EHA article (Ketelsen et al. 2019, §§1, 2, 4.1.2 and 7; Tiboni 2023, §1).

Can an accumulator store recovered energy in an industrial circuit?​

The federal report describes the idea but illustrates it with vehicles. ORNL's hybrid hydraulics case sizes the primary power source for the average power demand and uses hydraulic accumulators for energy storage; the accumulator can store energy during negative power flow, for example when a load is being lowered or a vehicle is braking, and its quantified example is a delivery truck. Tiboni names an accumulator installed in parallel with a fixed-displacement pump as an architecture that could be explored but that is "a poor fit" for the blanking press, and did not study it. This library found no neutral, quantified accumulator saving on a stationary industrial machine.

The component documents set the working conditions. Parker Hannifin's manual for one series of its bladder accumulators lists energy storage and auxiliary power among the intended uses. It states that when the accumulator is used to supplement pump flow, auxiliary power supply or leakage compensation, precharge "is usually set at approximately 90% of minimum system pressure"; it recommends a precharge of at least 25% of maximum system pressure and warns that damage to the bladder may occur if this ratio is not maintained or is exceeded; and it says the hydraulic circuit should be designed to discharge all the hydraulic fluid from the accumulator automatically when the equipment is turned off.

Every energy figure in the rest of this answer is engineering reasoning, not a sourced result. A 10 t mass lowered 2 m releases m × g × h = 10,000 kg × 9.81 m/s² × 2 m ≈ 196 kJ, about 0.05 kWh, before cylinder, valve and accumulator losses. Whatever is stored at shutdown is dumped by the automatic discharge the manual calls for. Each accumulator added for efficiency is also stored energy to control at lockout, as the companion article on stored energy and LOTO describes, and the mixed-power machine article covers sizing a shared power unit to simultaneous flow, with an accumulator for short peaks (Love et al. 2012, ORNL/TM-2011/14, p. 10; Tiboni 2023, §5; Parker Hannifin HY10-2300-M1, 2025, pp. 3, 17 and 21).

What sensing and controls does an efficient hydraulic power unit need?​

As engineering reasoning, an efficient architecture is only as good as the signals that run it, and the intelligence layer of a variable-speed or load-sensing power unit includes the items below; each sourced statement is attributed to its document.

  • Pressure feedback. As engineering reasoning, supply and load pressures for the speed or displacement loop, and gauge points that confirm zero at the accumulator and its safety block. HYDAC recommends a safety and shut-off block that combines pressure measurement, pressure relief, shut-off, manual or electric bleed-down, and locking.
  • Electrical power measurement. Tiboni measured the press's absorbed current and active power upstream of the frequency converter, the measurement that checked the model to within about 8%.
  • Oil temperature and filter condition. Oil temperature sets viscosity, which the pump data sheet ties to its optimal efficiency range, and Rexroth warns that contamination can, under certain circumstances, make control spools stick in position, a risk that filtration reduces but does not rule out; the machine builder should test whether additional measures are required to bring the driven consumer into a safe position.
  • Electro-hydraulic pressure setting. Rexroth's electro-hydraulic pressure control sets pump pressure by solenoid current and, as the signal drops toward zero, limits pressure to the maximum through a hydraulic cut-off, described as a fail-safe function in case of power failure; the data sheet says the control's swivel-time characteristic was optimized for use as a fan drive system.
  • Safety-related parts kept separate. ISO 13849-1:2023 specifies a methodology and provides related requirements, recommendations and guidance for the design and integration of safety-related parts of control systems that perform safety functions, for high demand and continuous modes of operation, and it does not apply to low demand mode of operation. As engineering reasoning, a speed or pressure loop that saves energy remains a control function.

Tiboni's review also reports that application-independent variations in power consumption can be associated with hydraulic system malfunctions. As engineering reasoning, trending kilowatts per cycle beside temperature and filter differential pressure turns the efficiency design into a monitored quantity. The PLC side of a slow pressure loop is covered in the closed-loop hydraulic control article. UTEC Industrial, a Rockwell Automation Recognized System Integrator, builds this layer on Allen-Bradley ControlLogix and CompactLogix controllers with VFD and servo drives, PanelView and FactoryTalk operator interfaces, and EtherNet/IP networks (HYDAC PN#02068196, 2011, p. 4; Tiboni 2023, §§1 and 3.2; Bosch Rexroth RE 92711, 2021, pp. 17, 47 and 48; ISO 13849-1:2023).

