Hydro turbine control

[Grendel]

Hi,

We have a Francis hydro turbine driving an asynchronous generator.

The wicket gates are manual and there is really no control.

We open the wicket gates, the turbine starts running and at approximate synchronous speed, we switch the generator on the grid. After that we keep the upper water level at setpoint by manually opening and closing the wicket gates.

Now because of difference in head and flow, the Francis has low efficiency outside design range, so we are in the process of removing the Francis and installing a Kaplan turbine.

The asynchronous generator will be replaced by a synchronous one.

I would like to install a PLC that controls the installation.

This PLC has input from water level at three points, frequency and current, wicket gate position and runner blade position.

I would like 2 modes of operation.

1) Grid tied

2) Island.

The PLC needs to be able to:

1) Start up the system.

2) Run the system grid tied and control the wicket gates to keep upper water level at set point.

3) Run Island mode and control the wicket gates to keep frequency at changing loads.

4) Run a flushing program in order to clear the turbine of debris.

5) In the future be able to start a grid cleaner when water level difference over the grid becomes too great.

6) Safely shut down the system in case of a fault or reverse current.

All in all not very complicated, but I am not familiar with programming PLC's and I have no idea what I need.

Would anybody here that is familiar with plc's be so kind as to help me out a bit?

It would be a great help if we could discuss the setup and sort out what I need.

Kind regards,

Hubert

 

[Tony S]

You’ve a fair number of variables to take in to account and so it won’t be cheap. It’s not just the cost of the PLC, the peripheral sensors / actuators significantly add to the cost.

RoB is possibly the guy to give an idea of costs.

 

[Big Blue Bulb]

Two parameters not mentioned in your list are reactive power (measuring) and excitation (controlling). The PLC is controlling real power flow and monitoring for reverse power, would it also make sense to give it supervision of the VAR flow or is that to be preset within the alternator controls?

 

[Grendel]

Hi, yes that would make sense, but I am not too concerned about the generator part of the setup.

I have a large AVR that is capable of delivering full excitation current, so I am planning to use that in stead of the built in exciter.

The AVR comes with a set of controls and safeties for parallel operation and so on.

My biggest problem is the PLC, because the last time I learned programming was at school ages ago and I remember it was in Pascal. Never used it since.

So, I'll probably manage to make the thing switch some relays in the right order after some studying, but for the water level and frequency control, I think I need a built in PID function.

I know exactly what I want this thing to do and what sensors I need to do it, but to get this from my brain into this little black box is beyond my skills.........

 

[Grendel]

You’ve a fair number of variables to take in to account and so it won’t be cheap. It’s not just the cost of the PLC, the peripheral sensors / actuators significantly add to the cost.

RoB is possibly the guy to give an idea of costs.

I already have a hydraulic unit, valves, cylinders and so on to control the wicket gates and runner blades.

I only need a PLC that can do what I want and a bunch of sensors. These are not cheap, especially the ones for water level.

The most expensive part, believe it or not, was the rotary union that sits on top of the shaft and feeds oil to the runner blade adjusting cylinder.

 

[Richard-the-Ninth]

You don't "programme" a PLC, you "upload to it"

I use a couple of PLC's (MItsubishi) and I also have the software, all you do is tell the software what you want it to do, "test it" on the PC then upload it to the PLC.

(sorry I am not offering my services) But my point being you don't really need to know that much of programming just "logic" and since you can test it on the PC you can adjust it till its correct.

 

[Tony S]

You don't "programme" a PLC, you "upload to it"

Do you? Quite frequently I had to made program changes on the fly to keep production going.

Even more fun is changing a dodgy program without telling the operator. It now does exactly what it’s supposed to do and you’re their little hero.

My life in R&D was ruled by a “critical flow” spreadsheet that took up the entire width of my office wall, it was 4 A0 drawings sellotaped together. I was constantly looking for ways of shaving off a couple of seconds here and there.

I never did get the steak promised to me if I could get a part of the process to end within 2 seconds of its cycle time (high inertia load). ½ a second early and the boss wouldn’t stump up.

Another episode, the boss collared me (I used hide from him) to speed up part of the line, “you mean like that?” I walked away laughing.

The bottom line:

You have to know the process and know what you’re doing.

 

[Rob.]

Ok first of all. What's your IO count. (How many digital inputs/outputs. Analogue in/out)?

How are they distributed? Can you contain everything in one cabinet or would it be better for you having a little here and there remote.

Are you requiring a full control system to be built? Or integrated into existing hardwired control.

You mention safety. If you want a safety rated plc you will need this separately to control just the safety functionality. Completely isolated and separate from the other. 

Do you require any form of data management or logging. If so, you may want to look at a front end Scada system.

Everything you've asked for is bog standard work for me. Ask away.

Ok first of all. What's your IO count. (How many digital inputs/outputs. Analogue in/out)?

How are they distributed? Can you contain everything in one cabinet or would it be better for you having a little here and there remote.

Are you requiring a full control system to be built? Or integrated into existing hardwired control.

You mention safety. If you want a safety rated plc you will need this separately to control just the safety functionality. Completely isolated and separate from the other. 

Do you require any form of data management or logging. If so, you may want to look at a front end Scada system.

Everything you've asked for is bog standard work for me. Ask away.

Edit: just noticed you're in the Netherlands, I'd recommend looking at Siemens or beckhoff. They're widely used around there and will have spares and tech support readily available.

