Keeping the Lights on with Python

Video thumbnail (Frame 0) Video thumbnail (Frame 2055) Video thumbnail (Frame 5093) Video thumbnail (Frame 6874) Video thumbnail (Frame 7905) Video thumbnail (Frame 12449) Video thumbnail (Frame 14460) Video thumbnail (Frame 15580) Video thumbnail (Frame 16729) Video thumbnail (Frame 18503)
Video in TIB AV-Portal: Keeping the Lights on with Python

Formal Metadata

Keeping the Lights on with Python
Title of Series
Part Number
Number of Parts
CC Attribution - NonCommercial - ShareAlike 3.0 Unported:
You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal and non-commercial purpose as long as the work is attributed to the author in the manner specified by the author or licensor and the work or content is shared also in adapted form only under the conditions of this license.
Release Date

Content Metadata

Subject Area
Scott Reeve - Keeping the Lights on with Python We are using Python to help the National Grid in the UK to balance electricity production and usage. We do this by installing Python powered devices at customers sites that allow us to monitor and set control criteria to automatically turn on and off power consuming and producing devices when there is a mismatch between electricity supply and demand. In this talk we will be talking about how and why we have used Python, as well as where in our system we would like to use Python. ----- We are using Python to help the National Grid in the UK to balance electricity production and usage. We do this by installing Python powered devices at customers sites that allow us to monitor and set control criteria to automatically turn on and off power consuming and producing devices when there is a mismatch between electricity supply and demand. These devices talk to our Python powered cloud based system using the 3g network, giving us near real-time monitoring of our customers assets. Our entire infrastructure is written in Python, from our billing systems, data analytics systems and customer portal all the way through to our on site industrial system interfaces. In this talk we will be talking about how and why we have used Python, where we have had problems, as well as where in our system we would like to use Python and why we cannot. We will also be talking about what we are going to do next, moving our system from near real time monitoring to near real-time control, using Python for both system modelling and control. We will discuss how we are using Python to creating a system that monitors the balance between electricity supply and demand many times per second and is able to provide a corrective control based on the sum of the output of a dynamic set of our customer sites and the challenges that presents.
Service (economics) Demon Arm Service (economics) Red Hat Bit Line (geometry) Food energy Software Utility software Utility software Hydraulic jump Hydraulic jump Computer architecture
Disintegration Workstation <Musikinstrument> 1 (number) Feldtheorie Water vapor Field (computer science) Product (business) Frequency Different (Kate Ryan album) Velocity System programming Physical system Execution unit Electric generator Scaling (geometry) Chemical equation Gender Structural load Projective plane Line (geometry) Electric power transmission Type theory Frequency Computer animation Software Factory (trading post) Quicksort
Point (geometry) Metre Trail Ewe language Game controller Group action Dynamical system Dependent and independent variables Multiplication sign 3 (number) Shift operator Product (business) Neuroinformatik 2 (number) Frequency Response time (technology) Mathematics Term (mathematics) Ring (mathematics) Software Energy level Aerodynamics Physical system Task (computing) Dependent and independent variables Electric generator Continuous track Moment (mathematics) Thresholding (image processing) Electric power transmission Product (business) Type theory Process (computing) Frequency Fluid statics Computer animation Personal digital assistant Logic Factory (trading post) System programming Website Asynchronous Transfer Mode
Point (geometry) Code Set (mathematics) Mereology Event horizon Computer programming Front and back ends Mixture model Frequency Natural number Energy level Diagram Software testing Selectivity (electronic) Computer architecture Social class Metropolitan area network Dependent and independent variables Frequency response Particle system Malware Computer animation Right angle Data logger Quicksort Freeware Diagram Resultant Spectrum (functional analysis) Library (computing)
Game controller Code Mass Mereology Coprocessor Computer programming Revision control Mixture model Goodness of fit Term (mathematics) Single-precision floating-point format Software Computer hardware Diagram Summierbarkeit Condition number Game controller Real number Structural load Software developer Computer program Plastikkarte Variable (mathematics) Formal language Microprocessor Computer configuration Computer animation Software Website Software testing Right angle
Dynamical system Group action State of matter Correspondence (mathematics) Mereology Vibration Computer programming Frequency Protein folding Term (mathematics) Profil (magazine) Energy level System programming Software testing System on a chip Information security Physical system Area Installation art Dependent and independent variables Scaling (geometry) Mathematical analysis Moment of inertia Bit Instance (computer science) Flow separation 10 (number) Frequency response Type theory Arithmetic mean Process (computing) Personal digital assistant Blog Factory (trading post) Network topology Website Musical ensemble
welcome all to using patented less on arms G that is his constant lapping
thanks and so this tool that is really about well I guess all the celebration the last talk with excellent about the Arctic and so like I guess is a complement in that we talk very little about the architecture but it's more about talking about the problem that we're solving and why we liked using Python so much to celebrate and so on the technical director at line jump and we've been going is a complete proper for a couple of years and so long so that line jump is the utility in the UK and we're able to trade energy and the thing that we're interested in a in his death aggregation services and we're going to talk a bit more on that in a minute but we work with small companies and in Association National and so the national grid in the UK and the high volatility network and just for England and Wales and they work to maintain that and also they try to keep supply and demand of electricity balanced and that's our where we come in so and
