welcome everyone to the part two of our optimized master class series great to see many returning and apologies for the small technical issues last week and welcome to those joining for the first time so in part one we looked at why load management is the foundation of what is broadly spoken as smart ev charging and the first building block for running cost efficient scalable and grid friendly networks if you missed it you can of course catch up on our website the on-demand class today we're zooming in on what we believe at flexi charge we are specialists of and are recognized for it it's actually what sets us apart from other software vendors is the local load management this is where resilience and performance happen in the real world not just in the cloud so it's about real -time optimization faster response rock solid uptime and connecting and integrating battery solars and grid signals so today robert our cto will exactly explain how it is how it works and how cpos can use it today to future prove their operation as always if you have questions you can ask them in the chats section or in the question sections we'll keep an eye on this all right and now enough introduction let's dive into it robert the floor is yours yeah thank you chris um thanks for the introduction welcome yeah welcome welcome everybody or welcome back so yeah welcome back um yeah also apologies uh from from from my side about um we had a few technical issues last week basically it's a it's a funny story because somehow it relates to um let's say the topic today because uh there were um internet connectivity issues um on my side um i'm sitting in germany and yeah so to that I can relate later because one of the things or the topic today is local load management and one of the key advantages for local load management is obviously that it does not rely on your internet connection at site so these problems we had for the webinar obviously you don't want that for your charging infrastructure and you want a reliable operation so yeah I'm sharing my screen let's dive into it is the screen visible Chris go ahead yeah as Chris said the topic is local load management we dive into a few things first of all I'm going to give a quick introduction into what load management architecture are actually available and which are are used today independent of what what we are offering at Flexi charge we obviously offer all but there are advantages disadvantages among the different architectures then we'll dive into local load management that's a topic for today and first question is where can you where can you deploy it where can it live in your in your technology stack locally obviously then we look into what are the fundamentals of a load management architecture so everybody talks load management what at the end of the day is actually load management because there's a there's a there's a not complex but there is a technology stack behind it and one of the fundamental things in load management is actually load management is actually a control and control theory and I'm gonna dive a bit into that so you get a bit of a basic understanding how it works because then it's easier to understand why latencies and real-time control actually matter and where load management actually provides benefits over for example cloud-based load management and the latencies and real-time control is is is is what is really really important here in in in in in in control then as a decentralization is something we're going to look into it's an implicit advantage of local load management reliability security grid operator perspective there are some interesting developments in Germany and some recommendations from yeah authorities how energy energy and energy and load management should be operated at a critical infrastructure and charging infrastructures then we look a bit into the interfacing options for for for for local load management and then conclusions okay let's let's start with the architecture introduction we had that in our previous master classes we touched upon it several times so one of the options how you can implement load load management is based on the load management is based on a on a cloud setup so you have a number of charging stations you connect them to your back end cpms charge point management system and then the charge point management system provides some form of load management often you have to be careful here whether that is a load management or smart charging from the cloud but in principle it can be provided by the cloud flexi charge obviously it also has a solution for for cloud based load management but the key point here is that everything all the control the distribution and the decision how much power charge power is allocated to each charging station to each charge point to each ev is decided in the cloud somewhere on the server so when you say cloud that's a bit abstract so behind the scenes obviously there's some software running somewhere on the server in the internet right so that's what we refer to as cloud based load management then you can have something which we call a hybrid setup we see this also these days more and more more and more but a few cpms systems they also provide ways of how you can integrate for example local assets in and in particular for example local external load into a cloud based load management topology architecture in that case again the charges are connected to your cpms or through the flexi charge harmony cloud based solution often this is via mobile connection and then you have a parallel path so imagine the scenario you have a number of charging stations in front of the building the charging stations are connected to your cpms the charging stations are connected by sim card which is a classic use case to the cpms to the charge point management system but electrically the charging stations are connected to a same grid connection point as for example a building and in that case you're not talking static load management anymore we are talking dynamic load management and we are talking that a building and the charging infrastructure a number of charging infrastructure a number of charging charging stations are sharing a common one grid connection obviously which has a limit and then the goal of dynamic load management is obviously to um to measure the consumption in real time of the building and then whatever is left allocate for charging right so in this case you need integration of the meter or you need to get the information how to charge how to charge the charging right so in real time you need to make the information so much the building is currently