> For the complete documentation index, see [llms.txt](https://delta-robotics.gitbook.io/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://delta-robotics.gitbook.io/docs/hardware/hydraflex.md).

# HydraFlex™ (In Development)

Delta Robotics hydraulic artificial muscle line in development, how it works, how it differs from AeroFlex, and the 500 psi research targets behind it

{% hint style="info" %}
**Development snapshot, not a datasheet.** This page reflects the current R\&D state. Hardware, materials, and test procedures will change quickly, so expect sharp corners and file feedback early. HydraFlex is not shipping and is not for general purchase or integration yet.
{% endhint %}

## What is a hydraulic artificial muscle?

A McKibben-style artificial muscle is a flexible inner bladder wrapped in a braided sleeve, terminated with end fittings and pressurized to generate an axial pulling force. For HydraFlex, the pressure medium is hydraulic fluid instead of air. That trades some plumbing discipline for higher stiffness (fluid barely compresses) and higher force potential in the same diameter.

## Why hydraulics are interesting

### Higher force density

Hydraulics run at much higher pressures than typical pneumatic setups, which puts more force in the same form factor.

### Better controllability

Low compressibility makes pressure control tighter and more predictable than air.

### Portable, decoupled power

As with pneumatics, the heavy parts stay off-joint: the pump, reservoir, optional accumulator, valves, and filtration live off-board while lightweight hoses run to the muscles.

## HydraFlex vs AeroFlex

HydraFlex and AeroFlex muscles are mechanically very similar. Both use a bladder, braid, and end fittings, share geometry-driven contraction physics, and follow the same modular "muscle as a part" approach.

What changes with HydraFlex: the operating pressure target is much higher, the material requirements get harder at the bladder, braid, and termination, and the supporting system shifts from a compressor and regulator to a pump, reservoir, and valves.

If you have not read the AeroFlex overview yet, start there.

{% content-ref url="/pages/e24x9ck04uixfOvWaqP3" %}
[AeroFlex™ (In Development)](/docs/hardware/aeroflex.md)
{% endcontent-ref %}

## Design targets

These are targets for the R\&D program, not promises. HydraFlex is not ready for guaranteed specs.

| Parameter          | Target                       | Why it exists                 |
| ------------------ | ---------------------------- | ----------------------------- |
| Operating pressure | 500 psi (≈ 3.45 MPa)         | High force in compact form    |
| System footprint   | Compact pump + plumbing      | On-board or wearable-friendly |
| Muscle form factor | Modular sizes                | Repeatable, swappable builds  |
| Characterization   | Force/pressure/length curves | Control loops need data       |
| Failure behavior   | Predictable + documented     | No mystery bursts             |

{% hint style="warning" %}
At 500 psi this is well below industrial hydraulics, but it still stores enough energy to cause serious injury. Design and test accordingly, and never exceed the rated pressure of any component.
{% endhint %}

## What we are building

HydraFlex is a muscle plus the ecosystem around it. Today that means bench rigs, off-the-shelf plumbing, and a lot of material testing.

### High-pressure braided muscles

The core focus: a bladder that survives pressure, flexing, and cycling; braided sleeving that holds geometry without shredding the bladder; and end fittings that do not become the weak link.

### A compact pump pack

We start with off-the-shelf pumps, fittings, hoses, and valves, then shrink and standardize once the requirements are real.

### A modular integration story

The endgame is standard muscle sizes, standard fluid interfaces, and reusable mounting hardware.

## What we are researching

### Muscle materials at 500 psi

The muscle is the hard part. We are testing bladder materials and wall thicknesses, braids (carbon fiber versus aramid-class fibers, different picks and weaves), and termination methods (clamps, ferrules, overmolds, and hybrids).

### End fittings and terminations

A high-pressure muscle fails at the interface first. We track leak rate under static pressure and under bending and cycling, and we want repeatable assembly without hand-tuned torque.

### Compact hydraulics

A small pump system needs mandatory pressure relief, filtration and fluid cleanliness, and a serviceable layout.

### Sensing and control

To behave like a real actuator, HydraFlex needs pressure sensing close to the muscle, in-line or joint force sensing, and control modes for pressure, force, and position.

## Where HydraFlex is headed

The roadmap, roughly in order:

1. Prove a repeatable 500 psi muscle build
2. Publish force, pressure, and length characterization
3. Package a compact pump and valve module
4. Release reference designs for joints, limbs, and wearables

## Who this is for

HydraFlex is for you if you want high force density without rigid gear trains, you build at body scale (wearables, limbs, animatronics), and you have the lab discipline to work with pressurized fluid. If you are new to fluidic muscles, start with AeroFlex.

## Further reading

Hydraulic McKibben muscles have a deep research history. Good starting points:

* [Very High Force Hydraulic McKibben Artificial Muscle with a PBO Cord Sleeve (Advanced Robotics, 2010)](https://doi.org/10.1163/016918609X12586209967366)
* [Development of Power Robot Hand with Shape Adaptability Using Hydraulic McKibben Muscles (IEEE ICRA, 2010)](https://doi.org/10.1109/ROBOT.2010.5509489)
* [Development of Contraction and Extension Artificial Muscles with Different Braid Angles (Journal of Robotics and Mechatronics, 2011)](https://doi.org/10.20965/jrm.2011.p0582)
* [A Proposal of a New Rotational-Compliant Joint with Oil-Hydraulic McKibben Artificial Muscles (Advanced Robotics, 2018)](https://doi.org/10.1080/01691864.2018.1464946)
* [Trends in Hydraulic Actuators and Components in Legged and Tough Robots: A Review (Advanced Robotics, 2018)](https://doi.org/10.1080/01691864.2018.1455606)
* [Safety-Enhanced Control Strategy of a Power Soft Robot Driven by Hydraulic Artificial Muscles (ROBOMECH Journal, 2021)](https://doi.org/10.1186/s40648-021-00194-5)
* [Experimental Comparison of Antagonistic Hydraulic Muscle Actuation (Mechatronics, 2022)](https://doi.org/10.1016/j.mechatronics.2022.102737)
* [Experimental Validation of a 7-DOF Power Soft Robot Driven by Hydraulic Artificial Muscles (IEEE RA-L, 2024)](https://doi.org/10.1109/LRA.2024.3394219)

## Related pages

* [AeroFlex™ (In Development)](/docs/hardware/aeroflex.md): The pneumatic sibling, and the best place to start.
* [Hardware](/docs/hardware.md): The rest of the Delta hardware line.
* [Safety essentials](/docs/get-started/safety.md): Hazards that apply to pressurized systems too.