if we go back to to 18 70 and more 1933 in the UK when 1st National Lab power network was created you had a big mostly coal-powered power stations that created a large amount of power which is then distributed through high-voltage lines across the country going into lower voltage lines eventually down to the factories smaller factories and consumers now that works well back in 18 70 and 19 thirties but there things that changing now we've got concerns about climate so we don't have to a hold fired power stations or some other big stations for that matter their new types of fat power generation coming in like wind and solar generation we've got some wave power and other things like aerobic gender digested where you stand poised to products and animal manure to generate gasses and then the power engines which produce lectures were got landfill gas which is methane coming off the waste and to to produce electricity again and just now uh things like actually coming and so the power network has to change to meet these new challenges so this is the kind of thing that we're starting to see so we're getting and much more distributed power systems and you can soul of an old system arose projects in there that all sorts of different ways of producing power in their work on a much smaller scale the amount of electricity they generate a smaller they can come down these of high-voltage lines a much more local systems so we need to start controlling the more likely to balance balance the grid and the way and what I mean by that great so if you
think back to the original power stations you had a big and coal fired boiling water turning through turbines that would turn a huge self and then you go back to the magnetic field theory if you turn itself the magnetic field you can generate electricity and the speed of that so is proportional to the frequency of the AC current you get out so what we're trying to do is to keep the shaft moving at a constant velocity so we get 50 hertz so what we have to think of if there's too much load on the shelf saying you've got to me like syllable for example power station is broken down then the shaft is going to start slowing down there's not enough power to push the frequency of electricity will go down and vice versa if it's the middle of the nite and no one's really using electricity and suddenly you wind starts growing very strongly in Scotland this happens sometimes and you get too much wind power so there's more supply than demand and said the frequency starts to go up so
let's on summarized here so when supplies into 2 bonds which is where we want to be everything is at 50 and then as supply has outstripped by demand the frequency goes down and when supply is higher than demand in the frequency goes up and really what we're trying to do is keep living 49 . 5 and 15 . 5 in reality we keep in a much tighter bound than that but when we start getting to these points then we start having problems and electronic equipment will stop working in some cases and if things get really bad to get a complete blackout very not all we want to so that's what we coming so as I
said earlier the national grids job is to balance the supply and demand so they have products in which we work with customers so we work with lots of small customers you on big enough to go to the national grid directly so we coordinate their response time and then sell them as a group together to deliver different products to the national grid and along with then working on for the last year think that to and implement for the last year with customers in a static response so when things again quite bad so when frequency goes down to 49 . 7 about 50 . 3 we get customers produce electricity and to turn things on if the frequency is gone which they use electricity we get to turn things off and viceversa when the frequency of the good goes to high we get people to produce electricity to turn their production down for customers we use electricity to turn their usage of electricity of 1 100 such so this a lot more and how we do that um but then also in the future we're going to be starting to look at things like and dynamic products where we track the grid frequency so we keep in a much tighter bound and then you get 2 seconds to respond to a changing grid frequency and then enhanced response which is responding very quickly and that's coming out in the future and the next 1 to 2 years and then you have to respond to a change in the because you've been 1 2nd so the only thing capable of doing that moment is probably about and turn the power that we have there so what exactly do we do and wages Python coming up because we work with lots of small customers we have to go in and still a panel on each of their sites and the same with generators or with factory equipment and we have the PLC in there furthermore logic controller which is an industrial computer effectively that that you use for automation tasks and we have data being backed with 3 G mode and we have what's called the about the pi type device where what Python 6 um customer site level and also we're starting to use things like what to control and measure customer criminal so we have power meters which sits on the power lines out equipment so that we can measure what's going on or how much electricity that using and when we do this we we monitor power every 2nd so we can see second-by-second how much power at a cost of producing or using and we measure frequency on site so we we know what happened in terms of frequency on-site 10 times a 2nd and we also measure what our system is during intersect so we can keep track of the weather it's performance of now
but as you know last talk very similar architecture to the 1 being at in the last talk I'm not talk about it quite so much and other than to say that the the data logger there is pure Python and everything to the right the is also pretty much Python except the front-end stuff and everything to the left of the data logger so that's of controlled and we'll see there in a mixture of C and then we'll see code results on a 2nd uh which is some level of of programming by diagram really and and then we're going just to which he communicates through a particle called much faster than the 19 seventies so with that got those next class set and we can also control lot there and a sort of point out the data logger doesn't just log data also sets of the L C so you have a semi-autonomous response to 2 events so via the LCC the part that responds with frequency goes out of the spectrum was talking about earlier and it's also the PLC that that controls the African acid we would want to dynamically columns frequency response so that's kind of how we do things why do we use point