consuming or you can measure also at the grid connection what is the load there but you need to transport that information to your cloud server and we call it a hybrid setup if the charging stations are connected via sim card mobile connection or any internet connection directly to the cloud server and we use for example the flexi charge gateway to transport power meter information power meter information external load information also to the cloud server we merge it in the cloud and then the whole dynamic load management the whole control of the system is at site sorry at in the cloud and there are power meters internet connected power meters several ways how you can transport power meter data into the cloud but that is commonly referred to you we refer to it as a hybrid setup the left hand side you have the cloud setup and then finally the last category of setup is basically the local load management and there's also different ways and they differ basically by the networking architecture right on the left hand side is basically um you have again the classic case you have a charging station or you have charging station at the right hand side you have a grid connection right there needs to be obviously some incentives to for load management yeah so let's assume you have more physical charging power installed at the side which is often the case at public charging hubs so let's say you have eight iptronic high power chargers 400 kilowatt but you only have whatever one megawatt or or 1.5 megawatt grid connection and in that case you have a grid connection and in that case you need some form of a load management system so and typical scenario is chargers are connected via mobile network sim card to your cpms they communicate via ocpp obviously you do all the authentication all the operations all the charging charging charge charging management the requirement so the connect direct connections monization it is so to say via octop but then you install locally is chargers port empirical ambassadors wireless operating system of an operation program and then you can connect to the charging stations with a log input and you reach关 employee at the deny operative with your OCPP data stream, right? So you are not having a proxy or something in between, which is a cache, but you have a completely separate communication pass via Modbus to the charging stations, right? And then the gateway or the local controller can be connected via an external router. In FlexiCharge case, there's a router inside the FlexiCharge gateway. It is connected to some load management operation platform, dashboard control system, which is, for example, the FlexiCharge Harmony platform. And then, I mean, this becomes very, very, very crucial. And we get to that point later when you, for example, need to provide DNO control. So network operator needs to be able to measure what is the current consumption at the site and needs to, in case of grid stability issues or to prevent blackouts or whatever, needs to, for example, have an interface to reduce the load at site. There's the, on the right-hand side, there's a, there's a, there's a different way. You can also obviously have the whole communication of the charging stations and the, and the gateway via one router. So then you don't have the, the, the Wi-Fi connection and not the Wi-Fi, the mobile connection. In this case, you would have the, in this case, you would have a complete, let's say, charging stations connected to a switch by a LAN local network. You have a local controller. You might have a router to provide internet access for the charging stations and for the gateway to connect to the Harmony platform. In this case, the, the, like the load management operation solution and to the CPMS. So that's for the charger. In this case, you can do Modbus connection to the chargers, but you can also do, for example, OCPP connection to the charger. So you can, you can, you can, you can control or you can, can, can allocate charge power via Modbus, like on the left-hand side, or you can even do it by, by OCPP. There's advantages, there's advantages, disadvantages, whether you do it by Modbus or OCPP. We'll cover that in one of the next masterclasses. And then again, you can provide a DNO control interface to be able to, to, to act and react on activation reduction signals from, from the DNO. So important fact, important point, local load management. And now that's a key happens locally. So there's a local controller, which is providing load management independent of your OCPP stream and also independent of your, um, internet, uh, connection. So that's a bit closing loop towards the, the issues we had in the webinar left last week. So now let's look into the fundamentals of, no, uh, first point is where can you, is there alternatives where you, where you can deploy your local load management at sites? And there are obviously, um, advantages. Everybody knows there's charging station manufacturers in the DC world. We see that. Yeah, you see it. You have some vendors which provide load management. It's often from the cloud. It's often from the cloud. I haven't seen. Yeah. Yeah. There are some also with local load management, but it's very, let's say, limited. Yeah. So in the AC world and in particular for residential use cases and often also for, let's say, multi apartments. Um, so for the whole residential segment, there are, and we all know them, there are providers, which, which, which, which, which, have charging station manufacturers, which provide local, local load management in the charging station. So this is just to say, okay, where in your stack. Can load management live? And one way is to have it with your charging station. Obviously it's a plug and play solution. Yeah. But it's closed and it's often proprietary system. So if you buy from vendor a, and when the a provides you with load management, you have a, special vendor, vendor, vendor. So if you have a supply chain of different AC and DC charging station manufacturers, you will have. Maybe you chose them based on that. They will have load management, but then you have to maintain three, four, five different load management solution, which are often close proprietary. And that. can create some vendor login um there they often have a limited feature set it's often difficult to integrate with other local assets so for example if you want to integrate a stationary local battery at the site as a charging hub then you rely on that charging station manufacturer vendor to integrate with a particular battery yeah that's a drawback you can also have and get load management with other assets so peripheral