 

[Rob.]

You don't "programme" a PLC, you "upload to it"

I use a couple of PLC's (MItsubishi) and I also have the software, all you do is tell the software what you want it to do, "test it" on the PC then upload it to the PLC.

(sorry I am not offering my services) But my point being you don't really need to know that much of programming just "logic" and since you can test it on the PC you can adjust it till its correct

Technically wrong, the ideas in the right place and I think you're just misusing some terminology. 

You write software on a programming unit (laptop etc..).

Compile to machine code (nearly all software packages do this for you) and then download it to the CPU of a plc. 

Not all you can get to simulate on a pc without buying additional emulators.

 

[Grendel]

Hi Rob,

I don't want a second PLC for safety. To be honest, the only thing that can go wrong is that the generator overspeeds in island mode, or that the main breaker does not switch off in case of a power failure.

Let me make a list of all the IO signals and write a sequence of events.

If you would then be so kind as to have a look at it?

Regards,

Hubert

 

[Tony S]

The I/O listing would be interesting to see and to have a think about implementation.

PID feed back loops in Modicon got me so annoyed I wrote my own algorithms.

But since I retired I’ve kept clear of PLC’s. “Tempus fugit” and so I haven’t kept track of developments.

 

[Rob.]

@Tony S PIDs are covered with standard function blocks now. Only really have to modified them if you start going into cascading PIDs. They're still tricky to set up correctly. 

@Grendel if you can get me an IO list and a rough FDS. I can give you a budget price on what you'd looking to pay for hardware and someone to write the software. 

I can't give you a price to bring everything in line with current standards, this would be far too much to do over a forum. 

Pm me if you like. Although I'm sure others are interested, so if there's nothing private we should try to keep it in here.

 

[Tony S]

@Tony S

 PIDs are covered with standard function blocks now. Only really have to modified them if you start going into cascading PIDs. They're still tricky to set up correctly. 

RoB I know standard function blocks are there, they’ve been around for yonks but I didn’t want to use them. I’d got other things in mind, mainly screwing up the holier than thou development team.

 

[Grendel]

Sorry for the format, it was copied from an excel sheet.

I probably need a couple more for future use, or in case I have forgotten something.

List of signals:     

# Name FunctionType in/out   

1 Upper waterlevel before grid ? in   

2 Upper waterlevel after grid ? in   

3 Lower water level ? in   

4 Position guide vanes ? in   

5 Position runner blades ? in   

6 End switch sluice 1 up analog in   

7 End switch sluice 1 down analog in   

8 End switch sluice 2 up analog in   

9 End switch sluice 2 down analog in   

10 Open guide vanes analog out   

11 Close guide vanes analog out   

12 Open runner blades analog out   

13 Close runner blades analog out   

14 Open sluice 1 analog out   

15 Close sluice 1 analog out   

16 Open sluice 2 analog out   

17 Close sluice 2 analog out   

18 Start hydro pump analog out   

19 Sluice 1 cylinder open analog out   

20 Sluice 1 cylinder close analog out   

21 Sluice 2 cylinder open analog out   

22 Sluice 2 cylinder close analog out   

23 Switch on synchro relay analog out   

24 Synchronisation "relay in" signal analog in   

25 Setpoint upper water level analog in   

26 Voltages digital in Via power quality meter

27 Currents digital in Via power quality meter

28 Frequency generator digital in Via power quality meter

29 Inverter power relay analog out   

30 Mains relay analog out   

31 Generator relay analog out   

I'll draw a situation sketch.

 

[Grendel]

This is basically what the system looks like.

 

[Grendel]

 

[Grendel]

Now, the operation should look something like this:

Now, I am not sure how often we will need the Off grid modes.

That depends on future rules, regulations and taxes regarding power generation. In any case, the system should be capable of doing this.

Also, I would like to run a cable from the switch board to my router, so I can monitor and control the system from my computer.

 

[Rob.]

RoB I know standard function blocks are there, they’ve been around for yonks but I didn’t want to use them. I’d got other things in mind, mainly screwing up the holier than thou development team.

I just know you worked with modicon gear. An from what limited work I've done with them, originally you had to write in your own PID algorithm. 

Most of the standard ones now are ok for basic stuff. However on a few fast acting processes I've had to develop my own. For instance a chemical plant where the reaction can go exothermal and jump 400 deg in 4 seconds. That was done with a stand alone safety controller as a tertiarry level protection system.

@Grendel. Are your AI and AO signals 4-20ma or 0-10v? Will the dig in dig out all be 24vdc? 

I also apologise for a delay in response, I'm currently on a project working 8hours behind UK time. 

Your graphset is ok, however i would look at adding in what happens in fault conditions, should you not meet the requisite to jump a step. 

 

[Sidewinder]

Won't this need a grid protection relay?

Also, not sure that you can get away with no safety unless it's  a job at home as DIY.

Inputs 6-9 are down as analogue?

Won't they just be digital end point switches?

 

[Grendel]

@rob,

No problem. I am working abroad myself at the moment and I still have time.

I can always control the system manually and install the control system later.

Dig signals are 24 V. I will look into the sensors later. Most probably 4-20 mA.

@Sidewinder

I planned a relay between the grid and our system. Most power quality meters can be programmed and alarm levels set. They will also switch off that relay in case of power down.

Yes you are right. Should be digital. My mistake.