uh well it's foster develop this is been developed within 2 years very often prefer in that's given its introduction now in the last few years which I I I come from the C and C + + background I don't think we could have done this if would you see and there's a huge selection of free libraries and and preaching to the converted I guess this set ACE transformed the way that we can do things especially with things like more and the testing and I'm a big fan of like that we used part test in all our testing and stay largely we use quite so I can come to Europe nature
so where we've been having problems with Python and that's the question we have been having massive problems with Python we've been moving to racing recently and will cut down on load on our embedded processors depresses the customer sites by 5 10 fold by using an Cairo so that's been really great and we think it's not quite elegant it could be that that seems to be changing a lot of each new version of Python so really enjoying using that my guess the biggest problem we have is that finding good Python developers in London terms most people can sympathize with so that's kind of really problem all
things with using where would we like to use Python but as I said before I like to talk yesterday and microgrid investment used in microprocessors and the diagram on the right is part of the the PLC software that rather earlier I spent days and days without around PLC supplies debugging this software uh it's it's tricky and it's very very tricky this is the kind of code you get a new card protest answers around the very easily so you end up having to find create every single situation by mixture of the forcing variables within the program and present conditions where violent hardware to time-tested is going to respond you want to respond in each and if anyone has any ideas for getting us out of using the LCC cannot be really really pleased to hear them especially to using Python and as also say we've been using embedded control the seat bag again that's not quite nice in the
so what are we going to do next and this is a test case from last week and I spoke about dynamic frequency response of if you look at the 2nd laughter and you see this in the text profile was injected into 1 of our analysis is the site so this is a solar panel array with the 800 kilowatt actually attached to it so the battery is charged by the solar array during the day and then we can discharge it when the customer can get the best prices for that is the so what we've done here is we've got charged systematic a SOC stands the state of knowledge and so 140 at means 70 % charge set to the systems attached and so what we've done is injected different frequencies into that and if you look at the top 1 the kilowatt hour so that's that's a lot out of the back the mirror is the frequency so frequency goes down the article about he goes up and pretty much the frequency as he goes up and down and you can see the battery discharge charges against that we use and different ways so once you've done that next job is gonna be delivering dynamic response and which I haven't mentioned now this is the response to a local level so this is just a battery responding to the grid and what we're looking to do next is well is to have a response across several pieces of equipment so we might have engine and a factory and baby solar farms as well and what we're looking to do is have each 1 of them produce their own individual response which when summed together creates this kind of group response this means that a lot more uh customers can get involved in this program and we can give that larger of a lump of response power to the national grid so it kind of helps if you come and that the more part and as well and for thank listening questions sentence to the Watkins skins is reading them you've used was this so we had our 1st sight schools last year we're now in the tens of installs um probably by the end of the year will be moving up to hundreds of schools and so on from there and in terms of power and sites range from about 100 kilowatts of small sites up to the a couple of megawatts of big site so if you compare that with the power station power station might be hundred megawatts of whereas with revising very quickly at so it's getting bigger and bigger emotions I understood this correctly from what you said earlier but is it the case that when uh demand falls so much that the frequencies of risk of rising all will demand rises the frequencies of folding that you will get a large users of electricity choose more or less yeah that's right and is the possibility of doing this not I guess you can do that we very easy with huge users but with embedded systems in all kinds of devices in people's homes you could presumably get all the things like which isn't the heating systems and so on to use more or less approach uses of possibilities yeah yes so when we started and the industrial scale because that it's easier to have a viable business that scale but as we refine our technology we can move into smaller and smaller producers and so in the end hopefully yes will be of use and supply so if you think of a fridge for instance which is only on some of the time but if you had a thousand produce you could predict the baseload that the aggregated thousand features used earlier yeah that's that's what we're aiming toward from performance was security consideration because it remains in these systems are affected you could discover was the blogger that music and yes I could cure it is a big consideration yet so yeah we can turn on of these big factories in area that is something we think about a lot of that yet I don't think the talk about a bit too much but I yeah that and say is it the case that when the freedom to drop you respond but what's what happens if you and somebody else's bonds in somebody else's bonds is there risk of you had you not like your you that that is very true actually and has happened so we will employed by the national grid so they know how much the response will be so they contract out set responsibility you right and that has been seen on a few occasions where every corresponds at once and you go from having low frequency having high frequency etc. yeah that's that's 1 for the national grid to sort luckily at the that's they moving to the young dynamic responses because of course they're much more granular so they make correct things much more quickly so you don't get that yeah yeah and I should probably also mention that when we're moving away from these rotating shaft type systems that power generators in the coal fired power stations we're losing a lot of inertia in the system which is stabilizing effect of big itself turning acts like a flywheel whereas with things like wind they don't have a formula such because their electronic inverted so yeah as the UK reduces that would we to become more stable so we have to be more careful to stop this swinging up and down the tree frequency emotions along this yes answer questions if anyone wants afterward so yeah alone