Where is hydraulic efficiency decided in the chain from design to monitoring?​

The power architecture is chosen before the first hose is made up, and Tiboni concludes that the choice of power generation architecture "has a very large impact on the energy efficiency and operating costs." In Tiboni's sizing, choosing regeneration and the power architecture changed the pump flow, the motor size and the active power, and the variable-speed design needed a dynamic sizing that included motor inertia. Rexroth's data sheet adds build-stage duties: hydraulic lines designed to avoid exciting the system's natural frequency, and connecting elements and lines that the machine or system manufacturer must match to the application conditions with the necessary safety factors. As engineering reasoning, the links of the chain divide the work as follows:

  • Design and engineering set the duty cycle, the architecture (fixed, compensated, load-sensing, variable-speed, or pump-controlled), regeneration, and whether an accumulator earns its stored energy.
  • Parts machining and fabrication produce the manifolds, pump mounts, reservoir and frames, where line routing and mounting stiffness are set.
  • Assembly fixes hose and line lengths and the resonance behavior the data sheet warns about.
  • Drives, controls and tuning set the pump speed or compensator response and the margin above load pressure.
  • Monitoring confirms the kilowatts per cycle that the design predicted.

UTEC Industrial tests the power unit and its controls on the assembled machine during factory acceptance testing (FAT) and again at on-site commissioning. The proportional and servo valve article covers the metering valves whose pressure drop sets the throttling loss (Tiboni 2023, §§2.3 to 2.6, 3.1 and 5; Bosch Rexroth RE 92711, 2021, p. 47).

What should a specification for an energy-efficient hydraulic system define?​

As engineering reasoning drawn from the sources cited in this answer, "energy efficient" is not testable until the specification states the basis, and for each power unit it should define:

  • The duty cycle by phase: flow, pressure, and duration for each move and each dwell, which is the basis Tiboni used to size every architecture.
  • The architecture and the comparison basis: which alternatives were sized, on what cycle, and whether the claimed saving is modeled or measured.
  • The acceptance measurement: average active electrical power over the cycle, measured upstream of the drive at FAT, with oil temperature recorded.
  • The oil operating window: the viscosity band from the pump data sheet edition in hand, the cooler duty, and the required cleanliness code.
  • An independent pressure relief valve, since the data sheet says pressure controllers are not safeguards against pressure overload.
  • For variable-speed pumps, the minimum continuous speed and the motor's cooling at low speed and high pressure.
  • For accumulators, the precharge basis and the automatic discharge at shutdown.

Each item is a figure or a test that the builder can be held to (Tiboni 2023, §§2.1 and 3.2; Bosch Rexroth RE 92711, 2021, pp. 4 to 5 and 47; U.S. DOE AMO 2014, DOE/GO-102014-4356, p. 39; Ketelsen et al. 2019, §5; Parker Hannifin HY10-2300-M1, 2025, pp. 17 and 21).

Related Articles

References​

  • Love, L. J., Lanke, E., and Alles, P. Estimating the Impact (Energy, Emissions and Economics) of the U.S. Fluid Power Industry (ORNL/TM-2011/14). Oak Ridge National Laboratory, 2012.
  • Ketelsen, S., Padovani, D., Andersen, T. O., Ebbesen, M. K., and Schmidt, L. (2019). "Classification and Review of Pump-Controlled Differential Cylinder Drives." Energies, 12(7), 1293.
  • Tiboni, M. (2023). "Power Drive Architectures for Industrial Hydraulic Axes: Energy-Efficiency-Based Comparative Analysis." Applied Sciences, 13(18), 10066.
  • Bosch Rexroth AG. RE 92711: Axial Piston Variable Pump A10VSO Series 31 (data sheet, 2021-05-17). Bosch Rexroth, 2021.
  • U.S. DOE Advanced Manufacturing Office. Improving Motor and Drive System Performance: A Sourcebook for Industry, DOE/GO-102014-4356. U.S. Department of Energy, 2014.
  • Parker Hannifin HY10-2300-M1: Bladder Accumulators Maintenance and Installation Manual. Parker Hannifin Corporation, 2025.
  • HYDAC PN#02068196: Accumulators Operating and Installation Instructions. HYDAC Corporation, 2011.
  • ISO 13849-1:2023: Safety of machinery — Safety-related parts of control systems — Part 1: General principles for design. International Organization for Standardization, 2023.

Ready to Discuss a Material Handling System?​

UTEC Industrial designs, engineers, machines, fabricates, and installs custom material handling systems for heavy industry, from the stress-relieved structure and drives to the Allen-Bradley PLC controls, tuning, and monitoring that run them, at its Spokane Valley, WA facility. Send UTEC the application, loads, and duty cycle to start a system review.

Request a Quote →

Questions? Call (509) 922-1832 or email sales@utec.co

Last updated on