Autopilot

Hoping to build a Hot Air Balloon autopilot (no, I am not planning to defy science by flying against the wind!!). I need to build a controller that can automatically energise a relay that will cause a burner solenoid to operate in order to maintain level flight. I plan to use a BMP180/BME280 sensor to achieve the barometric pressure and use high/low thresholds from the required altitude i.e. + or - 1 Hectopascal, which is approximately +/- 9 metres. There are a few considerations - 1) a Hot Air Balloon does not react to burner input for 20-30 seconds, which complicates the firing algorithm 2) the length of burn will need to adjust to enable to minimal 'sine wave' possible (frequent short burns rather than less frequent longer burns). The above scenario would mean a pilot achieving level flight manually, then operating the autopilot. The system will need safety measures that would prevent the balloon bursting though the lower threshold without taking appropriate action. Later phases are likely to include the ability to climb or descend to specific altitudes, incorporate a GPS to show speed and track, enter current QNH pressure value (Mean Seal Level pressure) via a touchscreen or phone app and, later, adding LoRa functionality to pass data to the ground.

I would welcome any comments on this potential project.

Sounds entirely possible. I'd start with the altitude sensor, display, and user inputs. I assume your power source is a battery. You probably don't need a relay where a MOSFET will do for the burner solenoid.

Does the autopilot have any control over descent?

You could get better altitude control if your fuel valve was proportional rather than ON/OFF. This article: Homeroasters - Home Roasting Coffee Community - Discussion Forum: Proportional Solenoid Valve for Propane talks about a reasonably-priced ($70-$100) proportional valve.

Sadly, a proportional valve isn't an option, as restricting the flow of liquid propane would result in freezing of the valve. Whether I could use a vapor feed instead of liquid, that might be an option, but it would require a hardware change at the burner.

In phase 1, the autopilot would only be an altitude hold. Other functions would be manually operated.

I would imagine your hot air balloon would come under aircraft flight rulings and any modification would need to be certified.
Arduino could in this instance also come under life threatening situations and not allowed.

Further to this , altitude reference is zero at point of liftoff and will not be constant above the ground during the course of any flight.

What happens if it fails? Try this link for more information n the restrictions etc. http://asl.stanford.edu/wp-content/papercite-data/pdf/Suskho.Tedjarati.ea.AERO2017.pdf

This is no different than building an automobile autopilot using an Arduino. Do you have any clue as to your liability when the balloon crashes? Right now all the liability lies with the owner and pilot. How will your family react when a court awards the family of the dead passenger several million dollars and you are one of those being sued?
Even if you are the pilot and taking a friend for a flight and crash?
Who will do the inspections of your equipment and certify it?
Paul

bumpydog:
I would welcome any comments on this potential project.

Its an aircraft, so in most places in the World there would be significant certification and testing requirements for electronics that automatically fly an aircraft.

Having been in the odd hot air balloon, and noting how diligent the pilots are, the only purpose I could concieve of for a 'autopilot' of this type is if the pilot wanted a snooze.

The other issue of course is that if you need to ask for guidance in a hobby forum, do you think you have the skills and experience to implement this 'autopilot' in a safe and realible manner ?

srnet:
Its an aircraft, so in most places in the World there would be significant certification and testing requirements for electronics that automatically fly an aircraft.

Having been in the odd hot air balloon, and noting how diligent the pilots are, the only purpose I could concieve of for a 'autopilot' of this type is if the pilot wanted a snooze.

The other issue of course is that if you need to ask for guidance in a hobby forum, do you think you have the skills and experience to implement this 'autopilot' in a safe and realible manner ?

Let me be a little more specific. I am a commercial balloon pilot with 38 years experience and thousands of hours logged. I was the project manager for ten World Record flights in the smallest categories of hot air balloons. The balloons to which this autopilot will be deployed are 'certified' as Experimental. The purpose of an autopilot is two-fold - 1. maintaining a level flight optimises fuel economy (for duration records) and 2. it relieves the constant burden of the pilot having to reach up to operate the burner throughout a 10+ hour record flight when there are many other aspects of the flight that require total concentration.

bluejets:
I would imagine your hot air balloon would come under aircraft flight rulings and any modification would need to be certified.
Arduino could in this instance also come under life threatening situations and not allowed.

Further to this , altitude reference is zero at point of liftoff and will not be constant above the ground during the course of any flight.