assets next to the charging station for example a battery in the solar inverter again you have a plug-and-play solution it's very suitable for simple use cases we see that a lot in residential cases or like really really simple um destination charging sites um they often have a good specific feature set so that's that's clear so batteries they try to really build the load management around the battery so to also provide excuse me to to to provide all the use cases all the features around the battery often they miss specific features for charging stations um for example what we call adaptive charging so the system actually recognizes how much charge power has been allocated to a charging station and what is the actual charge power and different other things as well as soon as you talk about scheduling for example of charging sessions prioritization and things like that it gets to its limit right but some use cases simple use cases it could well work and it's limited integration with other vendors and assets obviously if you have different battery vendors you have the same problem problem with the charging stations again maybe a battery vendor a you can integrate with charger b but not with charger a um so it's it's it's messy so it's often limited um with respect to integrations with other assets and vendors um and often limited feature set and here i in particular mean the features that with respect to set with respect to to to charging yeah because we see it as a process from the perspective of a charge point operator then you can have a dedicated load management controller or a dedicated load management platform i would even call it um where again um you can have and you have plug and play solution you have a dedicated load management energy management solution which covers all use cases that's very very important so a dedicated load management controller or load management solution or suite like the flexi charge harmony platform it covers all use cases so it really takes the the how to say the the the top view right we have charging we have battery we have solar we have polarization we have building load so we we it tries to cover the the the the complete Hall issues weίναι Lord level appr to our display here because you bounce over the that reverend and then the trailer main like so it covers all use cases it's a complete fit feature set and if you go for example with flexi charge then you have an open system which obviously also provides interfacing options with all the relevant assets all the relevant assets at the site for charge point operator even covering dno network operator interfacing obviously the disadvantage and we have to be clear about it is you have an additional component not only additional hardware component but you also have an additional software component right so um that is that is one compromise you have to do but then um obviously it it um it uh it it comes with a with a whole bunch of of advantages specifically again if you're a cpo and you are you are having a bit of a supply chain um not only um you're counting on on on just one vendor okay so we'll focus on the dedicated load management controller and we are diving into fundamentals of a load management architecture now so what is actually the underlying technology in load management and i can tell you the underlying technology in load management is control so it's classic control you have a particular limit which is the grid limit yeah here at the at the at the left hand side or it can also be a temporary soft limit basically uh for example the the the the contracted maximum limit you have with your with your network operator or with your energy supplier but very often in the most and in the most of the cases it is the actual grid limit and then then you have assets behind the grid limit which is for example your charging stations you might have local batteries you might have a building and obviously you have the grid connection and what is happening is that in any And it's actually doing that by measuring, obviously, the whole time. What is the actual consumption? Is it at the grid limit? And if it's at the grid limit, then here this should symbolize the grid limit. Then it gets a feedback. So it has a limit. It calculates, am I above the limit or below the limit? If I'm above the limit, then I have an arrow. So then the system needs to correct for it. And it corrects for it by, for example, limiting or reducing the charge power of a charging station or charging stations or charging sessions. Or it limits the error by discharging a battery. Right? So the aim of a load management system is to always reduce the error here. So make the error actually zero or have a negative error. Yeah? If you have a negative error, then actually you're consuming less than you actually, then you're always below the grid limit. And the classic technical approach, engineering approach to cope with that problem is a controller. Yeah? So you deploy a controller and then all the, and I don't want to dive into it. We can have a whole semester on control theory, obviously, but there is a controller, which is a dynamic system, which is measuring very, very frequent. It gets feedback signals in this error changes all the time and it needs to control it. It's the same. Like when you fly on the plane and it's landing, right? It gets disturbance. And then the, the flight control is trying to keep the plane stable. Yeah? And this is exactly the same here. And the key with local load management. And this is important to understand is the complete control loop. So everything here, which is a dynamic system like on the plane is locally. It's locally. And the important fact, so from technical perspective, it means if everything is local, you have minimum. And latencies. And plus you have. And very limited. Asynchronous city. Which means things are happened. Deterministic and deterministic means that. The, the, when you put a set point and you get the feedback, so that basically one cycle here from, from putting a set point in and then getting a feedback and the error back to the control loop. And then getting a feedback and the error back to the controller. Deterministic means it happens always at a predefined time. In the predefined time period. So the system can basically say within a second, one control loop is done and that can rely on it. Right? This is what means something is deterministic. And if something is deterministic, then it's real time. So the definition of real time is that it's deterministic. And that you can ensure with the complete control loop. And that means your system is reliable and it is secure. Yeah. And if we, for example, now look into cloud-based load management, this is a completely different story. Because in the cloud-based system, you have the whole controller, the logic, yeah, making decisions, the algorithm. The algorithms executed in the cloud. You have the set points transported over van wide area network, the internet, you have the assets locally, and then you have the feedback again, transported over. Wide area network internet. And what is the problem here? The problem is obviously that there are. The disturbances. And so on and so forth. So to cope with that is really challenging. Yeah. And that means things are not deterministic. And they are specifically not deterministic because the, what we call round trip delay. So a charger sending something to a controller and then getting back to the set point, you cannot rely because it depends on network load, on routing, all kinds of things. Right? So therefore, this cannot be deterministic. And that means things are not deterministic. And it depends on network load, on network load. And so on and so forth. And so on and so forth. And so on and so forth. And so on and so forth. And so forth. And so forth. The key is better. And the user. And third. And now, for example, this is something to very large. you still have an active load management system. And in particular, this is important when you have buildings or when you operate logistic centers, depots. The most important thing is that your e-truck is charged and your load management is operational and your fuse doesn't blow if the internet connection is maybe temporary, not available. It's so important. Okay. Another advantage of local load management is actually decentralization and edge intelligence. So we didn't talk about it too much yet, but in some of our other masterclasses, you can look it up. In the case of FlexiCharge, every local gateway we install is part of an IoT architecture. So they are all interconnected in an IoT architecture. So they can provide reliable, secure, local load management, but still they are interconnected and we can provide overlaying central load and energy management. For example, building virtual power plants or having customers building their own logic on top of a central interface to all their load. local site controllers, right? So the advantage of this decentralization is also really, really important. So because you do not have a single point of failure, we are back a bit to reliability, what I said earlier. So there is no central server. There's no central backend or something where your complete load management relies on. Imagine you have hundreds of sites, maybe a thousand, thousands of chargers, and they all rely on. So the CPMS to provide load management that can go wrong. So here the advantage of the implicit advantage of local load management with the controller is there is no single point of failure, right? If one fails, all the 99 other sites, they will still run. Okay. So there is improved resilience and fault tolerance actually built in implicitly. It scales nicely, right? Because with every gateway, with every local controller you install, you install local processing power, local management, local memory. So it really scales, yeah? Without the need to whatever scale your CPMS or processing resources in the cloud. You have reduced latencies. We talked about it. You have real time. You have enhanced data privacy because you can really decide what data to transport, yeah? Upwards and which data you keep local. And often it is the case that you sample at really, really high data rates. You get lots of contacts, local contacts, and that local contacts and local information is actually processed locally and the load management is based on it, but it's not transported upwards to a data. So that's a database. Operational autonomy. Operational autonomy. Obviously, yes, if you have a blackout, that will still work and we'll come to that point in a second because it's really, really important from grid operator and grid stability perspective. There's local context awareness, of course. And finally, there's also reduced communication costs because local load management. And also in terms of user experience. And also in terms of user experience. As faster you can sample data. As faster you can get the data from a charger, for example, for actual charge power, right? Whether you read that every 15 seconds or every 30 or every one minute via OCPP meter values because you are constrained on communication costs, yeah? You can't send every 20 seconds. You can send every 20 seconds. You can send every 20 seconds or your 30 seconds or your 30 seconds. You can get OCPP data every 20 seconds or your 30 seconds or your 30 seconds or your 30 seconds. That's a really, really big advantage of local load management. quicker you read, as quicker it can react, act the load management. And for example, and that is directly related to the user. So for example, if you measure the actual charge power of an AV at any point in time, and now we have a DC charging process, which is everything, but not a constant charge power, it changes really dynamically, but the system can read it every second, right? From the charger, then it can quickly react and it allocates as much as it needs for charging, right? And then it can adapt and everybody else can actually get more. So it's the most optimum distribution of charge power you can actually get, right? And often it's also the case with respect to context awareness that you get more information. For example, if you integrate a charger by Modbus, then you get transported via OCPP. So that's the decentralization part. Now comes a really important fact. There has been an interesting position paper from the VDE, which is a German tech association based in Berlin. It's one of the largest associations in Europe. And they are focused on standardization, safety, development of technical guidelines. And they have published a position paper in March. And I can only recommend reading that. Unfortunately, there's only, I didn't find an English version yet, but I translated a few sections in here. And what they say is, if there is grid operator connection at the charging site, there's no way things have to happen local. So what you can see here. So what you can see here is, on the left-hand side, you basically have the, the, the, the, the phone B, which is a distribution network operator interface. And then, then you have the customer and the assets and you have a local router. And the thing they don't want is that something goes through the internet to a cloud-based load