No certification required (see my other reply). The altitude reference is not set at the ground height at lift-off. As the balloon will be flying above the transition layer, it will be referenced to standard QNH i.e. 1013.2HPa (as are all other aircraft in the sky above the transition layer). Regardless of this, these flights normally take place in Class G Airspace (uncontrolled) and would generally take place under high-pressure weather patterns where there will be no anticipated pressure changes during the flight. A change of 1HPa, and the resulting drop or climb of 30 feet is no big deal. Despite the device also having a manual override, I plan to add other safety measures i.e. alarm if lower altitude threshold is busted and not recovered or if variometer (rate of descent) exceeds a specific rate. This is in addition to the audible standard ‘sink rate’ alarm that is built into other instruments being carried. This autopilot is NOT a commercial offering. It is for one-off record or long distance flights with extremely experienced users.

Paul_KD7HB:
This is no different than building an automobile autopilot using an Arduino. Do you have any clue as to your liability when the balloon crashes? Right now all the liability lies with the owner and pilot. How will your family react when a court awards the family of the dead passenger several million dollars and you are one of those being sued?
Even if you are the pilot and taking a friend for a flight and crash?
Who will do the inspections of your equipment and certify it?
Paul

Thanks for your reply, Paul. As you will see in my other replies, this is for experimental one-off usage and the autopilot (maybe using that word encompassed all thoughts of the pilot paying no attention to what is going on) is a supplementary tool to optimise fuel consumption. At no stage will the pilot not be in full control of the aircraft and take appropriate measures should any device onboard fail. Every part of the fuel system on a balloon is fully redundant. In terms of inspections, I do this myself (as a European Aviation Safety Agency (EASA) Part-66 Flight Engineer) as well as a 38 year commercial hot air balloon pilot.

Where I lack experience is microcontroller programming, hence me being here.

This kind of thing will need a lot of testing. I would get a second Arduino and build a "balloon on a bench": a simulator that you can test your code against. Of course, it does mean that you may be chasing a bug in the simulator code, rather than the controller.

It would likely be worth having the sim have the ability to store flight data so you can play back the same scenario repeatedly as you refine the main program.

bumpydog:
Thanks for your reply, Paul. As you will see in my other replies, this is for experimental one-off usage and the autopilot (maybe using that word encompassed all thoughts of the pilot paying no attention to what is going on) is a supplementary tool to optimise fuel consumption. At no stage will the pilot not be in full control of the aircraft and take appropriate measures should any device onboard fail. Every part of the fuel system on a balloon is fully redundant. In terms of inspections, I do this myself (as a European Aviation Safety Agency (EASA) Part-66 Flight Engineer) as well as a 38 year commercial hot air balloon pilot.

Where I lack experience is microcontroller programming, hence me being here.

MY old company used to build all the circuit boards for Mt. High Oxygen and Supply in Redmond, Oregon. I know their system has 5 ft. altitude change detection. If you are using their system, then you may be able to get the altitude from their controller.
Paul

wildbill:
This kind of thing will need a lot of testing. I would get a second Arduino and build a "balloon on a bench": a simulator that you can test your code against. Of course, it does mean that you may be chasing a bug in the simulator code, rather than the controller.

It would likely be worth having the sim have the ability to store flight data so you can play back the same scenario repeatedly as you refine the main program.

True. I would also plan for the system to be checked in-flight using LEDs to tell the pilot when to burn, before adding the burner controls, to check functionality and performance. We are well-versed in over-the-top testing when dealing with world record attempts.

Paul_KD7HB:
MY old company used to build all the circuit boards for Mt. High Oxygen and Supply in Redmond, Oregon. I know their system has 5 ft. altitude change detection. If you are using their system, then you may be able to get the altitude from their controller.
Paul

These flights are for duration/distance and will, therefore, not require an oxygen system. We’d be happy to stay within a 30 ft. (+/- 1 Hpa) corridor, which is well within the capabilities of the BME280, for example.

My point is that you will likely be astounded by how many defects you can cram into a little bit of code. Given the time and money that true testing in a balloon will take, I would want to remove as many of those bugs as possible on the ground.

I went through something like this building yacht systems - the simulator was a life saver.

I am a hotair bvalloon pilot (and flight instructor) myself and know about hose vehickes well. I am highly interested in this poject nad would love to be involved. I am sure itr is possible. So far the only autopilotr working is the good old Comstock Auto Pilot (https://www.alifeintheair.com/comstock-autopilot-in-action) - so the expience is there. All modern burners are prepared for solenoid valves which are ready available by the manufacturers.

Please contact me, let’s do this!

And an additional comment: As there is anyhow a pilot ready and vigilant, even if the autopilot is engaged, we do notz have to fulfill a lavel of reliability which is comparable with commercial aviation. As soon as the balloon starts to climb fast or to descent, any pilot would immediately notice and disengagfe the autopilot and go on manual. So it really is worthwile a project.

Last comment, sorry guys: The purpose of such an autopilot is not for the piulot taking a snooze, but he might have to attend to ATC matters, or just pure simple wants to take some pictures, eat a sandwich or, in the worst case, attend to passengers feeling unwell.