management is coming back and then goes to the asset. Yeah. So that's exactly what I said earlier. Yeah. So that's exactly what I said earlier. That the local, the complete signal flow and the complete local control happens as a complete control happens locally. And that's basically what they say in this position paper. And they have, they, they have four sort of, um, standpoints here, which is operational resilience, resilience and, and grid stability. So that even in case of an internet outage, your system needs to be operational. And this is completely, I mean, the, the, the, the argumentation is crystal clear because what will happen in case of a blackout or what is very probable to happen in case of a blackout is that you have an internet outage. And if you have an internet outage, how the heck you want to ensure grid stability. If they control it. If they control it, they want to have that happening locally. So they interface to the local controller with the DNO telecontroller. And. And. They also, the, the whole DNO sort of architecture is based on completely separated wide area networks. Sometimes in, in it's, it's, it's, it's, it's, it's, it's, it's, it's even older radio frequency networks, which are completely independent on the internet. So they send their signals to dedicated, secured, independent networks. Um, to actually ensure that in any case, if there's grid stability close to blackout or whatever, they can still control the assets and the sites, which are DNO controlled. Um, and that is really important. Um, and that's basically what the saying here, prevent disruption in the execution of critical grid. stabilization measures, right? Right. Then. Obviously reliable control and feedback loops. All what I've said, real time control with minimum latencies, and then also timely execution on feedback to the grid operator. So. That. If the grid operator sends a signal to reduce load in the grid for, for the grid at the particular site, that's a grid operator gets a feedback in time. Um, and it's not relying on, um, and it's not relying on internet, maybe latencies or failures, which are very, very probable again. In times of when you're dealing with critical grid stabilization measures, security. I said it, I mean, avoid security risk. There is obviously, uh, uh, a, uh, a, a, a, a, a, a, a, a great, uh, attack vector. Maybe not even one. If you provide a load management through the cloud. of course you need to ensure adherence to strict cyber security standards and there are even now we are on the right hand side regulatory compliance coming yeah as part of the European Union Cyber Security Resilience Act not to talk about NIST 2 compliance and all this for energy relevant data and obviously with local load management local control you're using compliance with and future proofing system designs so that grid operator perspective is in my opinion it's very very important and we will see a lot with that because obviously at the end of the day charging infrastructure can be critical infrastructure and therefore you better get prepared and you better build your systems or your charging hubs or charging infrastructure as future proof systems okay last thing just to say to complete it and I don't know what the time is I think we've spent almost 45 minutes now so I'll come to an end quickly now there's different options how you can how you can integrate with with local or how you can interface your local assets with the local controller can be OCPP can be Modbus can be MQTT we often see custom interfacing for example if you have grid operator even in Germany these days you you still see analog for 220 millivolt current loops being being requested it's fine obviously reliable and secure but there's different ways of how to interface with your local controller and basically with the FlexiCharge gateway and the FlexiCharge suite in general we provide a nice abstraction layer here which we call FlexiCharge Harmony Connect which basically abstracts all the different communication protocols with the assets towards grid operators towards um um um um um um um to be independent of um of the internet you have cyber security via design you have blackout resilience very important for the, for the, for the grid operator regulatory compliance low latency control loops I mean you have even as, I tried to explain earlier I mean you have clear advantages of higher sampling more data more context, knowledge or awareness of the local controller which is obviously then also improving your load management generally the features you have mission critical reliable um in particular for sites where it really matters i mean public charging hub logistics e-truck bus charging depots and so on that's the force you really want that thing be operational independent and independent of your internet uh stability and and really mitigate risk for failures and that you can only reach with the local controller there's no doubt you have edge intelligence we talk that you have this implicit decentralized system architecture and then obviously enhanced data privacy privacy that's when you should when you should use a local controller um and local load management that's it thanks a lot this were like hopefully interesting 45 minutes um and if you have questions uh chris i guess we we still have a few few minutes um one thing one final point um we talked about we talked a lot about the dno network operator interfacing or i i covered it a few time and we have another masterclass now coming up chris will will probably tell us the date at some point the beginning of july and the beginning of july and that will be really focused on dno controls so what interfaces you can expect um how you can solve it um what is it um what are the constraints and so so on so that's what we're going to cover in our our next masterclass i'm looking forward to see you there that's it chris thanks yeah on that note thank you i haven't got any questions uh for now so but you can you know where to find us and uh thank you for for this very uh for this very thorough intense and deep dive uh session robert and yes on the 10th of july we'll uh focus on the next chapter within optimize um very important as you mentioned the grid regulation so it will be about what we call dno control uh 10th of july uh don't worry you will get a heads up all of the of you who had signed up uh so um you won't miss that unless you're on holidays and then you can catch it up afterwards so on that note um thank you very much um thank you very much um thank you very much and have a great afternoon everyone thank you robert yeah you're welcome bye bye